1 2014-05-22 Nick Clifton <nickc@redhat.com>
3 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
6 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
8 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
9 -> (T)A transformation to integer types.
11 2014-05-22 Teresa Johnson <tejohnson@google.com>
13 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
14 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
15 (gcov_rewrite): Use gcov_nonruntime_assert.
17 (gcov_write_words): Ditto.
18 (gcov_write_length): Ditto.
19 (gcov_read_words): Use gcov_nonruntime_assert, and remove
20 gcc_assert from IN_LIBGCOV code.
21 (gcov_read_summary): Use gcov_error to flag profile corruption.
22 (gcov_sync): Use gcov_nonruntime_assert.
23 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
24 (gcov_histo_index): Use gcov_nonruntime_assert.
25 (static void gcov_histogram_merge): Ditto.
26 (compute_working_sets): Ditto.
27 * gcov-io.h (gcov_nonruntime_assert): Define.
28 (gcov_error): Define for !IN_LIBGCOV
30 2014-05-22 Richard Biener <rguenther@suse.de>
32 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
33 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
34 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
35 and deallocation site.
36 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
37 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
38 passing through the incoming points-to set.
39 (handle_lhs_call): Use flags argument instead of recomputing it.
40 (find_func_aliases_for_call): Call handle_lhs_call with proper
43 2014-05-22 Jakub Jelinek <jakub@redhat.com>
45 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
46 all padding bits in REAL_VALUE_TYPE are cleared.
48 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
50 Cleanup and improve multipass_dfa_lookahead_guard
51 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
52 (core2i7_first_cycle_multipass_begin,)
53 (core2i7_first_cycle_multipass_issue,)
54 (core2i7_first_cycle_multipass_backtrack): Update signature.
56 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
57 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
58 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
60 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
61 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
63 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
65 * doc/tm.texi: Regenerate.
67 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
68 * haifa-sched.c (ready_try): Make signed to allow negative values.
69 (rebug_ready_list_1): Update.
70 (choose_ready): Simplify.
71 (sched_extend_ready_list): Update.
73 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
75 Remove IA64 speculation tweaking flags
76 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
77 speculation tuning flags.
78 (msched-prefer-non-data-spec-insns,)
79 (msched-prefer-non-control-spec-insns): Obsolete options.
80 * haifa-sched.c (choose_ready): Remove handling of
81 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
82 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
83 and PREFER_NON_DATA_SPEC.
84 * sel-sched.c (process_spec_exprs): Remove handling of
85 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
87 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
89 Improve scheduling debug output
90 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
91 (advance_one_cycle): Update.
92 (schedule_insn, queue_to_ready): Add debug printouts.
93 (debug_ready_list_1): New static function.
94 (debug_ready_list): Update.
95 (max_issue): Add debug printouts.
96 (dump_insn_stream): New static function.
97 (schedule_block): Use it. Also better indent printouts.
99 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
101 Fix sched_insn debug counter
102 * haifa-sched.c (schedule_insn): Update.
103 (struct haifa_saved_data): Add nonscheduled_insns_begin.
104 (save_backtrack_point, restore_backtrack_point): Update.
105 (first_nonscheduled_insn): New static function.
106 (queue_to_ready, choose_ready): Use it.
107 (schedule_block): Init nonscheduled_insns_begin.
108 (sched_emit_insn): Update.
111 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
113 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
115 (aarch64_secondary_reload) : LikeWise.
116 (aarch64_class_max_nregs) : Remove CORE_REGS.
117 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
118 (REG_CLASS_NAMES) : Likewise.
119 (REG_CLASS_CONTENTS) : LikeWise.
120 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
122 2014-05-21 Guozhi Wei <carrot@google.com>
125 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
127 (vqdmulhq_n_s16): Likewise.
129 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
131 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
133 2014-05-21 Marek Polacek <polacek@redhat.com>
136 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
138 2014-05-21 Martin Jambor <mjambor@suse.cz>
140 * doc/invoke.texi (Optimize Options): Document parameters
141 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
142 ipa-cp-array-index-hint-bonus.
144 2014-05-21 Mark Wielaard <mjw@redhat.com>
147 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
148 version >= 3 or not strict DWARF.
149 * langhooks.h (struct lang_hooks_for_types): Add
150 enum_underlying_base_type.
151 * langhooks.c (lhd_enum_underlying_base_type): New function.
152 * gcc/langhooks.h (struct lang_hooks_for_types): Add
153 enum_underlying_base_type.
154 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
155 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
156 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
158 2014-05-21 Richard Biener <rguenther@suse.de>
160 * doc/invoke.texi (-flto-partition=): Document one and
163 2014-05-21 John Marino <gnugcc@marino.st>
165 * config.gcc (*-*-dragonfly*): New target.
166 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
167 * configure: Regenerate.
168 * config/dragonfly-stdint.h: New.
169 * config/dragonfly.h: New.
170 * config/dragonfly.opt: New.
171 * config/i386/dragonfly.h: New.
172 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
174 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
176 * tree.def (VOID_CST): New.
177 * tree-core.h (TI_VOID): New.
178 * tree.h (void_node): New.
179 * tree.c (tree_node_structure_for_code, tree_code_size)
180 (iterative_hash_expr): Handle VOID_CST.
181 (build_common_tree_nodes): Initialize void_node.
183 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
185 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
187 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
189 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
192 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
193 flag_reorder_blocks_and_partition.
194 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
196 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
199 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
201 (*addc_r_t): Add new insn_and_split.
203 2014-05-21 Jakub Jelinek <jakub@redhat.com>
206 * omp-low.c (handle_simd_reference): New function.
207 (lower_rec_input_clauses): Use it. Defer adding reference
208 initialization even for reduction without placeholder if in simd,
209 handle it properly later on.
211 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
213 PR tree-optimization/60899
214 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
215 assume all static symbols will have definition wile parsing and
216 check the do have definition later in compilation; check that
217 variable referring symbol will be output before concluding that
218 reference is safe; be conservative for referring local statics;
219 be more precise about when comdat is output in other partition.
221 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
224 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
225 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
226 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
228 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
230 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
231 set writeonly flag only for vars actually written to.
233 2014-05-20 Dehao Chen <dehao@google.com>
235 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
236 and callee count to get clone count.
237 * tree-inline.c (expand_call_inline): Use callee count instead of bb
240 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
242 PR rtl-optimization/61243
243 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
245 2014-05-20 Xinliang David Li <davidxl@google.com>
247 * cgraphunit.c (walk_polymorphic_call_targets): Add
248 dbgcnt and fopt-info support.
249 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
250 * ipa-devirt.c (ipa_devirt): Ditto.
251 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
253 * ipa.c (walk_polymorphic_call_targets): Ditto.
254 * gimple-fold.c (fold_gimple_assign): Ditto.
255 (gimple_fold_call): Ditto.
256 * dbgcnt.def: New counter.
258 2014-05-20 DJ Delorie <dj@redhat.com>
260 * config/msp430/msp430.md (split): Don't allow subregs when
261 splitting SImode adds.
262 (andneghi): Fix subtraction logic.
263 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
265 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
267 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
269 * except.c (switch_to_exception_section, resolve_unique_section,
270 get_named_text_section, default_function_rodata_section,
271 align_variable, get_block_for_decl, default_section_type_flags):
272 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
273 * symtab.c (symtab_add_to_same_comdat_group,
274 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
275 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
277 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
278 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
279 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
280 (c6x_function_in_section_p): Likewise.
281 * config/darwin.c (machopic_select_section): Likewise.
282 * config/arm/arm.c (arm_function_in_section_p): Likewise.
283 * config/mips/mips.c (mips_function_rodata_section): Likewise.
284 * config/mep/mep.c (mep_select_section): LIkewise.
285 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
287 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
289 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
290 EH region of calls to pure functions that can throw an exception.
291 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
292 (copy_reference_ops_from_call): Also copy the EH region of the call if
293 it can throw an exception.
295 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
297 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
298 nested VEC_SELECTs that are inverses of each other.
300 2014-05-20 Richard Biener <rguenther@suse.de>
302 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
304 (extract_and_process_scc_for_name): not here.
305 (cond_dom_walker::before_dom_children): Only process
306 stmts that end the BB in interesting ways.
307 (run_scc_vn): Mark param uses as visited.
309 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
311 * config/arm/arm.md (arith_shiftsi): Do not predicate for
314 2014-05-20 Nick Clifton <nickc@redhat.com>
316 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
317 (msp430_gimplify_va_arg_expr): New function.
318 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
320 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
321 operand 0 in order to prevent confusion about the number of
324 2014-05-20 Richard Biener <rguenther@suse.de>
326 PR tree-optimization/61221
327 * tree-ssa-pre.c (el_to_update): Remove.
328 (eliminate_dom_walker::before_dom_children): Handle released
329 VDEFs by value-numbering them to the associated VUSE. Update
330 stmt immediately for substituted call address.
331 (eliminate): Remove delayed stmt updating code.
332 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
333 possibly late re-numbered vuses.
334 (vn_reference_lookup_2): Adjust.
335 (vn_reference_lookup_pieces): Likewise.
336 (vn_reference_lookup): Likewise.
338 2014-05-20 Richard Biener <rguenther@suse.de>
340 * config.gcc: Remove need_64bit_hwint.
341 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
342 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
344 * config.in: Regenerate.
345 * configure: Likewise.
347 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
349 * doc/extend.texi: Create Label Attributes section,
350 move all label attributes into it and reference it.
352 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
354 * arm.c (thumb1_reorg): When scanning backwards skip anything
355 that's not a proper insn.
357 2014-05-19 Richard Biener <rguenther@suse.de>
359 PR tree-optimization/61221
360 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
361 Do nothing for unreachable blocks.
362 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
363 Improve unreachability detection.
365 2014-05-19 Richard Biener <rguenther@suse.de>
367 PR tree-optimization/61209
368 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
370 2014-05-19 Nick Clifton <nickc@redhat.com>
372 * except.c (init_eh): Fix computation of builtin setjmp buffer
373 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
375 2014-05-19 Richard Biener <rguenther@suse.de>
377 PR tree-optimization/61184
378 * tree-vrp.c (is_negative_overflow_infinity): Use
379 TREE_OVERFLOW_P and do that check first.
380 (is_positive_overflow_infinity): Likewise.
381 (is_overflow_infinity): Likewise.
382 (vrp_operand_equal_p): Properly treat operands with
383 differing overflow as not equal.
385 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
387 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
388 shift simplification where it was intended.
390 2014-05-19 Christian Bruel <christian.bruel@st.com>
393 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
395 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
398 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
401 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
403 * reg-notes.def (CROSSING_JUMP): Likewise.
404 * rtl.h (rtx_def): Update comment for jump flag.
405 (CROSSING_JUMP_P): Define.
406 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
407 of a REG_CROSSING_JUMP note.
408 * cfghooks.c (tidy_fallthru_edges): Likewise.
409 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
410 * emit-rtl.c (try_split): Likewise.
411 * haifa-sched.c (sched_create_recovery_edges): Likewise.
412 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
413 * jump.c (redirect_jump_2): Likewise.
414 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
415 (relax_delay_slots): Likewise.
416 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
418 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
419 * config/sh/sh.md (jump_compact): Likewise.
420 * bb-reorder.c (rotate_loop): Likewise.
421 (pass_duplicate_computed_gotos::execute): Likewise.
422 (add_reg_crossing_jump_notes): Rename to...
423 (update_crossing_jump_flags): ...this.
424 (pass_partition_blocks::execute): Update accordingly.
426 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
428 * tree.h: Remove extraneous template <>.
430 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
432 * ipa.c (symtab_remove_unreachable_nodes): Remove
433 symbol from comdat group if its body was eliminated.
434 (comdat_can_be_unshared_p_1): Static symbols can always
436 * symtab.c (symtab_remove_from_same_comdat_group): Break out
438 (symtab_unregister_node): ... this one.
439 (verify_symtab_base): More strict checking of comdats.
440 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
442 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
444 * tree-pass.h (make_pass_ipa_comdats): New pass.
445 * timevar.def (TV_IPA_COMDATS): New timevar.
446 * passes.def (pass_ipa_comdats): Add.
447 * Makefile.in (OBJS): Add ipa-comdats.o
448 * ipa-comdats.c: New file.
450 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
452 * ipa.c (update_visibility_by_resolution_info): New function.
453 (function_and_variable_visibility): Use it.
455 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
457 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
459 (FOR_EACH_DEFINED_SYMBOL): New macro.
460 (varpool_first_static_initializer, varpool_next_static_initializer,
461 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
462 (symtab_in_same_comdat_p): Correctly deal with inline functions.
464 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
466 * ggc-page.c (ggc_handle_finalizers): Add comment.
468 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
470 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
471 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
472 (ggc_internal_cleared_alloc): Likewise.
473 * ggc-page.c (finalizer): New class.
474 (vec_finalizer): Likewise.
475 (globals::finalizers): New member.
476 (globals::vec_finalizers): Likewise.
477 (ggc_internal_alloc): Record the finalizer if any for the block being
479 (ggc_handle_finalizers): New function.
480 (ggc_collect): Call ggc_handle_finalizers.
481 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
483 (ggc_internal_cleared_alloc): Likewise.
484 (finalize): New function.
485 (need_finalization_p): Likewise.
486 (ggc_alloc): Install the type's destructor as the finalizer if it
488 (ggc_cleared_alloc): Likewise.
489 (ggc_vec_alloc): Likewise.
490 (ggc_cleared_vec_alloc): Likewise.
492 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
494 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
496 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
498 * alias.c (record_alias_subset): Adjust.
499 * bitmap.c (bitmap_element_allocate): Likewise.
500 (bitmap_gc_alloc_stat): Likewise.
501 * cfg.c (init_flow): Likewise.
502 (alloc_block): Likewise.
503 (unchecked_make_edge): Likewise.
504 * cfgloop.c (alloc_loop): Likewise.
505 (flow_loops_find): Likewise.
506 (rescan_loop_exit): Likewise.
507 * cfgrtl.c (init_rtl_bb_info): Likewise.
508 * cgraph.c (insert_new_cgraph_node_version): Likewise.
509 (cgraph_allocate_node): Likewise.
510 (cgraph_create_edge_1): Likewise.
511 (cgraph_allocate_init_indirect_info): Likewise.
512 * cgraphclones.c (cgraph_clone_edge): Likewise.
513 * cgraphunit.c (add_asm_node): Likewise.
514 (init_lowered_empty_function): Likewise.
515 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
516 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
517 (alpha_use_linkage): Likewise.
518 * config/arc/arc.c (arc_init_machine_status): Likewise.
519 * config/arm/arm.c (arm_init_machine_status): Likewise.
520 * config/avr/avr.c (avr_init_machine_status): Likewise.
521 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
522 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
523 * config/cris/cris.c (cris_init_machine_status): Likewise.
524 * config/darwin.c (machopic_indirection_name): Likewise.
525 (darwin_build_constant_cfstring): Likewise.
526 (darwin_enter_string_into_cfstring_table): Likewise.
527 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
528 * config/frv/frv.c (frv_init_machine_status): Likewise.
529 * config/i386/i386.c (get_dllimport_decl): Likewise.
530 (ix86_init_machine_status): Likewise.
531 (assign_386_stack_local): Likewise.
532 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
533 (i386_pe_maybe_record_exported_symbol): Likewise.
534 (i386_pe_record_stub): Likewise.
535 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
536 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
537 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
538 (m32c_note_pragma_address): Likewise.
539 * config/mep/mep.c (mep_init_machine_status): Likewise.
540 (mep_note_pragma_flag): Likewise.
541 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
542 (mips16_local_alias): Likewise.
543 (mips_init_machine_status): Likewise.
544 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
545 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
546 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
547 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
548 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
549 * config/pa/pa.c (pa_init_machine_status): Likewise.
550 (pa_get_deferred_plabel): Likewise.
551 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
552 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
553 (rs6000_init_machine_status): Likewise.
554 (output_toc): Likewise.
555 * config/s390/s390.c (s390_init_machine_status): Likewise.
556 * config/score/score.c (score_output_external): Likewise.
557 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
558 * config/spu/spu.c (spu_init_machine_status): Likewise.
559 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
560 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
561 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
562 * coverage.c (coverage_end_function): Likewise.
563 * dbxout.c (dbxout_init): Likewise.
564 * doc/gty.texi: Don't mention variable_size attribute.
565 * dwarf2cfi.c (new_cfi): Adjust.
566 (new_cfi_row): Likewise.
567 (copy_cfi_row): Likewise.
568 (create_cie_data): Likewise.
569 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
570 (new_loc_descr): Likewise.
571 (find_AT_string_in_table): Likewise.
572 (add_addr_table_entry): Likewise.
574 (add_var_loc_to_decl): Likewise.
575 (clone_die): Likewise.
576 (clone_as_declaration): Likewise.
577 (break_out_comdat_types): Likewise.
578 (new_loc_list): Likewise.
579 (add_loc_descr_to_each): Likewise.
580 (add_location_or_const_value_attribute): Likewise.
581 (add_linkage_name): Likewise.
582 (lookup_filename): Likewise.
583 (dwarf2out_var_location): Likewise.
584 (new_line_info_table): Likewise.
585 (dwarf2out_init): Likewise.
586 (mem_loc_descriptor): Likewise.
587 (loc_descriptor): Likewise.
588 (add_const_value_attribute): Likewise.
589 (tree_add_const_value_attribute): Likewise.
590 (comp_dir_string): Likewise.
591 (dwarf2out_vms_debug_main_pointer): Likewise.
592 (string_cst_pool_decl): Likewise.
593 * emit-rtl.c (set_mem_attrs): Likewise.
594 (get_reg_attrs): Likewise.
595 (start_sequence): Likewise.
596 (init_emit): Likewise.
597 (init_emit_regs): Likewise.
598 * except.c (init_eh_for_function): Likewise.
599 (gen_eh_region): Likewise.
600 (gen_eh_region_catch): Likewise.
601 (gen_eh_landing_pad): Likewise.
602 (add_call_site): Likewise.
603 * function.c (add_frame_space): Likewise.
604 (insert_temp_slot_address): Likewise.
605 (assign_stack_temp_for_type): Likewise.
606 (get_hard_reg_initial_val): Likewise.
607 (allocate_struct_function): Likewise.
608 (prepare_function_start): Likewise.
609 (types_used_by_var_decl_insert): Likewise.
610 * gengtype.c (variable_size_p): Remove function.
611 (enum alloc_quantity): Remove enum.
612 (write_typed_alloc_def): Remove function.
613 (write_typed_struct_alloc_def): Likewise.
614 (write_typed_typedef_alloc_def): Likewise.
615 (write_typed_alloc_defns): Likewise.
617 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
618 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
619 * ggc.h (ggc_alloc): new function.
620 (ggc_cleared_alloc): Likewise.
621 (ggc_vec_alloc): Template on type of vector element, and remove
622 element size argument.
623 (ggc_cleared_vec_alloc): Likewise.
624 * gimple.c (gimple_build_omp_for): Adjust.
625 (gimple_copy): Likewise.
626 * ipa-cp.c (get_replacement_map): Likewise.
627 (find_aggregate_values_for_callers_subset): Likewise.
628 (known_aggs_to_agg_replacement_list): Likewise.
629 * ipa-devirt.c (get_odr_type): Likewise.
630 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
631 (read_agg_replacement_chain): Likewise.
632 * loop-iv.c (get_simple_loop_desc): Likewise.
633 * lto-cgraph.c (input_node_opt_summary): Likewise.
634 * lto-section-in.c (lto_new_in_decl_state): Likewise.
635 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
636 (input_eh_region): Likewise.
637 (input_eh_lp): Likewise.
638 (input_cfg): Likewise.
639 * optabs.c (set_optab_libfunc): Likewise.
640 (init_tree_optimization_optabs): Likewise.
641 (set_conv_libfunc): Likewise.
642 * passes.c (do_per_function_toporder): Likewise.
643 * rtl.h: Don't use variable_size gty attribute.
644 * sese.c (if_region_set_false_region): Adjust.
645 * stringpool.c (gt_pch_save_stringpool): Likewise.
646 * target-globals.c (save_target_globals): Likewise.
647 * toplev.c (general_init): Likewise.
648 * trans-mem.c (record_tm_replacement): Likewise.
649 (split_bb_make_tm_edge): Likewise.
650 * tree-cfg.c (move_sese_region_to_fn): Likewise.
651 * tree-data-ref.h (lambda_vector_new): Likewise.
652 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
653 * tree-iterator.c (tsi_link_before): Likewise.
654 (tsi_link_after): Likewise.
655 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
656 * tree-ssa-loop-niter.c (record_estimate): Likewise.
657 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
658 * tree-ssa-operands.h: Don't use variable_size gty attribute.
659 * tree-ssa.c (init_tree_ssa): Adjust.
660 * tree-ssanames.c (set_range_info): Likewise.
661 (get_ptr_info): Likewise.
662 (duplicate_ssa_name_ptr_info): Likewise.
663 (duplicate_ssa_name_range_info): Likewise.
664 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
665 (unpack_ts_fixed_cst_value_fields): Likewise.
666 * tree.c (build_fixed): Likewise.
667 (build_real): Likewise.
668 (build_string): Likewise.
669 (decl_priority_info): Likewise.
670 (decl_debug_expr_insert): Likewise.
671 (decl_value_expr_insert): Likewise.
672 (decl_debug_args_insert): Likewise.
673 (type_hash_add): Likewise.
674 (build_omp_clause): Likewise.
675 * ubsan.c (decl_for_type_insert): Likewise.
676 * varasm.c (get_unnamed_section): Likewise.
677 (get_noswitch_section): Likewise.
678 (get_section): Likewise.
679 (get_block_for_section): Likewise.
680 (create_block_symbol): Likewise.
681 (build_constant_desc): Likewise.
682 (create_constant_pool): Likewise.
683 (force_const_mem): Likewise.
684 (record_tm_clone_pair): Likewise.
685 * varpool.c (varpool_create_empty_node): Likewise.
687 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
689 * dwarf2out.c (tree_add_const_value_attribute): Call
690 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
691 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
692 instead of ggc_internal_<x>alloc_stat.
693 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
694 (ggc_realloc): Likewise.
695 * ggc-none.c (ggc_internal_alloc): Likewise.
696 (ggc_internal_cleared_alloc): Likewise.
697 * ggc-page.c: Likewise.
698 * ggc.h (ggc_internal_alloc_stat): Likewise.
699 (ggc_internal_alloc): Remove macro.
700 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
701 (ggc_internal_cleared_alloc): Remove macro.
702 (GGC_RESIZEVEC): Adjust.
703 (ggc_resizevar): Remove macro.
704 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
705 (ggc_internal_cleared_vec_alloc_stat): Likewise.
706 (ggc_internal_vec_cleared_alloc): Remove macro.
707 (ggc_alloc_atomic_stat): Drop _stat suffix.
708 (ggc_alloc_atomic): Remove macro.
709 (ggc_alloc_cleared_atomic): Remove macro.
710 (ggc_alloc_string_stat): Drop _stat suffix.
711 (ggc_alloc_string): Remove macro.
712 (ggc_alloc_rtx_def_stat): Adjust.
713 (ggc_alloc_tree_node_stat): Likewise.
714 (ggc_alloc_cleared_tree_node_stat): Likewise.
715 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
716 (ggc_alloc_cleared_simd_clone_stat): Likewise.
717 * gimple.c (gimple_build_omp_for): Likewise.
718 (gimple_copy): Likewise.
719 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
720 * toplev.c (realloc_for_line_map): Adjust.
721 * tree-data-ref.h (lambda_vector_new): Likewise.
722 * tree-phinodes.c (allocate_phi_node): Likewise.
723 * tree.c (grow_tree_vec_stat): Likewise.
724 * vec.h (va_gc::reserve): Adjust.
726 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
728 * config/microblaze/microblaze.c (break_handler): New Declaration.
729 (microblaze_break_function_p,microblaze_is_break_handler): New.
730 (compute_frame_size): Use microblaze_break_function_p.
731 Add the test of break_handler.
732 (microblaze_function_prologue) : Add the test of variable
733 break_handler. Check the fnname by BREAK_HANDLER_NAME.
734 (microblaze_function_epilogue) : Add the test of break_handler.
735 (microblaze_globalize_label) : Add the test of break_handler.
736 Check the name by BREAK_HANDLER_NAME.
738 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
740 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
741 microblaze_is_break_handler test.
742 (call_internal1,call_value_intern): Use microblaze_break_function_p.
745 * config/microblaze/microblaze-protos.h
746 (microblaze_break_function_p,microblaze_is_break_handler):
749 * doc/extend.texi (MicroBlaze break_handler Functions): Document
750 new MicroBlaze break_handler functions.
752 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
754 * doc/extend.texi (Size of an asm): Move node text according
755 to its @menu entry position.
757 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
759 PR tree-optimization/61140
760 PR tree-optimization/61150
761 PR tree-optimization/61197
762 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
764 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
766 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
768 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
770 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
771 __SIZEOF_INT128__ is defined.
773 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
775 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
776 (rs6000_delegitimize_address): Use it.
778 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
780 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
781 inplace argument. Store the new address in the original MEM when true.
782 * emit-rtl.c (change_address_1): Likewise.
783 (adjust_address_1, adjust_automodify_address_1, offset_address):
785 * rtl.h (plus_constant): Add an inplace argument.
786 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
787 when true. Avoid generating (plus X (const_int 0)).
788 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
789 in-place. Pass true to plus_constant.
790 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
792 2014-05-16 Dehao Chen <dehao@google.com>
794 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
796 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
799 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
801 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
803 2014-05-16 Dehao Chen <dehao@google.com>
805 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
806 optimize_function_for_size_p.
807 * regs.h (REG_FREQ_FROM_BB): Likewise.
809 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
812 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
813 negt_reg_operand cases.
814 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
816 * config/sh/predicates.md (cbranch_treg_value): Simplify.
818 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
820 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
823 2014-05-16 David Malcolm <dmalcolm@redhat.com>
826 2014-04-29 David Malcolm <dmalcolm@redhat.com>
828 * tree-cfg.c (dump_function_to_file): Dump the return type of
829 functions, in a line to itself before the function body, mimicking
830 the layout of a C function.
832 2014-05-16 Dehao Chen <dehao@google.com>
834 * cfghooks.c (make_forwarder_block): Use direct computation to
835 get fall-through edge's count and frequency.
837 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
839 * config/arc/arc.c (arc_init): Fix typo in error message.
840 * config/i386/i386.c (ix86_expand_builtin): Likewise.
841 (split_stack_prologue_scratch_regno): Likewise.
842 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
843 word from error message.
845 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
847 * ira-costs.c: Fix typo in comment.
849 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
851 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
853 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
855 * varpool.c (dump_varpool_node): Dump write-only flag.
856 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
858 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
859 write-only variables.
860 * ipa.c (process_references): New function.
861 (set_readonly_bit): New function.
862 (set_writeonly_bit): New function.
863 (clear_addressable_bit): New function.
864 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
865 fix handling of aliases.
866 * cgraph.h (struct varpool_node): Add writeonly flag.
868 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
870 PR rtl-optimization/60969
871 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
872 Calculate costs for this case.
874 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
876 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
877 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
879 2014-05-16 Richard Biener <rguenther@suse.de>
881 PR tree-optimization/61194
882 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
883 bool patterns ending in a COND_EXPR.
885 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
887 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
889 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
891 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
892 where we were unable to cost an RTX.
894 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
896 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
899 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
900 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
902 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
904 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
905 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
907 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
908 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
910 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
911 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
913 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
916 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
917 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
919 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
922 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
923 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
925 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
926 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
928 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
929 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
931 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
934 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
935 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
937 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
938 ZERO_EXTEND and SIGN_EXTEND better.
940 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
941 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
943 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
946 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
947 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
949 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
950 costs when costing loads and stores to memory.
952 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
953 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
955 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
958 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
960 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
962 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
963 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
965 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
967 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
969 (aarch64_rtx_mult_cost): New.
970 (aarch64_rtx_costs): Use it, refactor as appropriate.
972 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
973 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
975 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
976 emit instructions, return number of instructions which would
978 (aarch64_add_constant): Update call to aarch64_build_constant.
979 (aarch64_output_mi_thunk): Likewise.
980 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
983 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
985 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
986 (TARGET_RTX_COSTS): Call it.
988 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
990 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
991 (cortexa57_vector_cost): Likewise.
992 (cortexa57_tunings): Use them.
994 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
996 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
997 (cpu_addrcost_table): Use it.
998 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
999 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1002 2014-05-16 Richard Biener <rguenther@suse.de>
1004 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1005 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1006 (visit_phi): Ignore edges marked as not executable.
1007 (class cond_dom_walker): New.
1008 (cond_dom_walker::before_dom_children): Value-number
1009 control statements and mark successor edges as not
1010 executable if possible.
1011 (run_scc_vn): First walk all control statements in
1012 dominator order, marking edges as not executable.
1013 * tree-inline.c (copy_edges_for_bb): Be not confused
1014 about random edge flags.
1016 2014-05-16 Richard Biener <rguenther@suse.de>
1018 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1020 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1023 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1024 (__TM_simple_begin): Use it.
1025 (__TM_begin): Likewise.
1027 2014-05-15 Martin Jambor <mjambor@suse.cz>
1030 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1031 type_preserved flag when the indirect edge is polymorphic.
1033 2014-05-15 Martin Jambor <mjambor@suse.cz>
1035 PR tree-optimization/61090
1036 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1037 build_ref_for_model.
1039 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1041 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1042 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1044 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1046 PR tree-optimization/61158
1047 * fold-const.c (fold_binary_loc): If X is zero-extended and
1048 shiftc >= prec, make sure zerobits is all ones instead of
1049 invoking undefined behavior.
1051 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1053 * regcprop.h: New file.
1054 * regcprop.c (skip_debug_insn_p): New decl.
1055 (replace_oldest_value_reg): Check skip_debug_insn_p.
1056 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1057 * shrink-wrap.c: Include regcprop.h.
1058 (prepare_shrink_wrap): Call
1059 copyprop_hardreg_forward_bb_without_debug_insn.
1061 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1063 * shrink-wrap.h: Update comment.
1064 * shrink-wrap.c: Update comment.
1065 (next_block_for_reg): Rename to live_edge_for_reg.
1066 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1067 (move_insn_for_shrink_wrap): Split live_edge.
1068 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1070 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1072 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1074 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1075 * config/sparc/sparc.md (fptype_ut699): New attribute.
1076 (in_branch_delay): Return false if -mfix-ut699 is specified and
1077 fptype_ut699 is set to single.
1078 (truncdfsf2): Add fptype_ut699 attribute.
1079 (fix_truncdfsi2): Likewise.
1080 (floatsisf2): Change fptype attribute.
1081 (fix_truncsfsi2): Likewise.
1082 (negtf2_notv9): Delete.
1083 (negtf2_v9): Likewise.
1084 (negtf2_hq): New instruction.
1085 (negtf2): New instruction and splitter.
1086 (negdf2_notv9): Rewrite.
1087 (abstf2_notv9): Delete.
1088 (abstf2_hq_v9): Likewise.
1089 (abstf2_v9): Likewise.
1090 (abstf2_hq): New instruction.
1091 (abstf2): New instruction and splitter.
1092 (absdf2_notv9): Rewrite.
1094 2014-05-14 Cary Coutant <ccoutant@google.com>
1097 * opts.c (common_handle_option): Don't special-case "-g".
1098 (set_debug_level): Default to at least level 2 with "-g".
1100 2014-05-14 DJ Delorie <dj@redhat.com>
1102 * config/msp430/msp430.c (msp430_builtin): Add
1103 MSP430_BUILTIN_DELAY_CYCLES.
1104 (msp430_init_builtins): Register void __delay_cycles(long long).
1105 (msp430_builtin_decl): Add it.
1106 (cg_magic_constant): New.
1107 (msp430_expand_delay_cycles): New.
1108 (msp430_expand_builtin): Call it.
1109 (msp430_print_operand_raw): Change integer printing from "int" to
1111 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1112 (delay_cycles_start): New.
1113 (delay_cycles_end): New.
1114 (delay_cycles_32): New.
1115 (delay_cycles_32x): New.
1116 (delay_cycles_16): New.
1117 (delay_cycles_16x): New.
1118 (delay_cycles_2): New.
1119 (delay_cycles_1): New.
1120 * doc/extend.texi: Document __delay_cycles().
1122 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1124 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1125 length attribute computation.
1127 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1130 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1132 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1135 * config/sparc/sparc.md: Fix types of low and high in DI constant
1136 splitter. Use gen_int_mode in some other splitters.
1138 2014-05-14 Martin Jambor <mjambor@suse.cz>
1141 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1143 2014-05-14 James Norris <jnorris@codesourcery.com>
1145 * omp-low.c (expand_parallel_call): Remove shadow variable.
1146 (expand_omp_taskreg): Likewise.
1148 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1150 * common/config/i386/i386-common.c
1151 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1152 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1153 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1154 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1155 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1156 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1157 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1158 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1159 xsavecintrin.h, xsavesintrin.h.
1160 (x86_64-*-*): Ditto.
1161 * config/i386/clflushoptintrin.h: New.
1162 * config/i386/xsavecintrin.h: Ditto.
1163 * config/i386/xsavesintrin.h: Ditto.
1164 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1165 (bit_XSAVES): Ditto.
1166 (bit_XSAVES): Ditto.
1167 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1168 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1170 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1171 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1172 OPTION_MASK_ISA_XSAVES.
1173 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1175 (PTA_CLFLUSHOPT) Define.
1176 (PTA_XSAVEC): Ditto.
1177 (PTA_XSAVES): Ditto.
1178 (ix86_option_override_internal): Handle new options.
1179 (ix86_valid_target_attribute_inner_p): Ditto.
1180 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1181 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1182 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1183 (bdesc_special_args): Add __builtin_ia32_xsaves,
1184 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1185 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1186 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1187 (ix86_expand_builtin): Handle new builtins.
1188 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1189 (TARGET_CLFLUSHOPT_P): Ditto.
1190 (TARGET_XSAVEC): Ditto.
1191 (TARGET_XSAVEC_P): Ditto.
1192 (TARGET_XSAVES): Ditto.
1193 (TARGET_XSAVES_P): Ditto.
1194 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1195 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1196 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1198 (ANY_XRSTOR64): Ditto.
1200 (xrstor): Change into <xrstor>.
1201 (xrstor_rex64): Change into <xrstor>_rex64.
1202 (xrstor64): Change into <xrstor>64
1204 * config/i386/i386.opt (mclflushopt): New.
1207 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1209 * doc/invoke.texi: Document new options.
1211 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1213 PR rtl-optimization/60866
1214 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1215 Default it to -1. Pass it down to init_simplejump_data.
1216 (init_simplejump_data): New parameter old_seqno. Pass it down
1217 to get_seqno_for_a_jump.
1218 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1219 initializing new jump seqno as a last resort. Add comment.
1220 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1221 jump and pass it down to sel_init_new_insn.
1222 (sel_redirect_edge_and_branch_force): Likewise.
1224 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1226 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1227 shifted values to avoid build warning.
1229 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1231 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1232 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1233 (cfg_layout_merge_blocks): Likewise.
1234 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1236 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1238 PR rtl-optimization/60901
1239 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1240 bb predecessor belongs to the same scheduling region. Adjust comment.
1242 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1244 * doc/sourcebuild.texi: (dfp_hw): Document.
1245 (p8vector_hw): Likewise.
1246 (powerpc_eabi_ok): Likewise.
1247 (powerpc_elfv2): Likewise.
1248 (powerpc_htm_ok): Likewise.
1249 (ppc_recip_hw): Likewise.
1252 2014-05-13 Cary Coutant <ccoutant@google.com>
1254 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1256 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1258 * gengtype-parse.c (require3): Eliminate in favor of...
1260 (require_template_declaration): Update to support optional single *
1263 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1264 (create_user_defined_type): Handle a single level of explicit
1265 pointerness within template arguments.
1266 (struct write_types_data): Add field "kind".
1267 (filter_type_name): Handle "*" character.
1268 (write_user_func_for_structure_ptr): Require a write_types_data
1269 rather than just a prefix string, so that we can look up the kind
1270 of the wtd and use it as an index into wrote_user_func_for_ptr,
1271 ensuring that such functions are written at most once. Support
1272 subclasses by invoking the marking function of the ultimate base class.
1273 (write_user_func_for_structure_body): Require a write_types_data
1274 rather than just a prefix string, so that we can pass this to
1275 write_user_func_for_structure_ptr.
1276 (write_func_for_structure): Likewise.
1277 (ggc_wtd): Add initializer of new "kind" field.
1278 (pch_wtd): Likewise.
1280 * gengtype.h (enum write_types_kinds): New.
1281 (struct type): Add field wrote_user_func_for_ptr to the "s"
1284 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1286 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1287 instead of const_binop.
1288 (fold_binary_loc): Likewise.
1290 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1292 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1293 calculation to match get_ref_base_and_extent.
1295 2014-05-13 Catherine Moore <clm@codesourcery.com>
1296 Sandra Loosemore <sandra@codesourcery.com>
1298 * configure.ac: Fix assembly for explicit JALR relocation check.
1299 * configure: Regenerate.
1301 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1303 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1304 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1305 Remove associated type declarations and initialisations.
1306 (arm_expand_neon_builtin): Likewise.
1307 (neon_emit_pair_result_insn): Delete.
1308 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1309 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1310 (neon_vzip<mode>): Likewise.
1311 (neon_vuzp<mode>): Likewise.
1313 2014-05-13 Richard Biener <rguenther@suse.de>
1316 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1317 it needs revisiting whether the call still may be tail-called.
1319 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1321 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1322 * rtl.h (block_symbol): Reduce number of fields to 2.
1323 (rtx_def): Add u2.symbol_ref_flags.
1324 (SYMBOL_REF_FLAGS): Use it.
1325 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1326 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1327 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1328 Lower index of SYMBOL_REF_DATA.
1329 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1330 Print SYMBOL_REF_FLAGS at the same time.
1331 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1333 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1335 * rtl.def (VAR_LOCATION): Remove "i" field.
1336 * rtl.h (rtx_def): Add u2.var_location_status.
1337 (PAT_VAR_LOCATION_STATUS): Use it.
1338 (gen_rtx_VAR_LOCATION): Declare.
1339 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1340 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1341 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1343 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1345 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1346 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1348 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1350 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1351 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1352 * rtl.h (rtx_def): Add insn_uid to u2 field.
1353 (RTX_FLAG_CHECK8): Delete in favor of...
1354 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1355 (INSN_DELETED_P): Update accordingly.
1356 (INSN_UID): Use u2.insn_uid.
1357 (INSN_CHAIN_CODE_P): Define.
1358 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1359 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1360 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1361 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1362 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1363 indices accordingly.
1364 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1365 Update indices for insn-chain rtxes.
1366 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1367 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1368 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1369 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1370 * combine.c (try_combine): Likewise.
1371 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1373 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1375 * rtl.def (REG): Remove middle field.
1376 * rtl.h (rtx_def): Add orignal_regno to u2.
1377 (ORIGINAL_REGNO): Use it instead of field 1.
1378 (REG_ATTRS): Lower field index accordingly.
1379 * gengtype.c (adjust_field_rtx_def): Remove handling of
1380 ORIGINAL_REGNO. Move REG_ATTRS index down.
1381 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1382 code that prints the REGNO.
1384 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1386 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1389 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1391 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1393 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1395 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1396 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1398 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1400 * config/aarch64/aarch64-protos.h
1401 (aarch64_hard_regno_caller_save_mode): New prototype.
1402 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1404 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1406 2014-05-13 Christian Bruel <christian.bruel@st.com>
1408 * target.def (mode_switching): New hook vector.
1409 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1410 (mode_exit, modepriority_to_mode): Likewise.
1411 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1412 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1413 * target.h: Include tm.h and hard-reg-set.h.
1414 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1415 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1416 * doc/tm.texi Regenerate.
1417 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1418 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1419 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1420 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1421 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1422 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1423 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1424 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1425 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1426 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1427 (ix86_emit_mode_set): Hookify.
1428 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1430 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1431 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1432 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1433 (epiphany_mode_priority_to_mode): Remove declaration.
1434 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1435 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1436 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1438 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1439 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1440 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1442 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1445 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1446 is const0_rtx, return immediately. Don't test count == 0 when
1449 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1451 * Makefile.in: add shrink-wrap.o.
1452 * config/i386/i386.c: include "shrink-wrap.h"
1453 * function.c: Likewise.
1454 (requires_stack_frame_p, next_block_for_reg,
1455 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1456 dup_block_and_redirect): Move to shrink-wrap.c
1457 (thread_prologue_and_epilogue_insns): Extract three code segments
1458 as functions in shrink-wrap.c
1459 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1461 * shrink-wrap.c: New file.
1462 * shrink-wrap.h: New file.
1464 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1466 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1467 reference to Solaris.
1469 2014-05-12 Mike Stump <mikestump@comcast.net>
1472 * genattrtab.c (filename): Add.
1473 (convert_set_attr_alternative): Improve error message.
1474 (check_defs): Restore read_md_filename for error messages.
1475 (gen_insn): Save filename.
1477 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1479 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1480 -fno-local-ivars and -fivar-visibility.
1481 * c-family/c.opt: Make -Wshadow also implicitly enable
1484 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1486 * doc/tm.texi: Remove reference to deleted macro.
1487 * doc/tm.texi.in: Likewise.
1489 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1492 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1495 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1497 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1498 * config/arm/aout.h (License): Same.
1499 * config/arm/bpabi.h (License): Same.
1500 * config/arm/elf.h (License): Same.
1501 * config/arm/linux-elf.h (License): Same.
1502 * config/arm/linux-gas.h (License): Same.
1503 * config/arm/netbsd-elf.h (License): Same.
1504 * config/arm/uclinux-eabi.h (License): Same.
1505 * config/arm/uclinux-elf.h (License): Same.
1506 * config/arm/vxworks.h (License): Same.
1508 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1510 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1511 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1512 number of operands to 3.
1513 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1514 * tree-nested.c (convert_nonlocal_omp_clauses,
1515 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1516 * gimplify.c (gimplify_scan_omp_clauses): Handle
1517 OMP_CLAUSE_LINEAR_STMT.
1518 * omp-low.c (lower_rec_input_clauses): Fix typo.
1519 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1520 cast between Fortran boolean_type_node and C _Bool if
1523 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1525 PR tree-optimization/61136
1526 * wide-int.h (multiple_of_p): Define a version that doesn't return
1528 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1529 integer_zerop/const_binop pair.
1530 (multiple_of_p): Likewise, converting both operands to widest_int
1533 2014-05-09 Teresa Johnson <tejohnson@google.com>
1535 * cgraphunit.c (analyze_functions): Use correct dump file.
1537 2014-05-09 Florian Weimer <fweimer@redhat.com>
1539 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1541 (stack_protect_return_slot_p): New function.
1542 (expand_used_vars): Call stack_protect_decl_p and
1543 stack_protect_return_slot_p for -fstack-protector-strong.
1545 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1548 * fold-const.c (fold_binary_loc): Changed width of mask.
1550 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1552 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1553 unsigned int initializers for regno_in, regno_out.
1555 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1558 * config/avr/avr.md (cc): Add new attribute set_vzn.
1559 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1560 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1561 with INC, DEC or NEG.
1562 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1563 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1564 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1566 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1569 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1571 * wide-int.cc (UTItype): Define.
1572 (UDWtype): Define for appropriate W_TYPE_SIZE.
1574 2014-05-09 Richard Biener <rguenther@suse.de>
1576 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1577 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1578 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1579 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1580 ssa_propagate): Adjust.
1582 2014-05-08 Jeff Law <law@redhat.com>
1584 PR tree-optimization/61009
1585 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1586 tri-state rather than a boolean. When a block is too big to
1587 thread through, inform caller via negative return value.
1588 (thread_across_edge): If a block was too big for normal threading,
1589 then it's too big for a joiner too, so remove temporary equivalences
1590 and return immediately.
1592 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1593 Matthias Klose <doko@ubuntu.com>
1596 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1598 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1601 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1603 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1606 * config/alpha/alpha.c: Include gimple-iterator.h.
1607 (alpha_gimple_fold_builtin): New function. Move
1608 ALPHA_BUILTIN_UMULH folding from ...
1609 (alpha_fold_builtin): ... here.
1610 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1612 2014-05-08 Wei Mi <wmi@google.com>
1615 * config/i386/i386.c (ix86_compute_frame_layout): Update
1616 preferred_stack_boundary for call, expanded from tls descriptor.
1617 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1618 to depend on SP register.
1619 (*tls_local_dynamic_base_32_gnu): Ditto.
1620 (*tls_local_dynamic_32_once): Ditto.
1621 (tls_global_dynamic_64_<mode>): Set
1622 ix86_tls_descriptor_calls_expanded_in_cfun.
1623 (tls_local_dynamic_base_64_<mode>): Ditto.
1624 (tls_global_dynamic_32): Set
1625 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1626 to depend on SP register.
1627 (tls_local_dynamic_base_32): Ditto.
1629 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1631 * config/arm/arm_neon.h: Update comment.
1632 * config/arm/neon-docgen.ml: Delete.
1633 * config/arm/neon-gen.ml: Delete.
1634 * doc/arm-neon-intrinsics.texi: Update comment.
1636 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1638 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1640 (vand, vorr, veor, vorn, vbic): Remove.
1641 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1643 (neon_vsub_unspec): Likewise.
1644 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1646 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1648 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1649 (vadd_s16): Likewise.
1650 (vadd_s32): Likewise.
1651 (vadd_f32): Likewise.
1652 (vadd_u8): Likewise.
1653 (vadd_u16): Likewise.
1654 (vadd_u32): Likewise.
1655 (vadd_s64): Likewise.
1656 (vadd_u64): Likewise.
1657 (vaddq_s8): Likewise.
1658 (vaddq_s16): Likewise.
1659 (vaddq_s32): Likewise.
1660 (vaddq_s64): Likewise.
1661 (vaddq_f32): Likewise.
1662 (vaddq_u8): Likewise.
1663 (vaddq_u16): Likewise.
1664 (vaddq_u32): Likewise.
1665 (vaddq_u64): Likewise.
1666 (vmul_s8): Likewise.
1667 (vmul_s16): Likewise.
1668 (vmul_s32): Likewise.
1669 (vmul_f32): Likewise.
1670 (vmul_u8): Likewise.
1671 (vmul_u16): Likewise.
1672 (vmul_u32): Likewise.
1673 (vmul_p8): Likewise.
1674 (vmulq_s8): Likewise.
1675 (vmulq_s16): Likewise.
1676 (vmulq_s32): Likewise.
1677 (vmulq_f32): Likewise.
1678 (vmulq_u8): Likewise.
1679 (vmulq_u16): Likewise.
1680 (vmulq_u32): Likewise.
1681 (vsub_s8): Likewise.
1682 (vsub_s16): Likewise.
1683 (vsub_s32): Likewise.
1684 (vsub_f32): Likewise.
1685 (vsub_u8): Likewise.
1686 (vsub_u16): Likewise.
1687 (vsub_u32): Likewise.
1688 (vsub_s64): Likewise.
1689 (vsub_u64): Likewise.
1690 (vsubq_s8): Likewise.
1691 (vsubq_s16): Likewise.
1692 (vsubq_s32): Likewise.
1693 (vsubq_s64): Likewise.
1694 (vsubq_f32): Likewise.
1695 (vsubq_u8): Likewise.
1696 (vsubq_u16): Likewise.
1697 (vsubq_u32): Likewise.
1698 (vsubq_u64): Likewise.
1699 (vand_s8): Likewise.
1700 (vand_s16): Likewise.
1701 (vand_s32): Likewise.
1702 (vand_u8): Likewise.
1703 (vand_u16): Likewise.
1704 (vand_u32): Likewise.
1705 (vand_s64): Likewise.
1706 (vand_u64): Likewise.
1707 (vandq_s8): Likewise.
1708 (vandq_s16): Likewise.
1709 (vandq_s32): Likewise.
1710 (vandq_s64): Likewise.
1711 (vandq_u8): Likewise.
1712 (vandq_u16): Likewise.
1713 (vandq_u32): Likewise.
1714 (vandq_u64): Likewise.
1715 (vorr_s8): Likewise.
1716 (vorr_s16): Likewise.
1717 (vorr_s32): Likewise.
1718 (vorr_u8): Likewise.
1719 (vorr_u16): Likewise.
1720 (vorr_u32): Likewise.
1721 (vorr_s64): Likewise.
1722 (vorr_u64): Likewise.
1723 (vorrq_s8): Likewise.
1724 (vorrq_s16): Likewise.
1725 (vorrq_s32): Likewise.
1726 (vorrq_s64): Likewise.
1727 (vorrq_u8): Likewise.
1728 (vorrq_u16): Likewise.
1729 (vorrq_u32): Likewise.
1730 (vorrq_u64): Likewise.
1731 (veor_s8): Likewise.
1732 (veor_s16): Likewise.
1733 (veor_s32): Likewise.
1734 (veor_u8): Likewise.
1735 (veor_u16): Likewise.
1736 (veor_u32): Likewise.
1737 (veor_s64): Likewise.
1738 (veor_u64): Likewise.
1739 (veorq_s8): Likewise.
1740 (veorq_s16): Likewise.
1741 (veorq_s32): Likewise.
1742 (veorq_s64): Likewise.
1743 (veorq_u8): Likewise.
1744 (veorq_u16): Likewise.
1745 (veorq_u32): Likewise.
1746 (veorq_u64): Likewise.
1747 (vbic_s8): Likewise.
1748 (vbic_s16): Likewise.
1749 (vbic_s32): Likewise.
1750 (vbic_u8): Likewise.
1751 (vbic_u16): Likewise.
1752 (vbic_u32): Likewise.
1753 (vbic_s64): Likewise.
1754 (vbic_u64): Likewise.
1755 (vbicq_s8): Likewise.
1756 (vbicq_s16): Likewise.
1757 (vbicq_s32): Likewise.
1758 (vbicq_s64): Likewise.
1759 (vbicq_u8): Likewise.
1760 (vbicq_u16): Likewise.
1761 (vbicq_u32): Likewise.
1762 (vbicq_u64): Likewise.
1763 (vorn_s8): Likewise.
1764 (vorn_s16): Likewise.
1765 (vorn_s32): Likewise.
1766 (vorn_u8): Likewise.
1767 (vorn_u16): Likewise.
1768 (vorn_u32): Likewise.
1769 (vorn_s64): Likewise.
1770 (vorn_u64): Likewise.
1771 (vornq_s8): Likewise.
1772 (vornq_s16): Likewise.
1773 (vornq_s32): Likewise.
1774 (vornq_s64): Likewise.
1775 (vornq_u8): Likewise.
1776 (vornq_u16): Likewise.
1777 (vornq_u32): Likewise.
1778 (vornq_u64): Likewise.
1780 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1782 * wide-int.cc (UTItype): Define.
1783 (UDWtype): Define for appropriate W_TYPE_SIZE.
1785 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1787 PR tree-optimization/59100
1788 * tree-ssa-phiopt.c: Include tree-inline.h.
1789 (neutral_element_p, absorbing_element_p): New functions.
1790 (value_replacement): Handle conditional binary operations with a
1791 neutral or absorbing element.
1793 2014-05-08 Richard Biener <rguenther@suse.de>
1795 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1796 folding the expression.
1797 (valueize_expr): Remove.
1798 (visit_reference_op_load): Do not valueize the result of
1800 (simplify_binary_expression): Likewise.
1801 (simplify_unary_expression): Likewise.
1803 2014-05-08 Richard Biener <rguenther@suse.de>
1805 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1806 looking at TYPE_ARG_TYPES.
1808 2014-05-08 Richard Biener <rguenther@suse.de>
1810 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1811 pointer propagation special-case.
1813 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1815 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1816 core part of address expressions.
1818 2014-05-08 Alan Modra <amodra@gmail.com>
1821 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1822 loads and stores when -mno-strict-align at any alignment.
1823 (expand_block_clear): Similarly. Also correct calculation of
1826 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1829 * tree-complex.c (expand_complex_move): Keep line info when expanding
1831 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1832 of complex expression. Use new argument to display correct location
1833 for values coming from phi statement.
1834 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1835 (warn_uninitialized_phi): Pass location of phi argument to
1837 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1838 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1840 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1842 * config/rs6000/predicates.md (indexed_address_mem): New.
1843 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1844 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1845 fpstore_ux, fpstore_u.
1846 (sign_extend, indexed, update): New.
1847 (cell_micro): Adjust.
1848 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1849 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1850 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1851 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1852 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1853 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1854 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1855 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1856 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1857 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1858 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1859 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1860 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1861 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1862 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1864 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1865 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1866 *vsx_extract_<mode>_store): Adjust.
1867 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1868 is_cracked_insn, insn_must_be_first_in_group,
1869 insn_must_be_last_in_group): Adjust.
1871 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1873 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1874 ppc440-fpstore): Adjust.
1875 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1876 ppc476-fpstore): Adjust.
1877 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1878 ppc601-fpstore): Adjust.
1879 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1881 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1883 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1884 ppc7450-fpstore): Adjust.
1885 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1886 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1887 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1889 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1890 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1891 cell-fpstore, cell-fpstore-update): Adjust.
1892 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1893 ppce300c3_store, ppce300c3_fpstore): Adjust.
1894 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1895 e500mc_fpstore): Adjust.
1896 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1897 e500mc64_store, e500mc64_fpstore): Adjust.
1898 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1899 e5500_fpstore): Adjust.
1900 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1901 e6500_fpstore): Adjust.
1902 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1904 * config/rs6000/power4.md (power4-load, power4-load-ext,
1905 power4-load-ext-update, power4-load-ext-update-indexed,
1906 power4-load-update-indexed, power4-load-update, power4-fpload,
1907 power4-fpload-update, power4-store, power4-store-update,
1908 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1910 * config/rs6000/power5.md (power5-load, power5-load-ext,
1911 power5-load-ext-update, power5-load-ext-update-indexed,
1912 power5-load-update-indexed, power5-load-update, power5-fpload,
1913 power5-fpload-update, power5-store, power5-store-update,
1914 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1916 * config/rs6000/power6.md (power6-load, power6-load-ext,
1917 power6-load-update, power6-load-update-indexed,
1918 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1919 power6-fpload-update, power6-store, power6-store-update,
1920 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1922 * config/rs6000/power7.md (power7-load, power7-load-ext,
1923 power7-load-update, power7-load-update-indexed,
1924 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1925 power7-fpload-update, power7-store, power7-store-update,
1926 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1928 * config/rs6000/power8.md (power8-load, power8-load-update,
1929 power8-load-ext, power8-load-ext-update, power8-fpload,
1930 power8-fpload-update, power8-store, power8-store-update-indexed,
1931 power8-fpstore, power8-fpstore-update): Adjust.
1932 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1934 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1935 titan_lsu_store, titan_lsu_fpstore): Adjust.
1936 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1938 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1941 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1942 unrolled byte insns. Emit address increments after move insns.
1944 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1946 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1947 const_gimple, rather than a gimple.
1948 (gimple_call_builtin_p): Likewise, for the three variants.
1950 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1951 (gimple_call_builtin_p): Likewise, for the three variants.
1953 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1955 PR tree-optimization/61095
1956 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1958 2014-05-07 Richard Biener <rguenther@suse.de>
1960 PR tree-optimization/61034
1961 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1962 (maybe_skip_until): Use translate to take into account
1963 lattices when trying to do disambiguations.
1964 (get_continuation_for_phi_1): Likewise.
1965 (get_continuation_for_phi): Adjust for added translate arguments.
1966 (walk_non_aliased_vuses): Likewise.
1967 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1968 (walk_non_aliased_vuses): Likewise.
1969 (call_may_clobber_ref_p_1): Declare.
1970 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1971 calls. Stop early if we are only supposed to disambiguate.
1972 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1974 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1976 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1977 Emit an error when the function has arguments.
1979 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1981 * cfgloop.h (unswitch_loops): Remove.
1982 * doc/passes.texi: Remove references to loop-unswitch.c
1983 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1985 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1987 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1988 check for loads group of length 3.
1989 (vect_permute_load_chain): New permutations for loads group of
1991 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1992 of vec_perm_shuffle for the new permutations.
1994 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1996 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1997 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1998 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1999 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2000 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2001 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2002 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2003 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2005 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2007 * loop-unswitch.c: Delete.
2009 2014-05-07 Richard Biener <rguenther@suse.de>
2011 * config.gcc: Always set need_64bit_hwint to yes.
2013 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2015 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2016 of using optimize_size.
2018 2014-05-06 Mike Stump <mikestump@comcast.net>
2020 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2022 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2024 * config/i386/sse.md (*mov<mode>_internal)
2025 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2026 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2027 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2028 (*<code><mode>3, *andnot<mode>3<mask_name>)
2029 (<mask_codefor><code><mode>3<mask_name>): Only consider
2030 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2032 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2035 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2037 * lra-constraints.c (valid_address_p): Move earlier in file.
2038 Add a constraint argument to the address_info version.
2039 (satisfies_memory_constraint_p): New function.
2040 (satisfies_address_constraint_p): Likewise.
2041 (process_alt_operands, curr_insn_transform): Use them.
2042 (process_address): Pass the constraint to valid_address_p when
2043 checking address operands.
2045 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2047 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2048 to their respective blocks. Fix inadvertent use of "node".
2050 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2052 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2054 (init_emit_once): ...here.
2055 * rtl.h (init_derived_machine_modes): Declare.
2056 * toplev.c (do_compile): Call it even if no_backend.
2058 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2059 Mike Stump <mikestump@comcast.net>
2060 Richard Sandiford <rdsandiford@googlemail.com>
2061 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2063 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2064 (rtx_equal_for_memref_p): Update comment.
2065 (adjust_offset_for_component_ref): Use wide-int interfaces.
2066 * builtins.c (get_object_alignment_2): Likewise.
2067 (c_readstr): Likewise.
2068 (target_char_cast): Add comment.
2069 (determine_block_size): Use wide-int interfaces.
2070 (expand_builtin_signbit): Likewise.
2071 (fold_builtin_int_roundingfn): Likewise.
2072 (fold_builtin_bitop): Likewise.
2073 (fold_builtin_bswap): Likewise.
2074 (fold_builtin_logarithm): Use signop.
2075 (fold_builtin_pow): Likewise.
2076 (fold_builtin_memory_op): Use wide-int interfaces.
2077 (fold_builtin_object_size): Likewise.
2078 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2079 nb_iterations_estimate.
2080 (record_niter_bound): Use wide-int interfaces.
2081 (get_estimated_loop_iterations_int): Likewise.
2082 (get_estimated_loop_iterations): Likewise.
2083 (get_max_loop_iterations): Likewise.
2084 * cfgloop.h: Include wide-int.h.
2085 (struct nb_iter_bound): Change bound to widest_int.
2086 (struct loop): Change nb_iterations_upper_bound and
2087 nb_iterations_estimate to widest_int.
2088 (record_niter_bound): Switch to use widest_int.
2089 (get_estimated_loop_iterations): Likewise.
2090 (get_max_loop_iterations): Likewise.
2091 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2092 update for wide-int.
2093 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2094 * combine.c (try_combine): Likewise.
2095 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2096 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2098 (aarch64_float_const_representable_p): Likewise.
2099 * config/arc/arc.c: Include wide-int.h.
2100 (arc_can_use_doloop_p): Use wide-int interfaces.
2101 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2102 (vfp3_const_double_index): Likewise.
2103 * config/avr/avr.c (avr_out_round): Likewise.
2104 (avr_fold_builtin): Likewise.
2105 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2106 (bfin_can_use_doloop_p): Likewise.
2107 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2108 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2109 * config/i386/i386.c: Include wide-int.h.
2110 (ix86_data_alignment): Use wide-int interfaces.
2111 (ix86_local_alignment): Likewise.
2112 (ix86_emit_swsqrtsf): Update real_from_integer.
2113 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2114 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2115 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2116 (zero_constant): Likewise.
2117 (input_operand): Likewise.
2118 (splat_input_operand): Likewise.
2119 (non_logical_cint_operand): Change const_double to const_wide_int.
2120 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2121 (easy_altivec_constant): Remove comment.
2122 (paired_expand_vector_init): Use CONSTANT_P.
2123 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2124 (rs6000_emit_move): Update checks.
2125 (rs6000_aggregate_candidate): Use wide-int interfaces.
2126 (rs6000_expand_ternop_builtin): Likewise.
2127 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2128 (rs6000_assemble_integer): Likewise.
2129 (rs6000_hash_constant): Likewise.
2130 (output_toc): Likewise.
2131 (rs6000_rtx_costs): Likewise.
2132 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2133 * config/rs6000/rs6000-c.c: Include wide-int.h.
2134 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2135 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2136 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2137 Handle CONST_WIDE_INT.
2138 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2139 Use tree_fits_uhwi_p.
2140 * config/sparc/sparc.c: Include wide-int.h.
2141 (sparc_fold_builtin): Use wide-int interfaces.
2142 * config/vax/vax.c: Include wide-int.h.
2143 (vax_float_literal): Use real_from_integer.
2144 * coretypes.h (struct hwivec_def): New.
2146 (const_hwivec): New.
2147 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2148 (equiv_constant): Handle CONST_WIDE_INT.
2149 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2150 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2151 * dbxout.c (stabstr_U): Use wide-int interfaces.
2152 (dbxout_type): Update to use cst_fits_shwi_p.
2153 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2154 (TARGET_SUPPORTS_WIDE_INT): Add default.
2155 * dfp.c: Include wide-int.h.
2156 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2157 decimal_real_to_integer.
2158 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2159 decimal_real_to_integer.
2160 * doc/generic.texi (Constant expressions): Update for wide_int.
2161 * doc/rtl.texi (const_double): Likewise.
2162 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2163 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2164 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2165 (REAL_VALUE_FROM_INT): Remove.
2166 (TARGET_SUPPORTS_WIDE_INT): New.
2167 * doc/tm.texi: Regenerate.
2168 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2169 * double-int.h: Include wide-int.h.
2170 (struct wi::int_traits): New.
2171 * dwarf2out.c (get_full_len): New.
2172 (dw_val_equal_p): Add case dw_val_class_wide_int.
2173 (size_of_loc_descr): Likewise.
2174 (output_loc_operands): Likewise.
2175 (insert_double): Remove.
2176 (insert_wide_int): New.
2178 (print_die): Add case dw_val_class_wide_int.
2179 (attr_checksum): Likewise.
2180 (attr_checksum_ordered): Likewise.
2181 (same_dw_val_p): Likewise.
2182 (size_of_die): Likewise.
2183 (value_format): Likewise.
2184 (output_die): Likewise.
2185 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2187 (clz_loc_descriptor): Use wide-int interfaces.
2188 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2189 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2190 (round_up_to_align): Use wide-int interfaces.
2191 (field_byte_offset): Likewise.
2192 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2193 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2194 CONST_DOUBLE handling. Use wide-int interfaces.
2195 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2196 (gen_enumeration_type_die): Use add_AT_wide.
2197 (hash_loc_operands): Add case dw_val_class_wide_int.
2198 (compare_loc_operands): Likewise.
2199 * dwarf2out.h: Include wide-int.h.
2200 (wide_int_ptr): New.
2201 (enum dw_val_class): Add dw_val_class_wide_int.
2202 (struct dw_val_struct): Add val_wide.
2203 * emit-rtl.c (const_wide_int_htab): New.
2204 (const_wide_int_htab_hash): New.
2205 (const_wide_int_htab_eq): New.
2206 (lookup_const_wide_int): New.
2207 (const_double_htab_hash): Use wide-int interfaces.
2208 (const_double_htab_eq): Likewise.
2209 (rtx_to_double_int): Conditionally compile for wide-int.
2210 (immed_double_int_const): Rename to immed_wide_int_const and
2211 update for wide-int.
2212 (immed_double_const): Conditionally compile for wide-int.
2213 (init_emit_once): Use wide-int interfaces.
2214 * explow.c (plus_constant): Likewise.
2215 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2216 (lshift_value): Use wide-int interfaces.
2217 (expand_mult): Likewise.
2218 (choose_multiplier): Likewise.
2219 (expand_smod_pow2): Likewise.
2220 (make_tree): Likewise.
2221 * expr.c (convert_modes): Consolidate handling of constants.
2222 Use wide-int interfaces.
2223 (emit_group_load_1): Add note.
2224 (store_expr): Update comment.
2225 (get_inner_reference): Use wide-int interfaces.
2226 (expand_constructor): Update comment.
2227 (expand_expr_real_2): Use wide-int interfaces.
2228 (expand_expr_real_1): Likewise.
2229 (reduce_to_bit_field_precision): Likewise.
2230 (const_vector_from_tree): Likewise.
2231 * final.c: Include wide-int-print.h.
2232 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2233 * fixed-value.c: Include wide-int.h.
2234 (fixed_from_string): Use wide-int interfaces.
2235 (fixed_to_decimal): Likewise.
2236 (fixed_convert_from_real): Likewise.
2237 (real_convert_from_fixed): Likewise.
2238 * fold-const.h (mem_ref_offset): Return an offset_int.
2239 (div_if_zero_remainder): Remove code parameter.
2240 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2241 Use wide-int interfaces.
2242 (may_negate_without_overflow_p): Use wide-int interfaces.
2243 (negate_expr_p): Likewise.
2244 (fold_negate_expr): Likewise.
2245 (int_const_binop_1): Likewise.
2246 (const_binop): Likewise.
2247 (fold_convert_const_int_from_int): Likewise.
2248 (fold_convert_const_int_from_real): Likewise.
2249 (fold_convert_const_int_from_fixed): Likewise.
2250 (fold_convert_const_fixed_from_int): Likewise.
2251 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2252 (sign_bit_p): Use wide-int interfaces.
2253 (make_range_step): Likewise.
2254 (build_range_check): Likewise. Pass an integer of the correct type
2255 instead of using integer_one_node.
2256 (range_predecessor): Pass an integer of the correct type instead
2257 of using integer_one_node.
2258 (range_successor): Likewise.
2259 (merge_ranges): Likewise.
2260 (unextend): Use wide-int interfaces.
2261 (extract_muldiv_1): Likewise.
2262 (fold_div_compare): Likewise.
2263 (fold_single_bit_test): Likewise.
2264 (fold_sign_changed_comparison): Likewise.
2265 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2266 (fold_plusminus_mult_expr): Use wide-int interfaces.
2267 (native_encode_int): Likewise.
2268 (native_interpret_int): Likewise.
2269 (fold_unary_loc): Likewise.
2270 (pointer_may_wrap_p): Likewise.
2271 (size_low_cst): Likewise.
2272 (mask_with_tz): Likewise.
2273 (fold_binary_loc): Likewise.
2274 (fold_ternary_loc): Likewise.
2275 (multiple_of_p): Likewise.
2276 (tree_call_nonnegative_warnv_p): Update calls to
2277 tree_int_cst_min_precision and real_from_integer.
2278 (fold_negate_const): Use wide-int interfaces.
2279 (fold_abs_const): Likewise.
2280 (fold_relational_const): Use tree_int_cst_lt.
2281 (round_up_loc): Use wide-int interfaces.
2282 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2283 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2284 * gengtype.c: Remove include of double-int.h.
2285 (do_typedef): Use wide-int interfaces.
2286 (open_base_files): Add wide-int.h.
2287 (main): Add offset_int and widest_int typedefs.
2288 * gengtype-lex.l: Handle "^".
2289 (CXX_KEYWORD): Add "static".
2290 * gengtype-parse.c (require3): New.
2291 (require_template_declaration): Handle constant template arguments
2292 and nested templates.
2293 * gengtype-state.c: Don't include "double-int.h".
2294 * genpreds.c (write_one_predicate_function): Update comment.
2295 (write_tm_constrs_h): Add check for hval and lval use in
2297 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2298 (add_to_sequence): Likewise.
2299 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2300 and const_double_operand.
2301 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2303 * gimple-fold.c (get_base_constructor): Likewise.
2304 (fold_array_ctor_reference): Likewise.
2305 (fold_nonarray_ctor_reference): Likewise.
2306 (fold_const_aggregate_ref_1): Likewise.
2307 (gimple_val_nonnegative_real_p): Likewise.
2308 (gimple_fold_indirect_ref): Likewise.
2309 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2310 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2311 (struct slsr_cand_d): Change index to be widest_int.
2312 (struct incr_info_d): Change incr to be widest_int.
2313 (alloc_cand_and_find_basis): Use wide-int interfaces.
2314 (slsr_process_phi): Likewise.
2315 (backtrace_base_for_ref): Likewise. Return a widest_int.
2316 (restructure_reference): Take a widest_int instead of a double_int.
2317 (slsr_process_ref): Use wide-int interfaces.
2318 (create_mul_ssa_cand): Likewise.
2319 (create_mul_imm_cand): Likewise.
2320 (create_add_ssa_cand): Likewise.
2321 (create_add_imm_cand): Take a widest_int instead of a double_int.
2322 (slsr_process_add): Use wide-int interfaces.
2323 (slsr_process_cast): Likewise.
2324 (slsr_process_copy): Likewise.
2325 (dump_candidate): Likewise.
2326 (dump_incr_vec): Likewise.
2327 (replace_ref): Likewise.
2328 (cand_increment): Likewise. Return a widest_int.
2329 (cand_abs_increment): Likewise.
2330 (replace_mult_candidate): Take a widest_int instead of a double_int.
2331 (replace_unconditional_candidate): Use wide-int interfaces.
2332 (incr_vec_index): Take a widest_int instead of a double_int.
2333 (create_add_on_incoming_edge): Likewise.
2334 (create_phi_basis): Use wide-int interfaces.
2335 (replace_conditional_candidate): Likewise.
2336 (record_increment): Take a widest_int instead of a double_int.
2337 (record_phi_increments): Use wide-int interfaces.
2338 (phi_incr_cost): Take a widest_int instead of a double_int.
2339 (lowest_cost_path): Likewise.
2340 (total_savings): Likewise.
2341 (analyze_increments): Use wide-int interfaces.
2342 (ncd_with_phi): Take a widest_int instead of a double_int.
2343 (ncd_of_cand_and_phis): Likewise.
2344 (nearest_common_dominator_for_cands): Likewise.
2345 (insert_initializers): Use wide-int interfaces.
2346 (all_phi_incrs_profitable): Likewise.
2347 (replace_one_candidate): Likewise.
2348 (replace_profitable_candidates): Likewise.
2349 * godump.c: Include wide-int-print.h.
2350 (go_output_typedef): Use wide-int interfaces.
2351 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2352 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2353 (build_loop_iteration_domains): Likewise.
2354 * hooks.h: Include wide-int.h rather than double-int.h.
2355 (hook_bool_dint_dint_uint_bool_true): Delete.
2356 (hook_bool_wint_wint_uint_bool_true): Declare.
2357 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2358 (hook_bool_wint_wint_uint_bool_true): New.
2359 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2361 (ubsan_expand_si_overflow_mul_check): Likewise.
2362 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2363 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2364 (get_ancestor_addr_info): Likewise.
2365 (ipa_modify_call_arguments): Likewise.
2366 * loop-doloop.c (doloop_modify): Likewise.
2367 (doloop_optimize): Likewise.
2368 * loop-iv.c (iv_number_of_iterations): Likewise.
2369 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2370 (unroll_loop_constant_iterations): Likewise.
2371 (decide_unroll_runtime_iterations): Likewise.
2372 (unroll_loop_runtime_iterations): Likewise.
2373 (decide_peel_simple): Likewise.
2374 (decide_unroll_stupid): Likewise.
2375 * lto-streamer-in.c (streamer_read_wi): Add.
2376 (input_cfg): Use wide-int interfaces.
2377 (lto_input_tree_1): Likewise.
2378 * lto-streamer-out.c (streamer_write_wi): Add.
2379 (hash_tree): Use wide-int interfaces.
2380 (output_cfg): Likewise.
2381 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2382 (GTFILES): Add wide-int.h and signop.h.
2383 (TAGS): Look for .cc files too.
2384 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2385 * optabs.c (expand_subword_shift): Likewise.
2386 (expand_doubleword_shift): Likewise.
2387 (expand_absneg_bit): Likewise.
2388 (expand_copysign_absneg): Likewise.
2389 (expand_copysign_bit): Likewise.
2390 * postreload.c (reload_cse_simplify_set): Likewise.
2391 * predict.c (predict_iv_comparison): Likewise.
2392 * pretty-print.h: Include wide-int-print.h.
2394 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2395 * print-tree.c: Include wide-int-print.h.
2396 (print_node_brief): Use wide-int interfaces.
2397 (print_node): Likewise.
2398 * read-rtl.c (validate_const_wide_int): New.
2399 (read_rtx_code): Add CONST_WIDE_INT case.
2400 * real.c: Include wide-int.h.
2401 (real_to_integer2): Delete.
2402 (real_to_integer): New function, returning a wide_int.
2403 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2404 (ten_to_ptwo): Update call to real_from_integer.
2405 (real_digit): Likewise.
2406 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2407 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2408 (REAL_VALUE_TO_INT): Delete.
2409 (real_to_integer): Declare a wide-int form.
2410 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2411 * recog.c (const_int_operand): Improve comment.
2412 (const_scalar_int_operand): New.
2413 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2414 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2415 (split_double): Likewise.
2416 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2417 (rtx_size): Likewise.
2418 (rtx_alloc_stat_v): New.
2419 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2420 (cwi_output_hex): New.
2421 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2422 (cwi_check_failed_bounds): New.
2423 * rtl.def (CONST_WIDE_INT): New.
2424 * rtl.h: Include <utility> and wide-int.h.
2425 (struct hwivec_def): New.
2426 (CWI_GET_NUM_ELEM): New.
2427 (CWI_PUT_NUM_ELEM): New.
2428 (struct rtx_def): Add num_elem and hwiv.
2429 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2430 (CASE_CONST_UNIQUE): Likewise.
2431 (CASE_CONST_ANY): Likewise.
2432 (CONST_SCALAR_INT_P): Likewise.
2433 (CONST_WIDE_INT_P): New.
2435 (HWIVEC_CHECK): New.
2436 (cwi_check_failed_bounds): New.
2438 (HWIVEC_CHECK): New.
2439 (CONST_WIDE_INT_VEC) New.
2440 (CONST_WIDE_INT_NUNITS) New.
2441 (CONST_WIDE_INT_ELT) New.
2442 (rtx_mode_t): New type.
2443 (wi::int_traits <rtx_mode_t>): New.
2445 (wi::min_value): New.
2446 (wi::max_value): New.
2448 (const_wide_int_alloc): New.
2449 (immed_wide_int_const): New.
2450 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2451 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2452 * signop.h: New file.
2453 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2454 (simplify_const_unary_operation): Use wide-int interfaces.
2455 (simplify_binary_operation_1): Likewise.
2456 (simplify_const_binary_operation): Likewise.
2457 (simplify_const_relational_operation): Likewise.
2458 (simplify_immed_subreg): Likewise.
2459 * stmt.c (expand_case): Likewise.
2460 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2461 signop rather than a bool.
2462 * stor-layout.c (layout_type): Use wide-int interfaces.
2463 (initialize_sizetypes): Update calls to
2464 set_min_and_max_values_for_integral_type.
2465 (set_min_and_max_values_for_integral_type): Take a signop rather
2466 than a bool. Use wide-int interfaces.
2467 (fixup_signed_type): Update accordingly. Remove
2468 HOST_BITS_PER_DOUBLE_INT limit.
2469 (fixup_unsigned_type): Likewise.
2470 * system.h (STATIC_CONSTANT_P): New.
2471 (STATIC_ASSERT): New.
2472 * target.def (can_use_doloop_p): Take widest_ints rather than
2474 * target.h: Include wide-int.h rather than double-int.h.
2475 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2477 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2478 rather than INT_CST_LT_UNSIGNED.
2479 (can_use_doloop_if_innermost): Take widest_ints rather than
2481 * tree-affine.c: Include wide-int-print.h.
2482 (double_int_ext_for_comb): Delete.
2483 (wide_int_ext_for_comb): New.
2484 (aff_combination_zero): Use wide-int interfaces.
2485 (aff_combination_const): Take a widest_int instead of a double_int.
2486 (aff_combination_elt): Use wide-int interfaces.
2487 (aff_combination_scale): Take a widest_int instead of a double_int.
2488 (aff_combination_add_elt): Likewise.
2489 (aff_combination_add_cst): Likewise.
2490 (aff_combination_add): Use wide-int interfaces.
2491 (aff_combination_convert): Likewise.
2492 (tree_to_aff_combination): Likewise.
2493 (add_elt_to_tree): Take a widest_int instead of a double_int.
2494 (aff_combination_to_tree): Use wide-int interfaces.
2495 (aff_combination_remove_elt): Likewise.
2496 (aff_combination_add_product): Take a widest_int instead of
2498 (aff_combination_mult): Use wide-int interfaces.
2499 (aff_combination_expand): Likewise.
2500 (double_int_constant_multiple_p): Delete.
2501 (wide_int_constant_multiple_p): New.
2502 (aff_combination_constant_multiple_p): Take a widest_int pointer
2503 instead of a double_int pointer.
2504 (print_aff): Use wide-int interfaces.
2505 (get_inner_reference_aff): Take a widest_int pointer
2506 instead of a double_int pointer.
2507 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2508 * tree-affine.h: Include wide-int.h.
2509 (struct aff_comb_elt): Change type of coef to widest_int.
2510 (struct affine_tree_combination): Change type of offset to widest_int.
2511 (double_int_ext_for_comb): Delete.
2512 (wide_int_ext_for_comb): New.
2513 (aff_combination_const): Use widest_int instead of double_int.
2514 (aff_combination_scale): Likewise.
2515 (aff_combination_add_elt): Likewise.
2516 (aff_combination_constant_multiple_p): Likewise.
2517 (get_inner_reference_aff): Likewise.
2518 (aff_comb_cannot_overlap_p): Likewise.
2519 (aff_combination_zero_p): Use wide-int interfaces.
2520 * tree.c: Include tree.h.
2521 (init_ttree): Use make_int_cst.
2522 (tree_code_size): Removed code for INTEGER_CST case.
2523 (tree_size): Add INTEGER_CST case.
2524 (make_node_stat): Update comment.
2525 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2526 (build_int_cst_type): Use wide-int interfaces.
2527 (double_int_to_tree): Likewise.
2528 (double_int_fits_to_tree_p): Delete.
2529 (force_fit_type_double): Delete.
2530 (force_fit_type): New.
2531 (int_cst_hash_hash): Use wide-int interfaces.
2532 (int_cst_hash_eq): Likewise.
2533 (build_int_cst_wide): Delete.
2534 (wide_int_to_tree): New.
2535 (cache_integer_cst): Use wide-int interfaces.
2536 (build_low_bits_mask): Likewise.
2537 (cst_and_fits_in_hwi): Likewise.
2538 (real_value_from_int_cst): Likewise.
2539 (make_int_cst_stat): New.
2540 (integer_zerop): Use wide_int interfaces.
2541 (integer_onep): Likewise.
2542 (integer_all_onesp): Likewise.
2543 (integer_pow2p): Likewise.
2544 (integer_nonzerop): Likewise.
2545 (tree_log2): Likewise.
2546 (tree_floor_log2): Likewise.
2547 (tree_ctz): Likewise.
2548 (int_size_in_bytes): Likewise.
2549 (mem_ref_offset): Return an offset_int rather than a double_int.
2550 (build_type_attribute_qual_variant): Use wide_int interfaces.
2551 (type_hash_eq): Likewise
2552 (tree_int_cst_equal): Likewise.
2553 (tree_int_cst_lt): Delete.
2554 (tree_int_cst_compare): Likewise.
2555 (tree_fits_shwi_p): Use wide_int interfaces.
2556 (tree_fits_uhwi_p): Likewise.
2557 (tree_int_cst_sign_bit): Likewise.
2558 (tree_int_cst_sgn): Likewise.
2559 (tree_int_cst_min_precision): Take a signop rather than a bool.
2560 (simple_cst_equal): Use wide_int interfaces.
2561 (compare_tree_int): Likewise.
2562 (iterative_hash_expr): Likewise.
2563 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2565 (get_type_static_bounds): Use wide_int interfaces.
2566 (tree_int_cst_elt_check_failed): New.
2567 (build_common_tree_nodes): Reordered to set prec before filling in
2569 (int_cst_value): Check cst_and_fits_in_hwi.
2570 (widest_int_cst_value): Use wide_int interfaces.
2571 (upper_bound_in_type): Likewise.
2572 (lower_bound_in_type): Likewise.
2573 (num_ending_zeros): Likewise.
2574 (drop_tree_overflow): Likewise.
2575 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2576 (gen_conditions_for_pow_cst_base): Likewise.
2577 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2578 (group_case_labels_stmt): Use wide-int interfaces.
2579 (verify_gimple_assign_binary): Likewise.
2580 (print_loop): Likewise.
2581 * tree-chrec.c (tree_fold_binomial): Likewise.
2582 * tree-core.h (struct tree_base): Add int_length.
2583 (struct tree_int_cst): Change rep of value.
2584 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2585 (dr_may_alias_p): Likewise.
2586 (max_stmt_executions_tree): Likewise.
2587 * tree.def (INTEGER_CST): Update comment.
2588 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2589 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2590 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2591 (dequeue_and_dump): Use wide-int interfaces.
2592 * tree.h: Include wide-int.h.
2593 (NULL_TREE): Moved to earlier loc in file.
2594 (TREE_INT_CST_ELT_CHECK): New.
2595 (tree_int_cst_elt_check_failed): New.
2597 (TREE_INT_CST): Delete.
2598 (TREE_INT_CST_LOW): Use wide-int interfaces.
2599 (TREE_INT_CST_HIGH): Delete.
2600 (TREE_INT_CST_NUNITS): New.
2601 (TREE_INT_CST_EXT_NUNITS): Likewise.
2602 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2603 (TREE_INT_CST_ELT): Likewise.
2604 (INT_CST_LT): Delete.
2605 (tree_int_cst_elt_check): New (two forms).
2606 (type_code_size): Update comment.
2607 (make_int_cst_stat, make_int_cst): New.
2608 (tree_to_double_int): Delete.
2609 (double_int_fits_to_tree_p): Delete.
2610 (force_fit_type_double): Delete.
2611 (build_int_cstu): Replace with out-of-line function.
2612 (build_int_cst_wide): Delete.
2613 (tree_int_cst_lt): Define inline.
2614 (tree_int_cst_le): New.
2615 (tree_int_cst_compare): Define inline.
2616 (tree_int_cst_min_precision): Take a signop rather than a bool.
2617 (wi::int_traits <const_tree>): New.
2618 (wi::int_traits <tree>): New.
2619 (wi::extended_tree): New.
2620 (wi::int_traits <wi::extended_tree>): New.
2621 (wi::to_widest): New.
2622 (wi::to_offset): New.
2623 (wi::fits_to_tree_p): New.
2624 (wi::min_value): New.
2625 (wi::max_value): New.
2626 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2627 (copy_tree_body_r): Likewise.
2628 * tree-object-size.c (compute_object_offset): Likewise.
2629 (addr_object_size): Likewise.
2630 * tree-predcom.c: Include wide-int-print.h.
2631 (struct dref_d): Change type of offset to widest_int.
2632 (dump_dref): Call wide-int printer.
2633 (aff_combination_dr_offset): Use wide-int interfaces.
2634 (determine_offset): Take a widest_int pointer rather than a
2636 (split_data_refs_to_components): Use wide-int interfaces.
2637 (suitable_component_p): Likewise.
2638 (order_drefs): Likewise.
2639 (add_ref_to_chain): Likewise.
2640 (valid_initializer_p): Likewise.
2641 (determine_roots_comp): Likewise.
2642 * tree-pretty-print.c: Include wide-int-print.h.
2643 (dump_generic_node): Use wide-int interfaces.
2644 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2645 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2646 (move_fixed_address_to_symbol): Likewise.
2647 (move_hint_to_base): Likewise.
2648 (move_pointer_to_base): Likewise.
2649 (move_variant_to_index): Likewise.
2650 (most_expensive_mult_to_index): Likewise.
2651 (addr_to_parts): Likewise.
2652 (copy_ref_info): Likewise.
2653 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2654 (indirect_refs_may_alias_p): Likewise.
2655 (stmt_kills_ref_p_1): Likewise.
2656 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2657 * tree-ssa-ccp.c: Update comment at top of file. Include
2659 (struct prop_value_d): Change type of mask to widest_int.
2660 (extend_mask): New function.
2661 (dump_lattice_value): Use wide-int interfaces.
2662 (get_default_value): Likewise.
2663 (set_constant_value): Likewise.
2664 (set_value_varying): Likewise.
2665 (valid_lattice_transition): Likewise.
2666 (set_lattice_value): Likewise.
2667 (value_to_double_int): Delete.
2668 (value_to_wide_int): New.
2669 (get_value_from_alignment): Use wide-int interfaces.
2670 (get_value_for_expr): Likewise.
2671 (do_dbg_cnt): Likewise.
2672 (ccp_finalize): Likewise.
2673 (ccp_lattice_meet): Likewise.
2674 (bit_value_unop_1): Use widest_ints rather than double_ints.
2675 (bit_value_binop_1): Likewise.
2676 (bit_value_unop): Use wide-int interfaces.
2677 (bit_value_binop): Likewise.
2678 (bit_value_assume_aligned): Likewise.
2679 (evaluate_stmt): Likewise.
2680 (ccp_fold_stmt): Likewise.
2681 (visit_cond_stmt): Likewise.
2682 (ccp_visit_stmt): Likewise.
2683 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2684 (constant_pointer_difference): Likewise.
2685 (associate_pointerplus): Likewise.
2686 (combine_conversions): Likewise.
2687 * tree-ssa-loop.h: Include wide-int.h.
2688 (struct tree_niter_desc): Change type of max to widest_int.
2689 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2690 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2691 (remove_redundant_iv_tests): Likewise.
2692 (canonicalize_loop_induction_variables): Likewise.
2693 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2694 (constant_multiple_of): Take a widest_int pointer instead of
2695 a double_int pointer.
2696 (get_computation_aff): Use wide-int interfaces.
2697 (ptr_difference_cost): Likewise.
2698 (difference_cost): Likewise.
2699 (get_loop_invariant_expr_id): Likewise.
2700 (get_computation_cost_at): Likewise.
2701 (iv_elimination_compare_lt): Likewise.
2702 (may_eliminate_iv): Likewise.
2703 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2704 instead of double_int.
2705 (max_loop_iterations): Likewise.
2706 (max_stmt_executions): Likewise.
2707 (estimated_stmt_executions): Likewise.
2708 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2709 (split_to_var_and_offset): Use wide-int interfaces.
2710 (determine_value_range): Likewise.
2711 (bound_difference_of_offsetted_base): Likewise.
2712 (bounds_add): Take a widest_int instead of a double_int.
2713 (number_of_iterations_ne_max): Use wide-int interfaces.
2714 (number_of_iterations_ne): Likewise.
2715 (number_of_iterations_lt_to_ne): Likewise.
2716 (assert_loop_rolls_lt): Likewise.
2717 (number_of_iterations_lt): Likewise.
2718 (number_of_iterations_le): Likewise.
2719 (number_of_iterations_cond): Likewise.
2720 (number_of_iterations_exit): Likewise.
2721 (finite_loop_p): Likewise.
2722 (derive_constant_upper_bound_assign): Likewise.
2723 (derive_constant_upper_bound): Return a widest_int.
2724 (derive_constant_upper_bound_ops): Likewise.
2725 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2726 (record_estimate): Take a widest_int rather than a double_int.
2727 (record_nonwrapping_iv): Use wide-int interfaces.
2728 (double_int_cmp): Delete.
2729 (wide_int_cmp): New.
2730 (bound_index): Take a widest_int rather than a double_int.
2731 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2732 (maybe_lower_iteration_bound): Likewise.
2733 (estimate_numbers_of_iterations_loop): Likewise.
2734 (estimated_loop_iterations): Take a widest_int pointer than than
2735 a double_int pointer.
2736 (estimated_loop_iterations_int): Use wide-int interfaces.
2737 (max_loop_iterations): Take a widest_int pointer than than
2738 a double_int pointer.
2739 (max_loop_iterations_int): Use wide-int interfaces.
2740 (max_stmt_executions): Take a widest_int pointer than than
2741 a double_int pointer.
2742 (estimated_stmt_executions): Likewise.
2743 (n_of_executions_at_most): Use wide-int interfaces.
2744 (scev_probably_wraps_p): Likewise.
2745 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2747 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2749 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2750 double_ints. Adjust for trailing_wide_ints <3> representation.
2751 (set_nonzero_bits): Likewise.
2752 (get_range_info): Return wide_ints rather than double_ints.
2753 Adjust for trailing_wide_ints <3> representation.
2754 (get_nonzero_bits): Likewise.
2755 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2757 * tree-ssanames.h (struct range_info_def): Replace min, max and
2758 nonzero_bits with a trailing_wide_ints <3>.
2759 (set_range_info): Use wide_int_refs rather than double_ints.
2760 (set_nonzero_bits): Likewise.
2761 (get_range_info): Return wide_ints rather than double_ints.
2762 (get_nonzero_bits): Likewise.
2763 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2764 * tree-ssa-pre.c (phi_translate_1): Likewise.
2765 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2766 (acceptable_pow_call): Likewise.
2767 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2769 (vn_reference_fold_indirect): Likewise.
2770 (vn_reference_maybe_forwprop_address): Likewise.
2771 (valueize_refs_1): Likewise.
2772 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2773 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2774 tree_int_cst_lt and tree_int_cst_le.
2775 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2777 (streamer_alloc_tree): Likewise.
2778 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2779 (streamer_write_tree_header): Likewise.
2780 (streamer_write_integer_cst): Likewise.
2781 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2782 (build_constructors): Likewise.
2783 (array_value_type): Likewise.
2784 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2785 (vect_check_gather): Likewise.
2786 * tree-vect-generic.c (build_replicated_const): Likewise.
2787 (expand_vector_divmod): Likewise.
2788 * tree-vect-loop.c (vect_transform_loop): Likewise.
2789 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2790 (vect_do_peeling_for_alignment): Likewise.
2791 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2792 * tree-vrp.c: Include wide-int.h.
2793 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2794 (extract_range_from_assert): Use wide-int interfaces.
2795 (vrp_int_const_binop): Likewise.
2796 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2797 double_int pointers.
2798 (ranges_from_anti_range): Use wide-int interfaces.
2799 (quad_int_cmp): Delete.
2800 (quad_int_pair_sort): Likewise.
2801 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2802 (extract_range_from_unary_expr_1): Likewise.
2803 (adjust_range_with_scev): Likewise.
2804 (masked_increment): Take and return wide_ints rather than double_ints.
2805 (register_edge_assert_for_2): Use wide-int interfaces.
2806 (check_array_ref): Likewise.
2807 (search_for_addr_array): Likewise.
2808 (maybe_set_nonzero_bits): Likewise.
2809 (union_ranges): Pass an integer of the correct type instead of
2810 using integer_one_node.
2811 (intersect_ranges): Likewise.
2812 (simplify_truth_ops_using_ranges): Likewise.
2813 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2814 (range_fits_type_p): Likewise.
2815 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2817 (simplify_conversion_using_ranges): Use wide-int interfaces.
2818 (simplify_float_conversion_using_ranges): Likewise.
2819 (vrp_finalize): Likewise.
2820 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2821 (gimple_stringops_transform): Likewise.
2822 * varasm.c (decode_addr_const): Likewise.
2823 (const_hash_1): Likewise.
2824 (const_rtx_hash_1): Likewise
2825 (output_constant): Likewise.
2826 (array_size_for_constructor): Likewise.
2827 (output_constructor_regular_field): Likewise.
2828 (output_constructor_bitfield): Likewise.
2829 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2830 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2832 * gencheck.c: Define BITS_PER_UNIT.
2835 * wide-int-print.cc: New.
2836 * wide-int-print.h: New.
2838 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2840 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2842 2014-05-06 Richard Biener <rguenther@suse.de>
2844 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2845 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2846 (TODO_verify_all): Adjust.
2847 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2848 TODO_verify_stmts and TODO_verify_rtl_sharing.
2849 * bb-reorder.c: Likewise.
2850 * cfgexpand.c: Likewise.
2851 * cprop.c: Likewise.
2853 * function.c: Likewise.
2854 * fwprop.c: Likewise.
2856 * gimple-ssa-isolate-paths.c: Likewise.
2857 * gimple-ssa-strength-reduction.c: Likewise.
2858 * ipa-split.c: Likewise.
2859 * loop-init.c: Likewise.
2860 * loop-unroll.c: Likewise.
2861 * lower-subreg.c: Likewise.
2862 * modulo-sched.c: Likewise.
2863 * postreload-gcse.c: Likewise.
2864 * predict.c: Likewise.
2865 * recog.c: Likewise.
2866 * sched-rgn.c: Likewise.
2867 * store-motion.c: Likewise.
2868 * tracer.c: Likewise.
2869 * trans-mem.c: Likewise.
2870 * tree-call-cdce.c: Likewise.
2871 * tree-cfg.c: Likewise.
2872 * tree-cfgcleanup.c: Likewise.
2873 * tree-complex.c: Likewise.
2874 * tree-eh.c: Likewise.
2875 * tree-emutls.c: Likewise.
2876 * tree-if-conv.c: Likewise.
2877 * tree-into-ssa.c: Likewise.
2878 * tree-loop-distribution.c: Likewise.
2879 * tree-object-size.c: Likewise.
2880 * tree-parloops.c: Likewise.
2881 * tree-pass.h: Likewise.
2882 * tree-sra.c: Likewise.
2883 * tree-ssa-ccp.c: Likewise.
2884 * tree-ssa-copy.c: Likewise.
2885 * tree-ssa-copyrename.c: Likewise.
2886 * tree-ssa-dce.c: Likewise.
2887 * tree-ssa-dom.c: Likewise.
2888 * tree-ssa-dse.c: Likewise.
2889 * tree-ssa-forwprop.c: Likewise.
2890 * tree-ssa-ifcombine.c: Likewise.
2891 * tree-ssa-loop-ch.c: Likewise.
2892 * tree-ssa-loop-ivcanon.c: Likewise.
2893 * tree-ssa-loop.c: Likewise.
2894 * tree-ssa-math-opts.c: Likewise.
2895 * tree-ssa-phiopt.c: Likewise.
2896 * tree-ssa-phiprop.c: Likewise.
2897 * tree-ssa-pre.c: Likewise.
2898 * tree-ssa-reassoc.c: Likewise.
2899 * tree-ssa-sink.c: Likewise.
2900 * tree-ssa-strlen.c: Likewise.
2901 * tree-ssa-tail-merge.c: Likewise.
2902 * tree-ssa-uncprop.c: Likewise.
2903 * tree-switch-conversion.c: Likewise.
2904 * tree-tailcall.c: Likewise.
2905 * tree-vect-generic.c: Likewise.
2906 * tree-vectorizer.c: Likewise.
2907 * tree-vrp.c: Likewise.
2909 * var-tracking.c: Likewise.
2910 * bt-load.c: Likewise.
2911 * cfgcleanup.c: Likewise.
2912 * combine-stack-adj.c: Likewise.
2913 * combine.c: Likewise.
2914 * compare-elim.c: Likewise.
2915 * config/epiphany/resolve-sw-modes.c: Likewise.
2916 * config/i386/i386.c: Likewise.
2917 * config/mips/mips.c: Likewise.
2918 * config/s390/s390.c: Likewise.
2919 * config/sh/sh_treg_combine.cc: Likewise.
2920 * config/sparc/sparc.c: Likewise.
2923 * final.c: Likewise.
2924 * ifcvt.c: Likewise.
2925 * mode-switching.c: Likewise.
2926 * passes.c: Likewise.
2927 * postreload.c: Likewise.
2929 * reg-stack.c: Likewise.
2930 * regcprop.c: Likewise.
2931 * regrename.c: Likewise.
2934 2014-05-06 Richard Biener <rguenther@suse.de>
2937 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2938 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2940 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2943 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2945 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2946 Tom de Vries <tom@codesourcery.com>
2948 * target.def (call_fusage_contains_non_callee_clobbers): New
2950 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2952 (@node Miscellaneous Register Hooks): New node.
2953 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2954 * doc/tm.texi: Regenerate.
2956 2014-05-05 Marek Polacek <polacek@redhat.com>
2959 * opts.c (common_handle_option): Call error_at instead of warning_at.
2961 2014-05-05 Richard Biener <rguenther@suse.de>
2963 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2964 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2965 under the TODO_verify_il umbrella.
2967 2014-05-05 Richard Biener <rguenther@suse.de>
2969 * passes.c (execute_function_todo): Move TODO_verify_flow under
2970 the TODO_verify_ul umbrella.
2972 2014-05-05 Richard Biener <rguenther@suse.de>
2975 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2976 X & CST away from a CST that is the mask of a mode.
2978 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2980 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2981 int argument to enum machine_mode.
2982 (picochip_class_max_nregs): Ditto.
2983 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2984 (picochip_class_max_nregs): Ditto.
2986 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2988 * target.def: Add new target hook.
2989 * doc/tm.texi: Regenerate.
2990 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2991 * targhooks.c (default_keep_leaf_when_profiled): New function.
2993 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2994 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2996 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2998 PR tree-optimization/60363
2999 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3000 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3001 (update_destination_phis): New parameter.
3002 (create_edge_and_update_destination_phis): Ditto.
3003 (ssa_fix_duplicate_block_edges): Pass new arguments.
3004 (thread_single_edge): Ditto.
3006 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3008 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3009 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3010 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3011 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3012 Use RS6000_BTM_HARD_FLOAT.
3013 (BU_MISC_2): Likewise.
3014 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3015 RS6000_BTM_HARD_FLOAT.
3016 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3018 (rs6000_invalid_builtin): Add hard floating builtin support.
3019 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3020 hard float builtins.
3021 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3023 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3025 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3026 Add missing function* argument.
3028 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3030 * lra-constraints.c (valid_address_p): Move earlier in file.
3031 Add a constraint argument to the address_info version.
3032 (satisfies_memory_constraint_p): New function.
3033 (satisfies_address_constraint_p): Likewise.
3034 (process_alt_operands, curr_insn_transform): Use them.
3035 (process_address): Pass the constraint to valid_address_p when
3036 checking address operands.
3038 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3040 * config/mips/mips.c (mips_isa_rev): New variable.
3041 (mips_set_architecture): Set it.
3042 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3044 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3045 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3046 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3047 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3048 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3049 conditions in terms of mips_isa_rev.
3050 (mips_isa_rev): Declare.
3052 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3054 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3055 (prob_unlikely, prob_likely): Make variables const.
3057 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3059 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3061 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3063 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3065 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3067 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3068 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3069 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3071 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3072 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3074 Replace usage of ROUND_REG with sh_round_reg.
3075 Use CEIL instead of ROUND_ADVANCE.
3077 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3080 * config/sh/sh.c: Include stdlib headers before everything else.
3082 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3084 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3085 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3086 (gimplify_adjust_omp_clauses): Simd region is never
3087 directly nested in combined parallel. Instead, for linear
3088 with copyin/copyout, if in combined for simd loop, make decl
3089 firstprivate/lastprivate on OMP_FOR.
3090 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3091 expand_omp_for_static_chunk): When setting endvar, also set
3092 fd->loop.v to the same value.
3094 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3096 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3098 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3100 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3103 2014-05-02 Marek Polacek <polacek@redhat.com>
3105 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3107 2014-05-02 Kito Cheng <kito@0xlab.org>
3109 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3110 to a C expression marco.
3111 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3112 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3113 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3114 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3115 HONOR_REG_ALLOC_ORDER.
3116 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3118 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3120 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3122 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3124 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3126 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3128 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3129 (convert_scalar_cond_reduction): Likewise.
3130 (predicate_scalar_phi): Add recognition and transformation
3131 of simple conditioanl reduction to be vectorizable.
3133 2014-05-01 Marek Polacek <polacek@redhat.com>
3136 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3138 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3140 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3141 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3142 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3143 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3144 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3145 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3146 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3147 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3149 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3151 * config/arc/arc.opt (mlra): Move comment above option name
3152 to avoid mis-parsing as language options.
3154 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3156 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3157 * config/sol2.h: ... here.
3158 * config/sol2-10.h: Remove.
3160 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3161 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3162 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3163 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3164 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3165 * config/sol2.h: ... here.
3166 (SECTION_NAME_FORMAT): Don't redefine.
3167 (STARTFILE_ARCH32_SPEC): Rename to ...
3168 (STARTFILE_ARCH_SPEC): ... this.
3169 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3170 * config/sparc/sol2.h: ... here.
3171 (SECTION_NAME_FORMAT): Don't undef.
3172 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3173 (SUBTARGET_EXTRA_SPECS): Remove.
3174 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3176 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3177 (MD_STARTFILE_PREFIX): Remove.
3178 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3179 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3180 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3181 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3182 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3183 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3184 * config/i386/sol2.h: ... here.
3185 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3186 * config/i386/sol2-bi.h: Remove.
3187 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3188 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3190 * config/i386/t-sol2-64: Rename to ...
3191 * config/i386/t-sol2: ... this.
3192 * config/sparc/t-sol2-64: Rename to ...
3193 * config/sparc/t-sol2: ... this.
3195 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3196 sol2_tm_file_head, sol2_tm_file_tail.
3197 Include ${cpu_type}/sol2.h before sol2.h.
3199 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3200 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3201 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3202 Reflect i386/t-sol2-64 renaming.
3203 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3204 Reflect sparc/t-sol2-64 renaming.
3206 2014-04-30 Richard Biener <rguenther@suse.de>
3208 * passes.c (execute_function_todo): Move TODO_verify_stmts
3209 and TODO_verify_ssa under the TODO_verify_il umbrella.
3210 * tree-ssa.h (verify_ssa): Adjust prototype.
3211 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3212 we should verify SSA operands.
3213 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3214 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3215 whether we should verify whether not throwing stmts have EH info.
3216 * graphite-scop-detection.c (create_sese_edges): Adjust.
3217 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3218 * tree-eh.c (lower_try_finally_switch): Do not add the
3219 default case label twice.
3221 2014-04-30 Marek Polacek <polacek@redhat.com>
3223 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3224 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3225 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3226 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3228 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3230 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3231 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3232 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3233 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3234 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3235 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3236 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3237 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3239 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3241 * tree-cfg.c (dump_function_to_file): Dump the return type of
3242 functions, in a line to itself before the function body, mimicking
3243 the layout of a C function.
3245 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3247 PR tree-optimization/60971
3248 * tree-tailcall.c (process_assignment): Reject conversions which
3251 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3253 * calls.c (initialize_argument_information): Always treat
3254 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3255 (expand_call): Likewise.
3256 (emit_library_call_calue_1): Likewise.
3257 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3258 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3261 2014-04-29 Nick Clifton <nickc@redhat.com>
3263 * config/msp430/msp430.md (umulsidi): Fix typo.
3264 (mulhisi3): Enable even inside interrupt handlers.
3265 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3266 bigger return address pushed in large mode.
3268 2014-04-29 Nick Clifton <nickc@redhat.com>
3270 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3271 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3273 * config/m32r/m32r.c (init_reg_tables): Likewise.
3274 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3275 enum to hold the modes.
3277 2014-04-29 Richard Biener <rguenther@suse.de>
3279 * dominance.c (free_dominance_info): Add overload with
3281 (dom_info_state): Likewise.
3282 (dom_info_available_p): Likewise.
3283 * basic-block.h (free_dominance_info, dom_info_state,
3284 dom_info_available_p): Declare overloads.
3285 * passes.c (execute_function_todo): Verify that verifiers
3286 don't change dominator info state. Drop dominator info
3287 for IPA pass invocations.
3288 * cgraph.c (release_function_body): Restore asserts that
3289 dominator information is released.
3291 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3293 * doc/invoke.texi: Fix typo.
3294 * tree-vrp.c: Fix typos.
3295 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3297 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3299 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3301 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3303 * config/aarch64/aarch64-builtins.c
3304 (aarch64_types_storestruct_lane_qualifiers): New.
3305 (TYPES_STORESTRUCT_LANE): Likewise.
3306 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3307 (st3_lane): Likewise.
3308 (st4_lane): Likewise.
3309 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3310 (vec_store_lanesci_lane<mode>): Likewise.
3311 (vec_store_lanesxi_lane<mode>): Likewise.
3312 (aarch64_st2_lane<VQ:mode>): Likewise.
3313 (aarch64_st3_lane<VQ:mode>): Likewise.
3314 (aarch64_st4_lane<VQ:mode>): Likewise.
3315 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3316 * config/aarch64/arm_neon.h
3317 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3318 use new macro arguments.
3319 (__ST3_LANE_FUNC): Likewise.
3320 (__ST4_LANE_FUNC): Likewise.
3321 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3322 (V_THREE_ELEM): Likewise.
3323 (V_FOUR_ELEM): Likewise.
3325 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3327 * doc/gimple.texi: Replace the description of the now-defunct
3328 union gimple_statement_d with a diagram showing the
3329 gimple_statement_base class hierarchy and its relationships to
3330 the GSS_ and GIMPLE_ enums.
3332 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3334 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3335 * config/aarch64/aarch64.c
3336 (aarch64_cannot_change_mode_class): Weaken conditions.
3337 (aarch64_modes_tieable_p): New.
3338 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3340 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3342 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3343 (loadsync_<mode>): Change mode.
3344 (load_quadpti, store_quadpti): New.
3345 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3346 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3348 2014-04-28 Martin Jambor <mjambor@suse.cz>
3350 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3351 same alias type as the original statement.
3352 (subreplacement_assignment_data): New type.
3353 (handle_unscalarized_data_in_subtree): New type of parameter,
3354 generate new memory accesses with same alias type as the original
3356 (load_assign_lhs_subreplacements): Likewise.
3357 (sra_modify_constructor_assign): Generate new memory accesses with
3358 same alias type as the original statement.
3360 2014-04-28 Richard Biener <rguenther@suse.de>
3362 * tree-pass.h (TODO_verify_il): Define.
3363 (TODO_verify_all): Complete properly.
3364 * passes.c (execute_function_todo): Move existing loop-closed
3365 SSA verification under TODO_verify_il.
3366 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3367 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3368 Fix tree sharing issue.
3370 2014-04-28 Richard Biener <rguenther@suse.de>
3373 * builtins.def (DEF_C11_BUILTIN): Add.
3374 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3375 * coretypes.h (enum function_class): Add function_c11_misc.
3376 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3377 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3378 (call_may_clobber_ref_p_1): Likewise.
3379 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3380 (mark_all_reaching_defs_necessary_1): Likewise.
3381 (propagate_necessity): Likewise.
3382 (eliminate_unnecessary_stmts): Likewise.
3383 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3385 2014-04-28 Richard Biener <rguenther@suse.de>
3387 * tree-vrp.c (vrp_var_may_overflow): Remove.
3388 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3389 with overflow immediately bump to one before that value and
3390 let iteration figure out overflow status.
3392 2014-04-28 Richard Biener <rguenther@suse.de>
3394 * configure.ac: Do valgrind header checks unconditionally.
3395 Add --enable-valgrind-annotations.
3396 * system.h: Guard valgrind header inclusion with
3397 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3398 * alloc-pool.c (pool_alloc, pool_free): Use
3399 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3400 to guard possibly dead code.
3401 * config.in: Regenerated.
3402 * configure: Likewise.
3404 2014-04-28 Jeff Law <law@redhat.com>
3406 PR tree-optimization/60902
3407 * tree-ssa-threadedge.c
3408 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3409 over real defs when invalidating outputs from statements that do not
3410 produce useful outputs for threading.
3412 2014-04-28 Richard Biener <rguenther@suse.de>
3414 PR tree-optimization/60979
3415 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3416 SCOPs that end in a block with a successor with abnormal
3419 2014-04-28 Richard Biener <rguenther@suse.de>
3421 * tree-pass.h (execute_pass_list): Adjust prototype.
3422 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3423 (do_per_function): Change callback signature, push all actual
3424 work to the callbals.
3425 (do_per_function_toporder): Likewise.
3426 (execute_function_dump): Adjust.
3427 (execute_function_todo): Likewise.
3428 (clear_last_verified): Likewise.
3429 (verify_curr_properties): Likewise.
3430 (update_properties_after_pass): Likewise.
3431 (execute_pass_list_1): Split out from ...
3432 (execute_pass_list): ... here. Adjust.
3433 (execute_ipa_pass_list): Likewise.
3434 * cgraphunit.c (cgraph_add_new_function): Adjust.
3435 (analyze_function): Likewise.
3436 (expand_function): Likewise.
3437 * cgraph.c (release_function_body): Free dominance info
3438 here instead of asserting it was magically freed elsewhere.
3440 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3442 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3443 * configure: Regenerate.
3444 * config/sparc/sparc.opt (muser-mode): New option.
3445 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3447 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3448 * doc/invoke.texi (SPARC options): Document -muser-mode.
3450 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3452 * cselib.c (find_slot_memmode): Delete.
3453 (cselib_hasher): Change compare_type to a struct.
3454 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3456 (preserve_constants_and_equivs): Adjust for new compare_type.
3457 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3458 (wrap_constant): Delete.
3459 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3461 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3463 * doc/install.texi (Building with profile feedback): Remove
3466 2014-04-26 Tom de Vries <tom@codesourcery.com>
3468 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3471 2014-04-25 Cary Coutant <ccoutant@google.com>
3474 * dwarf2out.c (should_move_die_to_comdat): A type definition
3475 can contain a subprogram definition, but don't move it to a
3477 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3478 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3480 (clone_tree_hash): Rename to...
3481 (clone_tree_partial): ...this; change callers. Copy
3482 DW_TAG_subprogram DIEs as declarations.
3483 (copy_decls_walk): Don't copy children of a declaration into a
3486 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3489 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3492 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3494 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3495 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3497 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3500 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3502 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3504 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3506 PR tree-optimization/60930
3507 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3508 creating a multiply candidate by folding two constant
3509 multiplicands when the result overflows.
3511 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3513 PR tree-optimization/60960
3514 * tree-vect-generic.c (expand_vector_operation): Only call
3515 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3517 2014-04-25 Tom de Vries <tom@codesourcery.com>
3519 * expr.c (clobber_reg_mode): New function.
3520 * expr.h (clobber_reg): New function.
3522 2014-04-25 Tom de Vries <tom@codesourcery.com>
3524 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3527 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3528 Tom de Vries <tom@codesourcery.com>
3530 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3532 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3533 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3534 new argument to find_all_hard_reg_sets call.
3536 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3538 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3539 Use HOST_WIDE_INT_C for mask literal.
3540 (aarch_rev16_shleft_mask_imm_p): Likewise.
3542 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3545 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3547 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3549 PR preprocessor/56540
3550 * config/i386/i386-c.c (ix86_target_macros): Define
3551 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3553 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3555 * configure.ac (tga_func): Remove.
3556 (LIB_TLS_SPEC): Remove.
3557 * configure: Regenerate.
3558 * config.in: Regenerate.
3559 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3561 2014-04-25 Richard Biener <rguenther@suse.de>
3564 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3565 call stmt use/clobber sets during stmt walk instead of
3566 walking the possibly incomplete set of caller edges.
3568 2014-04-25 Richard Biener <rguenther@suse.de>
3571 * passes.c (apply_ipa_transforms): Inline into only caller ...
3572 (execute_one_pass): ... here. Properly bring in function
3573 bodies for nodes we want to apply IPA transforms to.
3575 2014-04-24 Cong Hou <congh@google.com>
3577 PR tree-optimization/60896
3578 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3579 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3580 (vect_mark_pattern_stmts): Set the def type of all statements in
3581 PATTERN_DEF_SEQ as vect_internal_def.
3583 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3585 * doc/extend.texi (PowerPC Built-in Functions): Document new
3586 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3587 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3589 * config/rs6000/predicates.md (const_0_to_3_operand): New
3590 predicate to match 0..3 integer constants.
3592 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3593 to support adding miscellaneous builtin functions.
3594 (BU_DFP_MISC_2): Likewise.
3595 (BU_P7_MISC_1): Likewise.
3596 (BU_P7_MISC_2): Likewise.
3597 (BU_P8V_MISC_3): Likewise.
3598 (BU_MISC_1): Likewise.
3599 (BU_MISC_2): Likewise.
3600 (DIVWE): Add extended divide builtin functions.
3603 (DIVWEUO): Likewise.
3607 (DIVDEUO): Likewise.
3608 (DXEX): Add decimal floating-point builtin functions.
3611 (DDEDPDQ): Likewise.
3613 (DENBCDQ): Likewise.
3620 (CDTBCD): Add new BCD builtin functions.
3624 (BCDADD_LT): Likewise.
3625 (BCDADD_EQ): Likewise.
3626 (BCDADD_GT): Likewise.
3627 (BCDADD_OV): Likewise.
3629 (BCDSUB_LT): Likewise.
3630 (BCDSUB_EQ): Likewise.
3631 (BCDSUB_GT): Likewise.
3632 (BCDSUB_OV): Likewise.
3633 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3634 (UNPACK_TD): Likewise.
3635 (PACK_TF): Likewise.
3636 (UNPACK_TF): Likewise.
3637 (UNPACK_TF_0): Likewise.
3638 (UNPACK_TF_1): Likewise.
3639 (PACK_V1TI): Likewise.
3640 (UNPACK_V1TI): Likewise.
3642 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3643 support for decimal floating point builtin functions.
3644 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3645 functions that take constant arguments.
3646 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3647 (rs6000_init_builtins): Setup long double, _Decimal64, and
3648 _Decimal128 types for new builtin functions.
3649 (builtin_function_type): Set the unsigned flags appropriately for
3650 the new builtin functions.
3651 (rs6000_opt_masks): Add support for decimal floating point builtin
3654 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3655 floating point builtin functions.
3656 (RS6000_BTM_COMMON): Likewise.
3657 (RS6000_BTI_long_double): Likewise.
3658 (RS6000_BTI_dfloat64): Likewise.
3659 (RS6000_BTI_dfloat128): Likewise.
3660 (long_double_type_internal_node): Likewise.
3661 (dfloat64_type_internal_node): Likewise.
3662 (dfloat128_type_internal_node): Likewise.
3664 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3665 2.07 bcd arithmetic instructions.
3666 (UNSPEC_BCDSUB): Likewise.
3667 (UNSPEC_BCD_OVERFLOW): Likewise.
3668 (UNSPEC_BCD_ADD_SUB): Likewise.
3669 (bcd_add_sub): Likewise.
3670 (BCD_TEST): Likewise.
3671 (bcd<bcd_add_sub>): Likewise.
3672 (bcd<bcd_add_sub>_test): Likewise.
3673 (bcd<bcd_add_sub>_test2): Likewise.
3674 (bcd<bcd_add_sub>_<code>): Likewise.
3675 (peephole2 for combined bcd ops): Likewise.
3677 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3678 decimal floating point builtin functions.
3679 (UNSPEC_DENBCD): Likewise.
3680 (UNSPEC_DXEX): Likewise.
3681 (UNSPEC_DIEX): Likewise.
3682 (UNSPEC_DSCLI): Likewise.
3683 (UNSPEC_DSCRI): Likewise.
3684 (D64_D128): Likewise.
3685 (dfp_suffix): Likewise.
3686 (dfp_ddedpd_<mode>): Likewise.
3687 (dfp_denbcd_<mode>): Likewise.
3688 (dfp_dxex_<mode>): Likewise.
3689 (dfp_diex_<mode>): Likewise.
3690 (dfp_dscli_<mode>): Likewise.
3691 (dfp_dscri_<mode>): Likewise.
3693 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3695 (UNSPEC_CDTBCD): Likewise.
3696 (UNSPEC_CBCDTD): Likewise.
3697 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3698 (UNSPEC_DIVEO): Likewise.
3699 (UNSPEC_DIVEU): Likewise.
3700 (UNSPEC_DIVEUO): Likewise.
3701 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3702 pack/unpack 128-bit types.
3703 (UNSPEC_PACK_128BIT): Likewise.
3704 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3705 (udiv<mode>3): Use idiv_ldiv mode attribute.
3706 (div<mode>3): Likewise.
3707 (addg6s): Add new BCD builtin functions.
3710 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3711 (div_extend): Likewise.
3712 (div<div_extend>_<mode>"): Likewise.
3713 (FP128_64): Add support for new builtin functions to pack/unpack
3715 (unpack<mode>): Likewise.
3716 (unpacktf_0): Likewise.
3717 (unpacktf_1): Likewise.
3718 (unpack<mode>_dm): Likewise.
3719 (unpack<mode>_nodm): Likewise.
3720 (pack<mode>): Likewise.
3721 (unpackv1ti): Likewise.
3722 (packv1ti): Likewise.
3724 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3726 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3729 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3731 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3732 * gimplify.c (omp_is_private): Change last argument's type to int.
3733 Only diagnose lastprivate if the simd argument is 1, only diagnose
3734 linear if the simd argument is 2.
3735 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3736 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3737 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3738 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3739 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3740 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3741 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3742 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3743 * tree-nested.c (convert_nonlocal_omp_clauses,
3744 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3746 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3749 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3752 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3754 * flag-types.h (enum ivar_visibility): Add.
3756 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3758 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3759 function * argument.
3761 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3763 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3765 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3766 Tom de Vries <tom@codesourcery.com>
3768 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3769 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3771 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3772 * emit-rtl.c (try_split): Same.
3774 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3775 Tom de Vries <tom@codesourcery.com>
3777 * common.opt (fuse-caller-save): New option.
3779 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3781 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3782 elements for big-endian.
3784 2014-04-24 Richard Biener <rguenther@suse.de>
3786 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3787 during TER and instead use the sepops interface for expanding
3788 non-GIMPLE_SINGLE_RHS.
3790 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3792 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3793 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3795 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3797 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3798 assembler 64-bit option.
3799 * configure: Regenerate.
3801 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3803 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3804 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3805 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3806 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3807 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3809 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3811 * config/aarch64/aarch64-builtins.c
3812 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3813 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3814 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3815 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3817 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3818 (Vrevsuff): New mode attribute.
3820 2014-04-24 Terry Guo <terry.guo@arm.com>
3822 * config/arm/arm.h (machine_function): Define variable
3823 after_arm_reorg here.
3824 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3825 (arm_split_constant): Update the way to access variable
3828 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3830 2014-04-23 Tom de Vries <tom@codesourcery.com>
3832 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3834 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3836 * is-a.h: Update comments to reflect the following changes to the
3837 "pointerness" of the API, making the template parameter match the
3838 return type, allowing use of is-a.h with typedefs of pointers.
3839 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3840 that the return type matches the parameter to the is_a_helper.
3842 (dyn_cast): Likewise.
3844 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3845 pointer from the is-a.h API.
3847 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3848 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3850 (is_a_helper <varpool_node>::test): Likewise, convert to...
3851 (is_a_helper <varpool_node *>::test): ...this.
3853 (varpool_first_variable): Update for removal of implicit pointer
3854 from the is-a.h API.
3855 (varpool_next_variable): Likewise.
3856 (varpool_first_static_initializer): Likewise.
3857 (varpool_next_static_initializer): Likewise.
3858 (varpool_first_defined_variable): Likewise.
3859 (varpool_next_defined_variable): Likewise.
3860 (cgraph_first_defined_function): Likewise.
3861 (cgraph_next_defined_function): Likewise.
3862 (cgraph_first_function): Likewise.
3863 (cgraph_next_function): Likewise.
3864 (cgraph_first_function_with_gimple_body): Likewise.
3865 (cgraph_next_function_with_gimple_body): Likewise.
3866 (cgraph_alias_target): Likewise.
3867 (varpool_alias_target): Likewise.
3868 (cgraph_function_or_thunk_node): Likewise.
3869 (varpool_variable_node): Likewise.
3870 (symtab_real_symbol_p): Likewise.
3871 * cgraphunit.c (referred_to_p): Likewise.
3872 (analyze_functions): Likewise.
3873 (handle_alias_pairs): Likewise.
3874 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3875 * gimple-ssa.h (gimple_vuse_op): Likewise.
3876 (gimple_vdef_op): Likewise.
3877 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3878 * gimple.c (gimple_build_asm_1): Likewise.
3879 (gimple_build_try): Likewise.
3880 (gimple_build_resx): Likewise.
3881 (gimple_build_eh_dispatch): Likewise.
3882 (gimple_build_omp_for): Likewise.
3883 (gimple_omp_for_set_clauses): Likewise.
3885 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3886 (is_a_helper <gimple_statement_asm *>::test): ...this.
3887 (is_a_helper <gimple_statement_bind>::test): Convert to...
3888 (is_a_helper <gimple_statement_bind *>::test): ...this.
3889 (is_a_helper <gimple_statement_call>::test): Convert to...
3890 (is_a_helper <gimple_statement_call *>::test): ...this.
3891 (is_a_helper <gimple_statement_catch>::test): Convert to...
3892 (is_a_helper <gimple_statement_catch *>::test): ...this.
3893 (is_a_helper <gimple_statement_resx>::test): Convert to...
3894 (is_a_helper <gimple_statement_resx *>::test): ...this.
3895 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3896 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3897 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3898 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3899 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3900 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3901 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3902 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3903 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3904 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3905 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3906 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3907 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3908 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3909 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3910 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3911 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3912 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3913 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3914 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3915 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3916 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3917 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3918 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3919 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3920 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3921 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3922 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3923 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3924 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3925 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3926 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3927 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3928 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3929 (is_a_helper <gimple_statement_phi>::test): Convert to...
3930 (is_a_helper <gimple_statement_phi *>::test): ...this.
3931 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3932 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3933 (is_a_helper <gimple_statement_try>::test): Convert to...
3934 (is_a_helper <gimple_statement_try *>::test): ...this.
3935 (is_a_helper <gimple_statement_wce>::test): Convert to...
3936 (is_a_helper <gimple_statement_wce *>::test): ...this.
3937 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3938 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3939 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3940 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3941 (is_a_helper <const gimple_statement_call>::test): Convert to...
3942 (is_a_helper <const gimple_statement_call *>::test): ...this.
3943 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3944 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3945 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3946 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3947 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3948 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3949 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3950 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3951 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3953 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3955 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3957 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3959 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3960 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3961 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3963 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3964 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3966 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3967 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3968 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3969 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3970 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3971 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3973 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3974 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3975 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3976 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3978 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3979 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3980 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3981 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3982 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3983 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3984 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3985 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3986 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3987 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3988 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3989 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3990 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3991 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3992 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3993 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3995 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3997 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3998 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4000 (gimple_use_ops): Update for removal of implicit pointer from the
4002 (gimple_set_use_ops): Likewise.
4003 (gimple_vuse): Likewise.
4004 (gimple_vdef): Likewise.
4005 (gimple_vuse_ptr): Likewise.
4006 (gimple_vdef_ptr): Likewise.
4007 (gimple_set_vuse): Likewise.
4008 (gimple_set_vdef): Likewise.
4009 (gimple_omp_return_set_lhs): Likewise.
4010 (gimple_omp_return_lhs): Likewise.
4011 (gimple_omp_return_lhs_ptr): Likewise.
4012 (gimple_call_fntype): Likewise.
4013 (gimple_call_set_fntype): Likewise.
4014 (gimple_call_set_internal_fn): Likewise.
4015 (gimple_call_use_set): Likewise.
4016 (gimple_call_clobber_set): Likewise.
4017 (gimple_bind_vars): Likewise.
4018 (gimple_bind_set_vars): Likewise.
4019 (gimple_bind_body_ptr): Likewise.
4020 (gimple_bind_set_body): Likewise.
4021 (gimple_bind_add_stmt): Likewise.
4022 (gimple_bind_block): Likewise.
4023 (gimple_bind_set_block): Likewise.
4024 (gimple_asm_ninputs): Likewise.
4025 (gimple_asm_noutputs): Likewise.
4026 (gimple_asm_nclobbers): Likewise.
4027 (gimple_asm_nlabels): Likewise.
4028 (gimple_asm_input_op): Likewise.
4029 (gimple_asm_input_op_ptr): Likewise.
4030 (gimple_asm_output_op): Likewise.
4031 (gimple_asm_output_op_ptr): Likewise.
4032 (gimple_asm_set_output_op): Likewise.
4033 (gimple_asm_clobber_op): Likewise.
4034 (gimple_asm_set_clobber_op): Likewise.
4035 (gimple_asm_label_op): Likewise.
4036 (gimple_asm_set_label_op): Likewise.
4037 (gimple_asm_string): Likewise.
4038 (gimple_catch_types): Likewise.
4039 (gimple_catch_types_ptr): Likewise.
4040 (gimple_catch_handler_ptr): Likewise.
4041 (gimple_catch_set_types): Likewise.
4042 (gimple_catch_set_handler): Likewise.
4043 (gimple_eh_filter_types): Likewise.
4044 (gimple_eh_filter_types_ptr): Likewise.
4045 (gimple_eh_filter_failure_ptr): Likewise.
4046 (gimple_eh_filter_set_types): Likewise.
4047 (gimple_eh_filter_set_failure): Likewise.
4048 (gimple_eh_must_not_throw_fndecl): Likewise.
4049 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4050 (gimple_eh_else_n_body_ptr): Likewise.
4051 (gimple_eh_else_e_body_ptr): Likewise.
4052 (gimple_eh_else_set_n_body): Likewise.
4053 (gimple_eh_else_set_e_body): Likewise.
4054 (gimple_try_eval_ptr): Likewise.
4055 (gimple_try_cleanup_ptr): Likewise.
4056 (gimple_try_set_eval): Likewise.
4057 (gimple_try_set_cleanup): Likewise.
4058 (gimple_wce_cleanup_ptr): Likewise.
4059 (gimple_wce_set_cleanup): Likewise.
4060 (gimple_phi_capacity): Likewise.
4061 (gimple_phi_num_args): Likewise.
4062 (gimple_phi_result): Likewise.
4063 (gimple_phi_result_ptr): Likewise.
4064 (gimple_phi_set_result): Likewise.
4065 (gimple_phi_arg): Likewise.
4066 (gimple_phi_set_arg): Likewise.
4067 (gimple_resx_region): Likewise.
4068 (gimple_resx_set_region): Likewise.
4069 (gimple_eh_dispatch_region): Likewise.
4070 (gimple_eh_dispatch_set_region): Likewise.
4071 (gimple_omp_critical_name): Likewise.
4072 (gimple_omp_critical_name_ptr): Likewise.
4073 (gimple_omp_critical_set_name): Likewise.
4074 (gimple_omp_for_clauses): Likewise.
4075 (gimple_omp_for_clauses_ptr): Likewise.
4076 (gimple_omp_for_set_clauses): Likewise.
4077 (gimple_omp_for_collapse): Likewise.
4078 (gimple_omp_for_index): Likewise.
4079 (gimple_omp_for_index_ptr): Likewise.
4080 (gimple_omp_for_set_index): Likewise.
4081 (gimple_omp_for_initial): Likewise.
4082 (gimple_omp_for_initial_ptr): Likewise.
4083 (gimple_omp_for_set_initial): Likewise.
4084 (gimple_omp_for_final): Likewise.
4085 (gimple_omp_for_final_ptr): Likewise.
4086 (gimple_omp_for_set_final): Likewise.
4087 (gimple_omp_for_incr): Likewise.
4088 (gimple_omp_for_incr_ptr): Likewise.
4089 (gimple_omp_for_set_incr): Likewise.
4090 (gimple_omp_for_pre_body_ptr): Likewise.
4091 (gimple_omp_for_set_pre_body): Likewise.
4092 (gimple_omp_parallel_clauses): Likewise.
4093 (gimple_omp_parallel_clauses_ptr): Likewise.
4094 (gimple_omp_parallel_set_clauses): Likewise.
4095 (gimple_omp_parallel_child_fn): Likewise.
4096 (gimple_omp_parallel_child_fn_ptr): Likewise.
4097 (gimple_omp_parallel_set_child_fn): Likewise.
4098 (gimple_omp_parallel_data_arg): Likewise.
4099 (gimple_omp_parallel_data_arg_ptr): Likewise.
4100 (gimple_omp_parallel_set_data_arg): Likewise.
4101 (gimple_omp_task_clauses): Likewise.
4102 (gimple_omp_task_clauses_ptr): Likewise.
4103 (gimple_omp_task_set_clauses): Likewise.
4104 (gimple_omp_task_child_fn): Likewise.
4105 (gimple_omp_task_child_fn_ptr): Likewise.
4106 (gimple_omp_task_set_child_fn): Likewise.
4107 (gimple_omp_task_data_arg): Likewise.
4108 (gimple_omp_task_data_arg_ptr): Likewise.
4109 (gimple_omp_task_set_data_arg): Likewise.
4110 (gimple_omp_taskreg_clauses): Likewise.
4111 (gimple_omp_taskreg_clauses_ptr): Likewise.
4112 (gimple_omp_taskreg_set_clauses): Likewise.
4113 (gimple_omp_taskreg_child_fn): Likewise.
4114 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4115 (gimple_omp_taskreg_set_child_fn): Likewise.
4116 (gimple_omp_taskreg_data_arg): Likewise.
4117 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4118 (gimple_omp_taskreg_set_data_arg): Likewise.
4119 (gimple_omp_task_copy_fn): Likewise.
4120 (gimple_omp_task_copy_fn_ptr): Likewise.
4121 (gimple_omp_task_set_copy_fn): Likewise.
4122 (gimple_omp_task_arg_size): Likewise.
4123 (gimple_omp_task_arg_size_ptr): Likewise.
4124 (gimple_omp_task_set_arg_size): Likewise.
4125 (gimple_omp_task_arg_align): Likewise.
4126 (gimple_omp_task_arg_align_ptr): Likewise.
4127 (gimple_omp_task_set_arg_align): Likewise.
4128 (gimple_omp_single_clauses): Likewise.
4129 (gimple_omp_single_clauses_ptr): Likewise.
4130 (gimple_omp_single_set_clauses): Likewise.
4131 (gimple_omp_target_clauses): Likewise.
4132 (gimple_omp_target_clauses_ptr): Likewise.
4133 (gimple_omp_target_set_clauses): Likewise.
4134 (gimple_omp_target_child_fn): Likewise.
4135 (gimple_omp_target_child_fn_ptr): Likewise.
4136 (gimple_omp_target_set_child_fn): Likewise.
4137 (gimple_omp_target_data_arg): Likewise.
4138 (gimple_omp_target_data_arg_ptr): Likewise.
4139 (gimple_omp_target_set_data_arg): Likewise.
4140 (gimple_omp_teams_clauses): Likewise.
4141 (gimple_omp_teams_clauses_ptr): Likewise.
4142 (gimple_omp_teams_set_clauses): Likewise.
4143 (gimple_omp_sections_clauses): Likewise.
4144 (gimple_omp_sections_clauses_ptr): Likewise.
4145 (gimple_omp_sections_set_clauses): Likewise.
4146 (gimple_omp_sections_control): Likewise.
4147 (gimple_omp_sections_control_ptr): Likewise.
4148 (gimple_omp_sections_set_control): Likewise.
4149 (gimple_omp_for_set_cond): Likewise.
4150 (gimple_omp_for_cond): Likewise.
4151 (gimple_omp_atomic_store_set_val): Likewise.
4152 (gimple_omp_atomic_store_val): Likewise.
4153 (gimple_omp_atomic_store_val_ptr): Likewise.
4154 (gimple_omp_atomic_load_set_lhs): Likewise.
4155 (gimple_omp_atomic_load_lhs): Likewise.
4156 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4157 (gimple_omp_atomic_load_set_rhs): Likewise.
4158 (gimple_omp_atomic_load_rhs): Likewise.
4159 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4160 (gimple_omp_continue_control_def): Likewise.
4161 (gimple_omp_continue_control_def_ptr): Likewise.
4162 (gimple_omp_continue_set_control_def): Likewise.
4163 (gimple_omp_continue_control_use): Likewise.
4164 (gimple_omp_continue_control_use_ptr): Likewise.
4165 (gimple_omp_continue_set_control_use): Likewise.
4166 (gimple_transaction_body_ptr): Likewise.
4167 (gimple_transaction_label): Likewise.
4168 (gimple_transaction_label_ptr): Likewise.
4169 (gimple_transaction_set_body): Likewise.
4170 (gimple_transaction_set_label): Likewise.
4172 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4173 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4174 * ipa-ref.c (ipa_record_reference): Likewise.
4175 * ipa-reference.c (analyze_function): Likewise.
4176 (ipa_reference_write_optimization_summary): Likewise.
4177 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4178 (address_taken_from_non_vtable_p): Likewise.
4179 (comdat_can_be_unshared_p_1): Likewise.
4180 * lto-cgraph.c (lto_output_ref): Likewise.
4181 (add_references): Likewise.
4182 (compute_ltrans_boundary): Likewise.
4183 (output_symtab): Likewise.
4184 (input_ref): Likewise.
4185 (input_cgraph_1): Likewise.
4186 (output_cgraph_opt_summary): Likewise.
4187 * lto-streamer-out.c (lto_output): Likewise.
4188 (output_symbol_p): Likewise.
4189 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4190 (lsei_start_function_in_partition): Likewise.
4191 (lsei_next_variable_in_partition): Likewise.
4192 (lsei_start_variable_in_partition): Likewise.
4193 * symtab.c (insert_to_assembler_name_hash): Likewise.
4194 (unlink_from_assembler_name_hash): Likewise.
4195 (symtab_unregister_node): Likewise.
4196 (symtab_remove_node): Likewise.
4197 (dump_symtab_node): Likewise.
4198 (verify_symtab_base): Likewise.
4199 (verify_symtab_node): Likewise.
4200 (symtab_make_decl_local): Likewise.
4201 (symtab_alias_ultimate_target): Likewise.
4202 (symtab_resolve_alias): Likewise.
4203 (symtab_get_symbol_partitioning_class): Likewise.
4204 * tree-phinodes.c (allocate_phi_node): Likewise.
4205 (reserve_phi_args_for_new_edge): Likewise.
4206 (remove_phi_args): Likewise.
4207 * varpool.c (varpool_node_for_asm): Likewise.
4208 (varpool_remove_unreferenced_decls): Likewise.
4210 2014-04-23 Jeff Law <law@redhat.com>
4212 PR tree-optimization/60902
4213 * tree-ssa-threadedge.c
4214 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4215 invalidate outputs from statements that do not produce useful
4216 outputs for threading.
4218 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4220 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4221 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4222 machine descriptions for Stack Smashing Protector.
4224 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4226 * aarch64.md (<optab>_rol<mode>3): New pattern.
4227 (<optab>_rolsi3_uxtw): Likewise.
4228 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4230 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4232 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4233 (arm_cortex_a12_tune): Likewise.
4235 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4237 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4239 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4241 * config/arm/arm.md (arm_rev16si2): New pattern.
4242 (arm_rev16si2_alt): Likewise.
4243 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4245 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4247 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4248 (rev16<mode>2_alt): Likewise.
4249 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4250 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4251 (aarch_rev16_shleft_mask_imm_p): Likewise.
4252 (aarch_rev16_p_1): Likewise.
4253 (aarch_rev16_p): Likewise.
4254 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4255 (aarch_rev16_shright_mask_imm_p): Likewise.
4256 (aarch_rev16_shleft_mask_imm_p): Likewise.
4258 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4260 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4261 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4263 (cortex_a53_extra_costs): Likewise.
4264 (cortex_a57_extra_costs): Likewise.
4265 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4266 (cortexa7_extra_costs): Likewise.
4267 (cortexa8_extra_costs): Likewise.
4268 (cortexa12_extra_costs): Likewise.
4269 (cortexa15_extra_costs): Likewise.
4270 (v7m_extra_costs): Likewise.
4271 (arm_new_rtx_costs): Handle BSWAP.
4273 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4275 * config/arm/arm.c (cortexa8_extra_costs): New table.
4276 (arm_cortex_a8_tune): New tuning struct.
4277 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4279 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4281 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4283 2014-04-23 Richard Biener <rguenther@suse.de>
4285 * Makefile.in (OBJS): Remove loop-unswitch.o.
4286 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4287 * passes.def (pass_rtl_unswitch): Likewise.
4288 * loop-init.c (gate_rtl_unswitch): Likewise.
4289 (rtl_unswitch): Likewise.
4290 (pass_data_rtl_unswitch): Likewise.
4291 (pass_rtl_unswitch): Likewise.
4292 (make_pass_rtl_unswitch): Likewise.
4293 * rtl.h (reversed_condition): Likewise.
4294 (compare_and_jump_seq): Likewise.
4295 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4297 * loop-unroll.c (compare_and_jump_seq): Likewise.
4299 2014-04-23 Richard Biener <rguenther@suse.de>
4301 PR tree-optimization/60903
4302 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4303 commented code block.
4304 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4305 loop flags to newly created BBs and edges.
4307 2014-04-23 Nick Clifton <nickc@redhat.com>
4309 * config/msp430/msp430.c (msp430_handle_option): Move function
4311 (msp430_option_override): Simplify mcu and mcpu option handling.
4312 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4313 support for -mhwmult command line option.
4314 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4315 -mhwmult command line option.
4316 (msp430_hwmult_enabled): Delete.
4317 (msp43o_output_labelref): Add support for -mhwmult command line option.
4318 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4319 (umulsidi3): Likewise.
4320 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4321 (mcpu, mlarge, msmall): Likewise.
4322 (mhwmult): New option.
4323 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4325 (msp430_is_f5_mcu): Remove prototype.
4326 (msp430_use_f5_series_hwmult): Add prototype.
4327 * config/msp430/msp430-opts.h: New file.
4328 * common/config/msp430: New directory.
4329 * common/config/msp430/msp430-common.c: New file.
4330 * config.gcc (msp430): Remove target_has_targetm_common.
4331 * doc/invoke.texi: Document -mhwmult command line option.
4333 2014-04-23 Nick Clifton <nickc@redhat.com>
4335 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4336 default-manifest.o if it can be found in the search path.
4337 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4339 2014-04-23 Terry Guo <terry.guo@arm.com>
4341 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4343 2014-04-23 Richard Biener <rguenther@suse.de>
4346 * tree-inline.c (declare_return_variable): Use mark_addressable.
4348 2014-04-23 Richard Biener <rguenther@suse.de>
4351 * loop-init.c (loop_optimizer_init): Make sure to apply
4352 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4354 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4357 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4359 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4360 if flag_sanitize_undefined_trap_on_error.
4361 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4362 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4363 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4364 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4365 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4366 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4367 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4368 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4369 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4370 * ubsan.c (ubsan_instrument_unreachable): Return
4371 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4372 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4373 if flag_sanitize_undefined_trap_on_error and
4374 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4375 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4376 instrument_bool_enum_load): Emit __builtin_trap () if
4377 flag_sanitize_undefined_trap_on_error and
4378 __builtin_handle_*_abort () if !flag_sanitize_recover.
4379 * doc/invoke.texi (-fsanitize-recover,
4380 -fsanitize-undefined-trap-on-error): Document.
4382 2014-04-22 Christian Bruel <christian.bruel@st.com>
4384 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4385 Force immediates to SImode.
4387 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4389 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4391 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4392 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4393 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4394 (nios2_fpu_insn): Add entry for round.
4395 (N2FPU_NO_ERRNO_P): Define.
4396 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4398 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4400 2014-04-22 Richard Henderson <rth@redhat.com>
4402 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4403 (add<GPI>3_compare0): Remove leading * from name.
4404 (add<GPI>3_carryin): Likewise.
4405 (sub<GPI>3_compare0): Likewise.
4406 (sub<GPI>3_carryin): Likewise.
4407 (<su_optab>mulditi3): New expander.
4408 (multi3): New expander.
4409 (madd<GPI>): Remove leading * from name.
4411 2014-04-22 Martin Jambor <mjambor@suse.cz>
4413 * cgraphclones.c (cgraph_function_versioning): Copy
4414 ipa_transforms_to_apply instead of asserting it is empty.
4416 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4419 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4420 on count_exp to get mode.
4422 2014-04-22 Andrew Pinski <apinski@cavium.com>
4424 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4425 Handle TLS for ILP32.
4426 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4427 (tlsie_small_<mode>): this and handle PTR.
4428 (tlsie_small_sidi): New pattern.
4429 (tlsle_small): Change to an expand to handle ILP32.
4430 (tlsle_small_<mode>): New pattern.
4431 (tlsdesc_small): Rename to ...
4432 (tlsdesc_small_<mode>): this and handle PTR.
4434 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4436 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4438 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4440 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4441 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4442 (aarch64_types_signed_poly_qualifiers): Likewise.
4443 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4444 (aarch64_types_poly_signed_qualifiers): Likewise.
4445 (TYPES_REINTERP_SS): Type macro added.
4446 (TYPES_REINTERP_SU): Likewise.
4447 (TYPES_REINTERP_SP): Likewise.
4448 (TYPES_REINTERP_US): Likewise.
4449 (TYPES_REINTERP_PS): Likewise.
4450 (aarch64_fold_builtin): New expression folding added.
4451 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4452 Declarations removed.
4453 (REINTERP_SS): Declarations added.
4454 (REINTERP_US): Likewise.
4455 (REINTERP_PS): Likewise.
4456 (REINTERP_SU): Likewise.
4457 (REINTERP_SP): Likewise.
4458 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4459 (vreinterpretq_p8_f64): Likewise.
4460 (vreinterpret_p16_f64): Likewise.
4461 (vreinterpretq_p16_f64): Likewise.
4462 (vreinterpret_f32_f64): Likewise.
4463 (vreinterpretq_f32_f64): Likewise.
4464 (vreinterpret_f64_f32): Likewise.
4465 (vreinterpret_f64_p8): Likewise.
4466 (vreinterpret_f64_p16): Likewise.
4467 (vreinterpret_f64_s8): Likewise.
4468 (vreinterpret_f64_s16): Likewise.
4469 (vreinterpret_f64_s32): Likewise.
4470 (vreinterpret_f64_s64): Likewise.
4471 (vreinterpret_f64_u8): Likewise.
4472 (vreinterpret_f64_u16): Likewise.
4473 (vreinterpret_f64_u32): Likewise.
4474 (vreinterpret_f64_u64): Likewise.
4475 (vreinterpretq_f64_f32): Likewise.
4476 (vreinterpretq_f64_p8): Likewise.
4477 (vreinterpretq_f64_p16): Likewise.
4478 (vreinterpretq_f64_s8): Likewise.
4479 (vreinterpretq_f64_s16): Likewise.
4480 (vreinterpretq_f64_s32): Likewise.
4481 (vreinterpretq_f64_s64): Likewise.
4482 (vreinterpretq_f64_u8): Likewise.
4483 (vreinterpretq_f64_u16): Likewise.
4484 (vreinterpretq_f64_u32): Likewise.
4485 (vreinterpretq_f64_u64): Likewise.
4486 (vreinterpret_s64_f64): Likewise.
4487 (vreinterpretq_s64_f64): Likewise.
4488 (vreinterpret_u64_f64): Likewise.
4489 (vreinterpretq_u64_f64): Likewise.
4490 (vreinterpret_s8_f64): Likewise.
4491 (vreinterpretq_s8_f64): Likewise.
4492 (vreinterpret_s16_f64): Likewise.
4493 (vreinterpretq_s16_f64): Likewise.
4494 (vreinterpret_s32_f64): Likewise.
4495 (vreinterpretq_s32_f64): Likewise.
4496 (vreinterpret_u8_f64): Likewise.
4497 (vreinterpretq_u8_f64): Likewise.
4498 (vreinterpret_u16_f64): Likewise.
4499 (vreinterpretq_u16_f64): Likewise.
4500 (vreinterpret_u32_f64): Likewise.
4501 (vreinterpretq_u32_f64): Likewise.
4503 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4505 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4506 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4507 (vreinterpret_p8_s8): Likewise.
4508 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4509 (vreinterpret_p8_s16): Likewise.
4510 (vreinterpret_p8_s32): Likewise.
4511 (vreinterpret_p8_s64): Likewise.
4512 (vreinterpret_p8_f32): Likewise.
4513 (vreinterpret_p8_u8): Likewise.
4514 (vreinterpret_p8_u16): Likewise.
4515 (vreinterpret_p8_u32): Likewise.
4516 (vreinterpret_p8_u64): Likewise.
4517 (vreinterpret_p8_p16): Likewise.
4518 (vreinterpretq_p8_s8): Likewise.
4519 (vreinterpretq_p8_s16): Likewise.
4520 (vreinterpretq_p8_s32): Likewise.
4521 (vreinterpretq_p8_s64): Likewise.
4522 (vreinterpretq_p8_f32): Likewise.
4523 (vreinterpretq_p8_u8): Likewise.
4524 (vreinterpretq_p8_u16): Likewise.
4525 (vreinterpretq_p8_u32): Likewise.
4526 (vreinterpretq_p8_u64): Likewise.
4527 (vreinterpretq_p8_p16): Likewise.
4528 (vreinterpret_p16_s8): Likewise.
4529 (vreinterpret_p16_s16): Likewise.
4530 (vreinterpret_p16_s32): Likewise.
4531 (vreinterpret_p16_s64): Likewise.
4532 (vreinterpret_p16_f32): Likewise.
4533 (vreinterpret_p16_u8): Likewise.
4534 (vreinterpret_p16_u16): Likewise.
4535 (vreinterpret_p16_u32): Likewise.
4536 (vreinterpret_p16_u64): Likewise.
4537 (vreinterpret_p16_p8): Likewise.
4538 (vreinterpretq_p16_s8): Likewise.
4539 (vreinterpretq_p16_s16): Likewise.
4540 (vreinterpretq_p16_s32): Likewise.
4541 (vreinterpretq_p16_s64): Likewise.
4542 (vreinterpretq_p16_f32): Likewise.
4543 (vreinterpretq_p16_u8): Likewise.
4544 (vreinterpretq_p16_u16): Likewise.
4545 (vreinterpretq_p16_u32): Likewise.
4546 (vreinterpretq_p16_u64): Likewise.
4547 (vreinterpretq_p16_p8): Likewise.
4548 (vreinterpret_f32_s8): Likewise.
4549 (vreinterpret_f32_s16): Likewise.
4550 (vreinterpret_f32_s32): Likewise.
4551 (vreinterpret_f32_s64): Likewise.
4552 (vreinterpret_f32_u8): Likewise.
4553 (vreinterpret_f32_u16): Likewise.
4554 (vreinterpret_f32_u32): Likewise.
4555 (vreinterpret_f32_u64): Likewise.
4556 (vreinterpret_f32_p8): Likewise.
4557 (vreinterpret_f32_p16): Likewise.
4558 (vreinterpretq_f32_s8): Likewise.
4559 (vreinterpretq_f32_s16): Likewise.
4560 (vreinterpretq_f32_s32): Likewise.
4561 (vreinterpretq_f32_s64): Likewise.
4562 (vreinterpretq_f32_u8): Likewise.
4563 (vreinterpretq_f32_u16): Likewise.
4564 (vreinterpretq_f32_u32): Likewise.
4565 (vreinterpretq_f32_u64): Likewise.
4566 (vreinterpretq_f32_p8): Likewise.
4567 (vreinterpretq_f32_p16): Likewise.
4568 (vreinterpret_s64_s8): Likewise.
4569 (vreinterpret_s64_s16): Likewise.
4570 (vreinterpret_s64_s32): Likewise.
4571 (vreinterpret_s64_f32): Likewise.
4572 (vreinterpret_s64_u8): Likewise.
4573 (vreinterpret_s64_u16): Likewise.
4574 (vreinterpret_s64_u32): Likewise.
4575 (vreinterpret_s64_u64): Likewise.
4576 (vreinterpret_s64_p8): Likewise.
4577 (vreinterpret_s64_p16): Likewise.
4578 (vreinterpretq_s64_s8): Likewise.
4579 (vreinterpretq_s64_s16): Likewise.
4580 (vreinterpretq_s64_s32): Likewise.
4581 (vreinterpretq_s64_f32): Likewise.
4582 (vreinterpretq_s64_u8): Likewise.
4583 (vreinterpretq_s64_u16): Likewise.
4584 (vreinterpretq_s64_u32): Likewise.
4585 (vreinterpretq_s64_u64): Likewise.
4586 (vreinterpretq_s64_p8): Likewise.
4587 (vreinterpretq_s64_p16): Likewise.
4588 (vreinterpret_u64_s8): Likewise.
4589 (vreinterpret_u64_s16): Likewise.
4590 (vreinterpret_u64_s32): Likewise.
4591 (vreinterpret_u64_s64): Likewise.
4592 (vreinterpret_u64_f32): Likewise.
4593 (vreinterpret_u64_u8): Likewise.
4594 (vreinterpret_u64_u16): Likewise.
4595 (vreinterpret_u64_u32): Likewise.
4596 (vreinterpret_u64_p8): Likewise.
4597 (vreinterpret_u64_p16): Likewise.
4598 (vreinterpretq_u64_s8): Likewise.
4599 (vreinterpretq_u64_s16): Likewise.
4600 (vreinterpretq_u64_s32): Likewise.
4601 (vreinterpretq_u64_s64): Likewise.
4602 (vreinterpretq_u64_f32): Likewise.
4603 (vreinterpretq_u64_u8): Likewise.
4604 (vreinterpretq_u64_u16): Likewise.
4605 (vreinterpretq_u64_u32): Likewise.
4606 (vreinterpretq_u64_p8): Likewise.
4607 (vreinterpretq_u64_p16): Likewise.
4608 (vreinterpret_s8_s16): Likewise.
4609 (vreinterpret_s8_s32): Likewise.
4610 (vreinterpret_s8_s64): Likewise.
4611 (vreinterpret_s8_f32): Likewise.
4612 (vreinterpret_s8_u8): Likewise.
4613 (vreinterpret_s8_u16): Likewise.
4614 (vreinterpret_s8_u32): Likewise.
4615 (vreinterpret_s8_u64): Likewise.
4616 (vreinterpret_s8_p8): Likewise.
4617 (vreinterpret_s8_p16): Likewise.
4618 (vreinterpretq_s8_s16): Likewise.
4619 (vreinterpretq_s8_s32): Likewise.
4620 (vreinterpretq_s8_s64): Likewise.
4621 (vreinterpretq_s8_f32): Likewise.
4622 (vreinterpretq_s8_u8): Likewise.
4623 (vreinterpretq_s8_u16): Likewise.
4624 (vreinterpretq_s8_u32): Likewise.
4625 (vreinterpretq_s8_u64): Likewise.
4626 (vreinterpretq_s8_p8): Likewise.
4627 (vreinterpretq_s8_p16): Likewise.
4628 (vreinterpret_s16_s8): Likewise.
4629 (vreinterpret_s16_s32): Likewise.
4630 (vreinterpret_s16_s64): Likewise.
4631 (vreinterpret_s16_f32): Likewise.
4632 (vreinterpret_s16_u8): Likewise.
4633 (vreinterpret_s16_u16): Likewise.
4634 (vreinterpret_s16_u32): Likewise.
4635 (vreinterpret_s16_u64): Likewise.
4636 (vreinterpret_s16_p8): Likewise.
4637 (vreinterpret_s16_p16): Likewise.
4638 (vreinterpretq_s16_s8): Likewise.
4639 (vreinterpretq_s16_s32): Likewise.
4640 (vreinterpretq_s16_s64): Likewise.
4641 (vreinterpretq_s16_f32): Likewise.
4642 (vreinterpretq_s16_u8): Likewise.
4643 (vreinterpretq_s16_u16): Likewise.
4644 (vreinterpretq_s16_u32): Likewise.
4645 (vreinterpretq_s16_u64): Likewise.
4646 (vreinterpretq_s16_p8): Likewise.
4647 (vreinterpretq_s16_p16): Likewise.
4648 (vreinterpret_s32_s8): Likewise.
4649 (vreinterpret_s32_s16): Likewise.
4650 (vreinterpret_s32_s64): Likewise.
4651 (vreinterpret_s32_f32): Likewise.
4652 (vreinterpret_s32_u8): Likewise.
4653 (vreinterpret_s32_u16): Likewise.
4654 (vreinterpret_s32_u32): Likewise.
4655 (vreinterpret_s32_u64): Likewise.
4656 (vreinterpret_s32_p8): Likewise.
4657 (vreinterpret_s32_p16): Likewise.
4658 (vreinterpretq_s32_s8): Likewise.
4659 (vreinterpretq_s32_s16): Likewise.
4660 (vreinterpretq_s32_s64): Likewise.
4661 (vreinterpretq_s32_f32): Likewise.
4662 (vreinterpretq_s32_u8): Likewise.
4663 (vreinterpretq_s32_u16): Likewise.
4664 (vreinterpretq_s32_u32): Likewise.
4665 (vreinterpretq_s32_u64): Likewise.
4666 (vreinterpretq_s32_p8): Likewise.
4667 (vreinterpretq_s32_p16): Likewise.
4668 (vreinterpret_u8_s8): Likewise.
4669 (vreinterpret_u8_s16): Likewise.
4670 (vreinterpret_u8_s32): Likewise.
4671 (vreinterpret_u8_s64): Likewise.
4672 (vreinterpret_u8_f32): Likewise.
4673 (vreinterpret_u8_u16): Likewise.
4674 (vreinterpret_u8_u32): Likewise.
4675 (vreinterpret_u8_u64): Likewise.
4676 (vreinterpret_u8_p8): Likewise.
4677 (vreinterpret_u8_p16): Likewise.
4678 (vreinterpretq_u8_s8): Likewise.
4679 (vreinterpretq_u8_s16): Likewise.
4680 (vreinterpretq_u8_s32): Likewise.
4681 (vreinterpretq_u8_s64): Likewise.
4682 (vreinterpretq_u8_f32): Likewise.
4683 (vreinterpretq_u8_u16): Likewise.
4684 (vreinterpretq_u8_u32): Likewise.
4685 (vreinterpretq_u8_u64): Likewise.
4686 (vreinterpretq_u8_p8): Likewise.
4687 (vreinterpretq_u8_p16): Likewise.
4688 (vreinterpret_u16_s8): Likewise.
4689 (vreinterpret_u16_s16): Likewise.
4690 (vreinterpret_u16_s32): Likewise.
4691 (vreinterpret_u16_s64): Likewise.
4692 (vreinterpret_u16_f32): Likewise.
4693 (vreinterpret_u16_u8): Likewise.
4694 (vreinterpret_u16_u32): Likewise.
4695 (vreinterpret_u16_u64): Likewise.
4696 (vreinterpret_u16_p8): Likewise.
4697 (vreinterpret_u16_p16): Likewise.
4698 (vreinterpretq_u16_s8): Likewise.
4699 (vreinterpretq_u16_s16): Likewise.
4700 (vreinterpretq_u16_s32): Likewise.
4701 (vreinterpretq_u16_s64): Likewise.
4702 (vreinterpretq_u16_f32): Likewise.
4703 (vreinterpretq_u16_u8): Likewise.
4704 (vreinterpretq_u16_u32): Likewise.
4705 (vreinterpretq_u16_u64): Likewise.
4706 (vreinterpretq_u16_p8): Likewise.
4707 (vreinterpretq_u16_p16): Likewise.
4708 (vreinterpret_u32_s8): Likewise.
4709 (vreinterpret_u32_s16): Likewise.
4710 (vreinterpret_u32_s32): Likewise.
4711 (vreinterpret_u32_s64): Likewise.
4712 (vreinterpret_u32_f32): Likewise.
4713 (vreinterpret_u32_u8): Likewise.
4714 (vreinterpret_u32_u16): Likewise.
4715 (vreinterpret_u32_u64): Likewise.
4716 (vreinterpret_u32_p8): Likewise.
4717 (vreinterpret_u32_p16): Likewise.
4718 (vreinterpretq_u32_s8): Likewise.
4719 (vreinterpretq_u32_s16): Likewise.
4720 (vreinterpretq_u32_s32): Likewise.
4721 (vreinterpretq_u32_s64): Likewise.
4722 (vreinterpretq_u32_f32): Likewise.
4723 (vreinterpretq_u32_u8): Likewise.
4724 (vreinterpretq_u32_u16): Likewise.
4725 (vreinterpretq_u32_u64): Likewise.
4726 (vreinterpretq_u32_p8): Likewise.
4727 (vreinterpretq_u32_p16): Likewise.
4729 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4731 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4733 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4735 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4736 (vqnegd_s64): Likewise.
4737 (vqabs_s64): Likewise.
4738 (vqabsd_s64): Likewise.
4740 2014-04-22 Richard Henderson <rth@redhat.com>
4742 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4743 computation to the top of the loop.
4745 2014-04-22 Renlin <renlin.li@arm.com>
4746 Jiong Wang <jiong.wang@arm.com>
4748 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4749 * config/aarch64/aarch64.c (aarch64_layout_frame)
4750 (aarch64_initial_elimination_offset): Likewise.
4752 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4754 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4757 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4759 * machmode.h (bitwise_mode_for_mode): Declare.
4760 * stor-layout.h (bitwise_type_for_mode): Likewise.
4761 * stor-layout.c (bitwise_mode_for_mode): New function.
4762 (bitwise_type_for_mode): Likewise.
4763 * builtins.c (fold_builtin_memory_op): Use it instead of
4764 int_mode_for_mode and build_nonstandard_integer_type.
4766 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4768 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4769 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4770 (*-*-solaris2*): Simplify.
4771 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4772 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4773 *-*-solaris2.9* handling.
4775 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4777 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4778 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4780 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4781 * configure: Regenerate.
4783 * config/i386/sol2-9.h: Remove.
4785 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4786 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4787 Remove Solaris 9 references.
4789 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4791 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4792 (floatuns<GPI:mode><GPF:mode>2): Remove.
4793 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4794 and floatuns conversions.
4795 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4796 and floatuns conversions.
4797 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4798 (w1,w2): New mode attributes for inequal width conversions.
4800 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4802 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4803 the output asm format.
4805 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4807 * config/aarch64/aarch64-simd.md
4808 (aarch64_cm<optab>di): Always split.
4809 (*aarch64_cm<optab>di): New.
4810 (aarch64_cmtstdi): Always split.
4811 (*aarch64_cmtstdi): New.
4813 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4815 PR tree-optimization/60823
4816 * omp-low.c (ipa_simd_modify_function_body): Go through
4817 all SSA_NAMEs and for those refering to vector arguments
4818 which are going to be replaced adjust SSA_NAME_VAR and,
4819 if it is a default definition, change it into a non-default
4820 definition assigned at the beginning of function from new_decl.
4821 (ipa_simd_modify_stmt_ops): Rewritten.
4822 * tree-dfa.c (set_ssa_default_def): When removing default def,
4823 check for NULL loc instead of NULL *loc.
4825 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4827 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4828 restrictions on core registers for DImode values in Thumb2.
4830 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4832 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4833 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4835 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4837 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4838 (*iordi_notzesidi_di): Likewise.
4839 (*iordi_notsesidi_di): Likewise.
4841 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4843 * config/arm/arm-protos.h (tune_params): New struct members.
4844 * config/arm/arm.c: Initialise tune_params per processor.
4845 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4846 for speed, based on new tune_params.
4848 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4850 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4851 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4852 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4853 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4854 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4855 (vrnda_f64): Likewise.
4856 (vrndi_f64): Likewise.
4857 (vrndm_f64): Likewise.
4858 (vrndn_f64): Likewise.
4859 (vrndp_f64): Likewise.
4860 (vrndx_f64): Likewise.
4862 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4864 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4865 GET_MODE_SIZE argument is enum machine_mode.
4867 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4870 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4871 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4873 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4876 * asan.c (asan_emit_stack_protection): Force the base to align to
4877 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4878 appropriate bits if STRICT_ALIGNMENT.
4879 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4881 (expand_used_vars): Leave a space in the stack frame for alignment
4882 if STRICT_ALIGNMENT.
4884 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4886 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4888 (gimple_store_p): Likewise.
4889 (gimple_assign_load_p): Likewise.
4890 (gimple_assign_cast_p): Likewise.
4891 (gimple_clobber_p): Likewise.
4893 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4894 rather than a gimple.
4895 (gimple_assign_cast_p): Likewise.
4897 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4900 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4901 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4903 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4904 more debug information for E500 if -mdebug=reg.
4906 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4909 * config/i386/i386.c (ix86_expand_builtin)
4910 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4911 register for target RTX.
4912 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4914 2014-04-18 Cong Hou <congh@google.com>
4916 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4917 the widen-mult pattern by handling two operands with different sizes,
4918 and operands whose size is smaller than half of the result type.
4920 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4922 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4923 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4924 (do_estimate_edge_time): Compute it.
4925 * ipa-inline.c (want_inline_small_function_p): Bypass
4926 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4928 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4930 * ipa-inline.c (spec_rem): New static variable.
4931 (dump_overall_stats): New function.
4932 (dump_inline_stats): New function.
4934 2014-04-18 Richard Henderson <rth@redhat.com>
4936 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4937 to GET_MODE_SIZE, not a reg_class_t.
4939 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4941 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4942 (vsx_xxmrglw_<mode>): Likewise.
4944 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4947 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4948 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4949 (rs6000_init_hard_regno_mode_ok): Likewise.
4951 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4953 * ipa-inline.c (inline_small_functions): Account only non-cold
4955 * doc/invoke.texi (inline-unit-growth): Update documentation.
4957 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4959 * config/rs6000/rs6000.md (addti3, subti3): New.
4961 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4964 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4965 comment. Check optimize_insn_for_size_p instead of
4966 optimize_insn_for_speed_p.
4968 2014-04-17 Martin Jambor <mjambor@suse.cz>
4970 * gimple-iterator.c (gsi_start_edge): New function.
4971 * gimple-iterator.h (gsi_start_edge): Declare.
4972 * tree-sra.c (single_non_eh_succ): New function.
4973 (disqualify_ops_if_throwing_stmt): Renamed to
4974 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4975 having one non-EH successor BB.
4976 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4977 generate loads into replacements.
4978 (sra_modify_assign): Likewise and and also use the simple path for
4980 (sra_modify_function_body): Commit statements on edges.
4982 2014-04-17 Richard Biener <rguenther@suse.de>
4985 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4986 comparison results and add clarifying comment.
4988 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4990 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4991 (blank_mode): Initialize it.
4992 (emit_mode_size_inline, emit_mode_nunits_inline,
4993 emit_mode_inner_inline): New functions.
4994 (emit_insn_modes_h): Call them and surround their output with
4995 #if GCC_VERSION >= 4001 ... #endif.
4996 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4997 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4998 mode_* arrays if the argument is __builtin_constant_p.
4999 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5000 is enum machine_mode.
5002 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5004 * passes.c (opt_pass::execute): Adjust.
5005 (pass_manager::execute_pass_mode_switching): Likewise.
5006 (early_local_passes::execute): Likewise.
5007 (execute_one_pass): Pass cfun to the pass's execute method.
5008 * tree-pass.h (opt_pass::execute): Add function * argument.
5009 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5010 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5011 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5012 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5013 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5014 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5015 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5016 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5017 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5018 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5019 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5020 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5021 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5022 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5023 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5024 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5025 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5026 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5027 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5028 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5029 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5030 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5031 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5032 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5033 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5034 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5035 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5036 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5037 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5040 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5042 * passes.c (opt_pass::gate): Take function * argument.
5043 (gate_all_early_local_passes): Merge into
5044 (early_local_passes::gate): this.
5045 (gate_all_early_optimizations): Merge into
5046 (all_early_optimizations::gate): this.
5047 (gate_all_optimizations): Mege into
5048 (all_optimizations::gate): this.
5049 (gate_all_optimizations_g): Merge into
5050 (all_optimizations_g::gate): this.
5051 (gate_rest_of_compilation): Mege into
5052 (rest_of_compilation::gate): this.
5053 (gate_postreload): Merge into
5054 (postreload::gate): this.
5055 (dump_one_pass): Pass cfun to the pass's gate method.
5056 (execute_ipa_summary_passes): Likewise.
5057 (execute_one_pass): Likewise.
5058 (ipa_write_summaries_2): Likewise.
5059 (ipa_write_optimization_summaries_1): Likewise.
5060 (ipa_read_summaries_1): Likewise.
5061 (ipa_read_optimization_summaries_1): Likewise.
5062 (execute_ipa_stmt_fixups): Likewise.
5063 * tree-pass.h (opt_pass::gate): Add function * argument.
5064 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5065 combine-stack-adj.c, combine.c, compare-elim.c,
5066 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5067 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5068 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5069 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5070 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5071 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5072 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5073 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5074 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5075 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5076 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5077 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5078 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5079 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5080 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5081 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5082 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5083 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5084 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5085 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5086 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5087 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5088 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5089 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5090 var-tracking.c, vtable-verify.c, web.c: Adjust.
5092 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5094 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5095 * configure: Regenerate.
5097 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5099 * passes.c (dump_one_pass): don't check pass->has_gate.
5100 (execute_ipa_summary_passes): Likewise.
5101 (execute_one_pass): Likewise.
5102 (ipa_write_summaries_2): Likewise.
5103 (ipa_write_optimization_summaries_1): Likewise.
5104 (ipa_read_optimization_summaries_1): Likewise.
5105 (execute_ipa_stmt_fixups): Likewise.
5106 * tree-pass.h (pass_data::has_gate): Remove.
5107 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5108 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5109 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5110 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5111 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5112 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5113 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5114 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5115 gimple-low.c, gimple-ssa-isolate-paths.c,
5116 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5117 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5118 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5119 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5120 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5121 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5122 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5123 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5124 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5125 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5126 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5127 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5128 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5129 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5130 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5131 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5132 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5133 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5134 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5135 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5136 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5137 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5140 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5142 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5144 * passes.c (pass_manager::register_dump_files_1): Merge into
5145 (pass_manager::register_dump_files): this, and remove its handling of
5146 properties since the pass always has the properties anyway.
5147 (pass_manager::pass_manager): Adjust.
5149 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5151 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5152 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5153 dealing with properties.
5154 (pass_manager::register_dump_files): Adjust.
5156 2014-03-20 Mark Wielaard <mjw@redhat.com>
5158 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5159 then represent the bound as normal constant value.
5161 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5164 Forward port from 4.8 branch
5165 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5167 * config/i386/bmiintrin.h (_blsi_u32): New.
5171 (_blsmsk_u32): Ditto.
5172 (_blsmsk_u64): Ditto.
5173 (_tzcnt_u32): Ditto.
5174 (_tzcnt_u64): Ditto.
5176 2014-04-17 Kito Cheng <kito@0xlab.org>
5178 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5180 2014-04-17 Richard Biener <rguenther@suse.de>
5183 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5184 boolean results for comparisons.
5186 2014-04-17 Richard Biener <rguenther@suse.de>
5188 PR tree-optimization/60836
5189 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5190 initial PHI args to be gimple values.
5192 2014-04-17 Richard Biener <rguenther@suse.de>
5194 PR tree-optimization/60841
5195 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5196 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5197 of stmts to SLP build.
5198 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5199 (vect_analyze_slp): Likewise.
5200 (vect_analyze_slp_instance): Likewise.
5201 (vect_build_slp_tree): Limit overall SLP tree growth.
5202 * tree-vectorizer.h (vect_analyze_data_refs,
5203 vect_analyze_slp): Adjust prototypes.
5205 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5207 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5210 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5212 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5213 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5214 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5215 for TARGET_SLOW_PSHUFB
5217 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5219 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5220 * config/i386/i386.c (intel_cost): Ditto.
5222 2014-04-17 Joey Ye <joey.ye@arm.com>
5224 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5226 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5228 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5231 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5233 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5234 (type_all_derivations_known_p): New predicate.
5235 (type_all_ctors_visible_p): New predicate.
5236 (type_possibly_instantiated_p): New predicate.
5237 (get_odr_type): Compute all_derivations_known.
5238 (dump_odr_type): Dump the flag.
5239 (maybe_record_type): Cleanup.
5240 (record_target_from_binfo): Add bases_to_consider array;
5241 record bases for types w/o instances and skip CXX destructor.
5242 (possible_polymorphic_call_targets_1): Add bases_to_consider
5243 and consider_construction parameters; check if type may have instance.
5244 (get_polymorphic_call_info): Set maybe_in_construction to true
5245 when we know nothing.
5246 (record_targets_from_bases): Skip CXX destructors; they are
5247 never called for types in construction.
5248 (possible_polymorphic_call_targets): Do not record target when
5249 type may not have instance.
5251 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5254 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5255 external aliases alive, too.
5257 2014-04-16 Andrew Pinski <apinski@cavium.com>
5259 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5262 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5264 * final.c (compute_alignments): Do not apply loop alignment to a block
5265 falling through to the exit.
5267 2014-04-16 Catherine Moore <clm@codesourcery.com>
5269 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5270 Adjust constraints for microMIPS store patterns.
5272 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5274 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5276 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5278 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5279 (append_use): Run at -O0.
5280 (append_vdef): Likewise.
5281 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5282 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5284 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5286 PR tree-optimization/60844
5287 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5288 (propagate_op_to_single_use, remove_visited_stmt_chain,
5289 linearize_expr, repropagate_negates, reassociate_bb): Use it
5290 instead of gsi_remove.
5292 2014-04-16 Martin Jambor <mjambor@suse.cz>
5294 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5295 ipa_transforms_to_apply.
5296 (cgraph_function_versioning): Assert that old_node has empty
5297 ipa_transforms_to_apply.
5298 * trans-mem.c (ipa_tm_create_version): Likewise.
5299 * tree-inline.c (tree_function_versioning): Do not duplicate
5300 ipa_transforms_to_apply.
5302 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5305 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5307 Pass necessary as flags on 64-bit Solaris/x86.
5308 Use lowercase relocs for x86_64-*-*.
5309 * configure: Regenerate.
5311 2014-04-15 Jan Hubicka <jh@suse.cz>
5313 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5314 (maybe_record_node, likely_target_p): Use it.
5316 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5319 Revert following patch
5321 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5324 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5325 software floating point or no floating point registers, do not
5326 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5327 in GPRs that occurs after we tested for GPRs that would never be
5330 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5331 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5332 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5333 specifically allow DDmode, since that does not use the SPE SIMD
5336 2014-03-21 Mark Wielaard <mjw@redhat.com>
5338 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5339 as unsigned or int depending on type and value used.
5341 2014-04-15 Richard Biener <rguenther@suse.de>
5343 PR rtl-optimization/56965
5344 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5345 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5347 * alias.c (true_dependence_1): Do not call
5348 nonoverlapping_component_refs_p.
5349 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5350 nonoverlapping_component_refs_p.
5351 (indirect_refs_may_alias_p): Likewise.
5353 2014-04-15 Teresa Johnson <tejohnson@google.com>
5355 * cfg.c (dump_bb_info): Fix flags check.
5356 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5358 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5360 PR rtl-optimization/60663
5361 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5364 2014-04-15 Richard Biener <rguenther@suse.de>
5366 * lto-streamer.h (LTO_major_version): Bump to 4.
5368 2014-04-15 Richard Biener <rguenther@suse.de>
5370 * common.opt (lto_partition_model): New enum.
5371 (flto-partition=): Merge separate options with a single with argument,
5372 add -flto-partition=one support.
5373 * flag-types.h (enum lto_partition_model): Declare.
5374 * opts.c (finish_options): Remove duplicate -flto-partition=
5376 * lto-wrapper.c (run_gcc): Adjust.
5378 2014-04-15 Richard Biener <rguenther@suse.de>
5380 * alias.c (ncr_compar): New function.
5381 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5383 2014-04-15 Richard Biener <rguenther@suse.de>
5385 * alias.c (record_component_aliases): Do not walk BINFOs.
5387 2014-04-15 Richard Biener <rguenther@suse.de>
5389 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5390 Add struct function argument and adjust.
5391 (find_func_aliases_for_call): Likewise.
5392 (find_func_aliases): Likewise.
5393 (find_func_clobbers): Likewise.
5394 (intra_create_variable_infos): Likewise.
5395 (compute_points_to_sets): Likewise.
5396 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5398 2014-04-15 Richard Biener <rguenther@suse.de>
5400 * tree.c (iterative_hash_expr): Use enum tree_code_class
5401 to store TREE_CODE_CLASS.
5402 (tree_block): Likewise.
5403 (tree_set_block): Likewise.
5404 * tree.h (fold_build_pointer_plus_loc): Use
5405 convert_to_ptrofftype_loc.
5407 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5410 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5413 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5415 * cfgloop.h (struct loop): Move force_vectorize down.
5416 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5417 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5418 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5419 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5420 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5421 * tree-core.h (enum annot_expr_kind): Add new kind values.
5422 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5423 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5425 * tree.def (ANNOTATE_EXPR): Tweak comment.
5427 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5429 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5432 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5434 * tree.h (TYPE_IDENTIFIER): Declare.
5435 * tree.c (subrange_type_for_debug_p): Use it.
5436 * godump.c (go_format_type): Likewise.
5437 * dwarf2out.c (is_cxx_auto, modified_type_die,
5438 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5439 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5441 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5444 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5446 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5448 * config/i386/i386.c (examine_argument): Return bool. Return true if
5449 parameter should be passed in memory.
5450 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5451 (construct_container): Update calls to examine_argument.
5452 (function_arg_advance_64): Ditto.
5453 (return_in_memory_32): Merge with ix86_return_in_memory.
5454 (return_in_memory_64): Ditto.
5455 (return_in_memory_ms_64): Ditto.
5457 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5459 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5460 * coverage.c (coverage_compute_profile_id): Handle externally visible
5463 2014-04-14 Martin Jambor <mjambor@suse.cz>
5465 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5466 DECL_DISREGARD_INLINE_LIMITS functions.
5468 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5471 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5473 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5476 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5477 optimize_insn_for_speed_p instead of
5478 optimize_function_for_speed_p.
5480 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5482 * doc/invoke.texi (free): Document AArch64.
5484 2014-04-14 Richard Biener <rguenther@suse.de>
5486 PR tree-optimization/60042
5487 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5488 (insert_into_preds_of_block): Do not prevent PHI insertion
5489 for REFERENCE exprs here ...
5490 (eliminate_dom_walker::before_dom_children): ... but prevent
5491 their use here under similar conditions when applied to the
5492 IL after PRE optimizations.
5494 2014-04-14 Richard Biener <rguenther@suse.de>
5496 * passes.def: Move early points-to after early SRA.
5498 2014-04-14 Richard Biener <rguenther@suse.de>
5500 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5501 check for which sign-changes we allow when forwarding
5502 a converted value into a switch.
5504 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5506 * stor-layout.c (place_field): Finalize non-constant offset for the
5509 2014-04-14 Richard Biener <rguenther@suse.de>
5511 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5513 (expand_switch_using_bit_tests_p): Likewise.
5514 (process_switch): Compute and pass on speed_p based on the
5516 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5517 optimize_bb_for_speed_p.
5519 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5521 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5522 * function.h (struct function): Rename has_force_vect_loops into
5523 has_force_vectorize_loops.
5524 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5525 (input_struct_function_base): Likewise.
5526 * lto-streamer-out.c (output_cfg): Likewise.
5527 (output_struct_function_base): Likewise.
5528 * omp-low.c (expand_omp_simd): Likewise.
5529 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5530 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5531 (version_loop_for_if_conversion): Likewise.
5532 (tree_if_conversion): Likewise.
5533 (main_tree_if_conversion): Likewise.
5534 (gate_tree_if_conversion): Likewise.
5535 * tree-inline.c (copy_loops): Likewise.
5536 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5537 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5538 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5539 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5540 * tree-vectorizer.c (vectorize_loops): Likewise.
5541 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5543 2014-04-14 Richard Biener <rguenther@suse.de>
5546 * lto-streamer-out.c (wrap_refs): New function.
5547 (lto_output): Wrap symbol references in global initializes in
5548 type-preserving MEM_REFs.
5550 2014-04-14 Christian Bruel <christian.bruel@st.com>
5552 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5554 2014-04-14 Christian Bruel <christian.bruel@st.com>
5556 * config/sh/sh.md (setmemqi): New expand pattern.
5557 * config/sh/sh.h (CLEAR_RATIO): Define.
5558 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5559 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5561 2014-04-14 Richard Biener <rguenther@suse.de>
5564 * fold-const.c (negate_expr_p): Don't negate directional rounding
5566 (fold_negate_expr): Likewise.
5568 2014-04-14 Richard Biener <rguenther@suse.de>
5570 PR tree-optimization/59817
5571 PR tree-optimization/60453
5572 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5573 recursion to catch all CHRECs in the scalar evolution and restrict
5574 the predicate for the remains appropriately.
5576 2014-04-12 Catherine Moore <clm@codesourcery.com>
5578 * config/mips/constraints.md: Add new register constraint "kb".
5579 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5580 (*movhi_internal): Likewise.
5581 (*movqi_internal): Likewise.
5582 * config/mips/mips.h (M16_STORE_REGS): New register class.
5583 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5584 (REG_CLASS_CONTENTS): Likewise.
5585 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5587 2014-04-11 Tobias Burnus <burnus@net-b.de>
5590 * doc/invoke.texi (-Wformat-signedness): Document it.
5591 (Wformat=2): Mention that this enables -Wformat-signedness.
5593 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5595 * common/config/epiphany/epiphany-common.c
5596 (epiphany_option_optimization_table): Enable section anchors by
5597 default at -O1 or higher.
5598 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5599 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5600 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5601 carries no extra cost.
5602 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5603 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5604 * config/epiphany/predicates.md (memclob_operand): New predicate.
5605 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5606 Use memclob_operand predicate and X constraint for operand 3.
5608 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5610 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5611 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5614 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5616 PR rtl-optimization/60651
5617 * mode-switching.c (optimize_mode_switching): Make sure to emit
5618 sets of a lower numbered entity before sets of a higher numbered
5619 entity to a mode of the same or lower priority.
5620 When creating a seginfo for a basic block that starts with a code
5621 label, move the insertion point past the code label.
5622 (new_seginfo): Document and enforce requirement that
5623 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5624 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5625 * doc/tm.texi: Regenerate.
5627 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5630 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5632 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5634 * BASE-VER: Set to 4.10.0.
5636 2014-04-11 Tobias Burnus <burnus@net-b.de>
5639 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5640 * doc/gcc.texi (Service): Update description in the @menu
5641 * doc/invoke.texi (Option Summary): Remove misplaced and
5644 2014-04-11 Steve Ellcey <sellcey@mips.com>
5645 Jakub Jelinek <jakub@redhat.com>
5648 * expr.c (convert_move): Use emit_store_flag_force instead of
5649 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5652 2014-04-11 Richard Biener <rguenther@suse.de>
5655 * varasm.c (assemble_alias): Avoid endless error reporting
5656 recursion by setting TREE_ASM_WRITTEN.
5658 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5660 * config/s390/s390.md: Add a splitter for NOT rtx.
5662 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5664 PR rtl-optimization/60663
5665 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5667 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5668 Jakub Jelinek <jakub@redhat.com>
5671 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5672 flag from decl_node to node.
5674 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5677 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5678 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5679 ameliorating the cases where it can be.
5681 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5684 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5686 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5687 (loadsync_<mode>): Change mode.
5688 (load_quadpti, store_quadpti): New.
5689 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5690 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5691 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5693 2014-04-09 Cong Hou <congh@google.com>
5696 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5699 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5701 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5702 instead of vnor to exploit possible fusion opportunity in the
5704 (altivec_expand_vec_perm_const_le): Likewise.
5706 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5708 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5709 (loadsync_<mode>): Change mode.
5710 (load_quadpti, store_quadpti): New.
5711 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5712 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5714 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5717 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5718 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5719 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5721 2014-04-08 Richard Biener <rguenther@suse.de>
5724 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5725 a 64bit widest int print double-int similar to on HWI64 hosts.
5727 2014-04-08 Richard Biener <rguenther@suse.de>
5729 PR tree-optimization/60785
5730 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5731 default defs properly.
5733 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5735 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5736 (Weffc++): Likewise.
5738 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5740 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5741 set completep to false rather than true.
5743 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5746 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5747 ARM_TARGET2_DWARF_FORMAT.
5749 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5752 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5753 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5756 2014-04-07 Richard Biener <rguenther@suse.de>
5758 PR tree-optimization/60766
5759 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5760 (may_eliminate_iv): Convert cand_value_at result to desired type.
5762 2014-04-07 Jason Merrill <jason@redhat.com>
5765 * common.opt (-fno-gnu-unique): Add.
5766 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5768 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5770 * haifa-sched.c: Fix outdated function reference and minor
5771 grammar errors in introductory comment.
5773 2014-04-07 Richard Biener <rguenther@suse.de>
5776 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5778 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5780 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5783 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5784 size accounting for thunks.
5785 (pa_asm_output_mi_thunk): Use final_start_function() and
5786 final_end_function() to output function start and end directives.
5788 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5790 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5791 device specific ISA/ feature information. Remove short_sp and
5792 errata_skip ds. Add avr_device_specific_features enum to have device
5794 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5795 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5796 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5797 updated device specific info.
5798 * config/avr/avr-mcus.def: Merge device specific details to
5799 dev_attribute field.
5800 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5802 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5803 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5804 assembler if RMW isa supported by current device.
5805 * config/avr/genmultilib.awk: Update as device info structure changed.
5806 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5808 2014-04-04 Cong Hou <congh@google.com>
5810 PR tree-optimization/60656
5811 * tree-vect-stmts.c (supportable_widening_operation):
5812 Fix a bug that elements in a vector with vect_used_by_reduction
5813 property are incorrectly reordered when the operation on it is not
5814 consistant with the one in reduction operation.
5816 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5818 PR rtl-optimization/60155
5819 * gcse.c (record_set_data): New function.
5820 (single_set_gcse): New function.
5821 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5822 (hoist_code): Likewise.
5823 (get_pressure_class_and_nregs): Likewise.
5825 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5827 * explow.c (probe_stack_range): Emit a final optimization blockage.
5829 2014-04-04 Anthony Green <green@moxielogic.com>
5831 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5834 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5837 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5838 flags are set only during streaming.
5839 * ipa.c (process_references, walk_polymorphic_call_targets,
5840 symtab_remove_unreachable_nodes): Drop bodies of always inline
5841 after early inlining.
5842 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5844 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5845 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5848 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5851 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5854 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5856 (cortex_a53_fdivd): Likewise.
5858 2014-04-04 Martin Jambor <mjambor@suse.cz>
5861 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5863 * cgraph.c (clone_of_p): Also return true if thunks match.
5864 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5865 cgraph_function_or_thunk_node and an obsolete comment.
5866 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5868 (build_function_decl_skip_args): Likewise.
5869 (set_new_clone_decl_and_node_flags): New function.
5870 (duplicate_thunk_for_node): Likewise.
5871 (redirect_edge_duplicating_thunks): Likewise.
5872 (cgraph_clone_node): New parameter args_to_skip, pass it to
5873 redirect_edge_duplicating_thunks which is called instead of
5874 cgraph_redirect_edge_callee.
5875 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5876 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5878 2014-04-04 Jeff Law <law@redhat.com>
5881 * config/arm/predicates.md (const_int_I_operand): New predicate.
5882 (const_int_M_operand): Similarly.
5883 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5885 (insv_t2, extv_reg, extzv_t2): Likewise.
5886 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5887 (pop_multiple_with_writeback_and_return): Likewise.
5888 (vfp_pop_multiple_with_writeback): Likewise
5890 2014-04-04 Richard Biener <rguenther@suse.de>
5893 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5894 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5896 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5897 * gimplify.c (gimple_add_tmp_var_fn): New function.
5898 * gimple-expr.h (create_tmp_reg_fn): Declare.
5899 * gimple-expr.c (create_tmp_reg_fn): New function.
5900 * gimple-low.c (record_vars_into): Don't change cfun.
5901 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5902 code generation without cfun.
5904 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5907 * Makefile.in (install-driver): Fix shell scripting.
5909 2014-04-03 Cong Hou <congh@google.com>
5911 PR tree-optimization/60505
5912 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5913 threshold of number of iterations below which no vectorization
5915 * tree-vect-loop.c (new_loop_vec_info):
5916 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5917 * tree-vect-loop.c (vect_analyze_loop_operations):
5918 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5919 * tree-vect-loop.c (vect_transform_loop):
5920 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5921 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5922 of iterations of the loop and see if we should build the epilogue.
5924 2014-04-03 Richard Biener <rguenther@suse.de>
5926 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5927 (streamer_tree_cache_create): Adjust.
5928 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5929 to allow optional nodes array.
5930 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5931 (streamer_tree_cache_append): Likewise.
5932 (streamer_tree_cache_create): Create nodes array optionally
5933 as specified by parameter.
5934 * lto-streamer-out.c (create_output_block): Avoid maintaining
5935 the node array in the writer cache.
5936 (DFS_write_tree): Remove assertion.
5937 (produce_asm_for_decls): Free the out decl state hash table early.
5938 * lto-streamer-in.c (lto_data_in_create): Adjust for
5939 streamer_tree_cache_create prototype change.
5941 2014-04-03 Richard Biener <rguenther@suse.de>
5943 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5944 set TREE_CHAIN to NULL_TREE.
5946 2014-04-03 Richard Biener <rguenther@suse.de>
5948 PR tree-optimization/60740
5949 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5950 over all GIMPLE_COND operands.
5952 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5954 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5955 (Weffc++): Remove Scott's numbering, merge lists and reference
5958 2014-04-03 Nick Clifton <nickc@redhat.com>
5960 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5963 2014-04-03 Martin Jambor <mjambor@suse.cz>
5965 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5966 mention gcc_unreachable before failing.
5967 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5970 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5973 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5974 inconsistent code and instead mark the context inconsistent.
5975 (possible_polymorphic_call_targets): For inconsistent contexts
5976 return empty complete list.
5978 2014-04-02 Anthony Green <green@moxielogic.com>
5980 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5981 (extendqisi2, extendhisi2): Define.
5982 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5983 (WCHAR_TYPE): Change to unsigned int.
5985 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5987 PR tree-optimization/60733
5988 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5989 insertion point for PHI candidates to be the end of the feeding
5990 block for the PHI argument.
5992 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5994 PR rtl-optimization/60650
5995 * lra-constraints.c (process_alt_operands): Decrease reject for
5996 earlyclobber matching.
5998 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6000 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6002 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6004 * config/spu/spu.c (pad_bb): Do not crash when the last
6005 insn is CODE_FOR_blockage.
6007 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6009 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6010 lies outside the target mode.
6012 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6015 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6016 software floating point or no floating point registers, do not
6017 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6018 in GPRs that occurs after we tested for GPRs that would never be
6021 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6022 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6023 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6024 specifically allow DDmode, since that does not use the SPE SIMD
6027 2014-04-02 Richard Biener <rguenther@suse.de>
6030 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6031 MODE_INTs. Properly use negv_optab.
6032 (expand_abs): Likewise.
6034 2014-04-02 Richard Biener <rguenther@suse.de>
6037 * Makefile.in (install-driver): Guard extra installs with special
6040 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6042 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6045 2014-04-01 Richard Henderson <rth@redhat.com>
6048 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6049 alternative enabled before register allocation.
6051 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6053 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6054 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6056 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6057 (nios2_got_address): Update nios2_large_got_address call site.
6058 (nios2_delegitimize_address): New function.
6059 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6060 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6061 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6063 2014-04-01 Martin Husemann <martin@duskware.de>
6065 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6068 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6070 PR rtl-optimization/60604
6071 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6072 check from register_operand.
6073 (register_operand): Redefine in terms of general_operand.
6074 (nonmemory_operand): Use register_operand for the non-constant cases.
6076 2014-04-01 Richard Biener <rguenther@suse.de>
6078 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6080 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6082 * doc/invoke.texi (mapp-regs): Clarify.
6084 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6086 * config/i386/avx512fintrin.h (__v32hi): Define type.
6087 (__v64qi): Likewise.
6088 (_mm512_set1_epi8): Define.
6089 (_mm512_set1_epi16): Define.
6090 (_mm512_set4_epi32): Define.
6091 (_mm512_set4_epi64): Define.
6092 (_mm512_set4_pd): Define.
6093 (_mm512_set4_ps): Define.
6094 (_mm512_setr4_epi64): Define.
6095 (_mm512_setr4_epi32): Define.
6096 (_mm512_setr4_pd): Define.
6097 (_mm512_setr4_ps): Define.
6098 (_mm512_setzero_epi32): Define.
6100 2014-03-31 Martin Jambor <mjambor@suse.cz>
6103 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6104 callsite_arguments_match_p. Updated all callers. Also check types of
6105 corresponding formal parameters and actual arguments.
6106 (not_all_callers_have_enough_arguments_p) Renamed to
6107 some_callers_have_mismatched_arguments_p.
6109 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6111 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6113 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6116 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6119 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6121 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6122 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6124 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6125 Use FMAMODE_NOVF512 mode iterator.
6126 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6127 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6128 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6129 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6131 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6132 Use VF_128_256 mode iterator.
6133 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6136 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6138 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6139 static chain if needed.
6141 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6144 * lra-constraints.c (index_part_to_reg): New.
6145 (process_address): Use it.
6147 2014-03-27 Jeff Law <law@redhat.com>
6148 Jakub Jelinek <jakub@redhat.com>
6151 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6152 gen_rtx_{PLUS,MULT} to build up the address expression.
6154 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6155 creating non-canonical RTL.
6157 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6160 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6161 functions; reorganize to make cheap checks first.
6162 (inline_small_functions): Do not estimate growth when dumping;
6164 * ipa-inline.h (inline_summary): Add min_size.
6165 (growth_likely_positive): New function.
6166 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6167 (set_cond_stmt_execution_predicate): Cleanup.
6168 (estimate_edge_size_and_time): Compute min_size.
6169 (estimate_calls_size_and_time): Likewise.
6170 (estimate_node_size_and_time): Likewise.
6171 (inline_update_overall_summary): Update min_size.
6172 (do_estimate_edge_time): Likewise.
6173 (do_estimate_edge_size): Update.
6174 (do_estimate_edge_hints): Update.
6175 (growth_likely_positive): New function.
6177 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6180 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6181 also if addr has VOIDmode.
6183 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6185 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6186 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6188 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6189 instructions as well as AdvancedSIMD loads.
6191 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6193 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6194 Use crypto_aese type.
6195 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6196 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6197 crypto_aese, crypto_aesmc. Move to types.md.
6198 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6200 * config/arm/iterators.md (crypto_type): Likewise.
6202 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6204 * cgraph.c: Include expr.h and tree-dfa.h.
6205 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6208 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6211 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6212 regs from checking multi-reg pseudos.
6214 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6216 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6218 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6220 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6221 if it would clobber the stack pointer, even temporarily.
6223 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6225 * mode-switching.c: Make small adjustments to the top comment.
6227 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6229 * config/rs6000/constraints.md (wD constraint): New constraint to
6230 match the constant integer to get the top DImode/DFmode out of a
6231 vector in a VSX register.
6233 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6234 match the constant integer to get the top DImode/DFmode out of a
6235 vector in a VSX register.
6237 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6240 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6243 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6244 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6246 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6247 Optimize vec_extract of 64-bit values, where the value being
6248 extracted is in the top word, where we can use scalar
6249 instructions. Add direct move and store support. Combine the big
6250 endian/little endian vector select load support into a single insn.
6251 (vsx_extract_<mode>_internal1): Likewise.
6252 (vsx_extract_<mode>_internal2): Likewise.
6253 (vsx_extract_<mode>_load): Likewise.
6254 (vsx_extract_<mode>_store): Likewise.
6255 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6256 combined into vsx_extract_<mode>_load.
6257 (vsx_extract_<mode>_one_le): Likewise.
6259 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6260 define the top 64-bit vector element.
6262 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6265 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6266 Document vec_vbpermq builtin.
6269 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6270 enable use of xxsldwi and xxpermdi builtin functions.
6271 (vec_xxpermdi): Likewise.
6273 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6274 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6276 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6278 PR rtl-optimization/60650
6279 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6281 (find_spills_for): New.
6282 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6283 Spill all pseudos on the second iteration.
6285 2014-03-27 Marek Polacek <polacek@redhat.com>
6288 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6291 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6293 * config/s390/s390.c (s390_can_use_return_insn): Check for
6294 call-saved FPRs on 31 bit.
6296 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6299 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6300 if they need regimplification, just drop them instead of
6301 calling gimple_regimplify_operands on them.
6303 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6306 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6307 (aarch64_frame_pointer_required): Adjust logic.
6308 (aarch64_can_eliminate): Adjust logic.
6309 (aarch64_override_options_after_change): Adjust logic.
6311 2014-03-27 Dehao Chen <dehao@google.com>
6313 * ipa-inline.c (early_inliner): Update node's inline info.
6315 2014-03-26 Dehao Chen <dehao@google.com>
6317 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6318 compiler inserted conditional jumps for NAN float check.
6320 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6322 * ubsan.h (ubsan_create_data): Change second argument's type
6323 to const location_t *.
6324 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6326 (ubsan_create_data): Change second argument to const location_t *PLOC.
6327 Create Loc field whenever PLOC is non-NULL.
6328 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6329 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6333 * real.c (real_to_integer2): Change type of low to UHWI.
6335 2014-03-26 Tobias Burnus <burnus@net-b.de>
6337 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6338 (CILK_SELF_SPECS): New define.
6339 (driver_self_specs): Use it.
6341 2014-03-26 Richard Biener <rguenther@suse.de>
6343 * tree-pretty-print.c (percent_K_format): Implement special
6344 case for LTO and its stripped down BLOCK tree.
6346 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6349 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6351 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6352 one range is range_int_cst_p, but not both, at least optimize
6353 addition/subtraction of 0 and multiplication by 0 or 1.
6354 * gimple-fold.c (gimple_fold_call): Fold
6355 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6356 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6357 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6359 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6361 PR rtl-optimization/60452
6362 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6363 <case REG>: Return 1 for invalid offsets from the frame pointer.
6365 2014-03-26 Marek Polacek <polacek@redhat.com>
6368 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6371 2014-03-26 Marek Polacek <polacek@redhat.com>
6374 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6377 2014-03-26 Marek Polacek <polacek@redhat.com>
6380 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6381 multiplication in unsigned type.
6383 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6385 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6387 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6389 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6391 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6394 * cif-code.def (UNREACHABLE) New code.
6395 * ipa-inline.c (inline_small_functions): Skip edges to
6396 __builtlin_unreachable.
6397 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6398 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6399 predicate to __bulitin_unreachable.
6400 (set_cond_stmt_execution_predicate): Fix issue when
6401 invert_tree_comparison returns ERROR_MARK.
6402 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6403 propagate to inline clones.
6404 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6406 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6407 * cgraphclones.c (cgraph_clone_node): If call destination is already
6408 ureachable, do not redirect it back.
6409 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6412 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6414 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6415 Do not modify inline clones.
6417 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6419 * config/i386/i386.md (general_sext_operand): New mode attr.
6420 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6421 don't generate (sign_extend (const_int)).
6422 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6423 operands[2]. Use We constraint instead of <i> and
6424 <general_sext_operand> predicate instead of <general_operand>.
6425 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6426 * config/i386/constraints.md (We): New constraint.
6427 * config/i386/predicates.md (x86_64_sext_operand,
6428 sext_operand): New predicates.
6430 2014-03-25 Martin Jambor <mjambor@suse.cz>
6433 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6434 inconsistent devirtualizations to __builtin_unreachable.
6436 2014-03-25 Marek Polacek <polacek@redhat.com>
6439 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6441 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6443 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6444 order of elements for big-endian.
6446 2014-03-25 Richard Biener <rguenther@suse.de>
6449 * gimplify-me.c (gimple_regimplify_operands): Update the
6452 2014-03-25 Martin Jambor <mjambor@suse.cz>
6455 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6456 (lto_output_varpool_node): Likewise.
6457 (input_overwrite_node): Likewise.
6458 (input_varpool_node): Likewise.
6460 2014-03-25 Richard Biener <rguenther@suse.de>
6462 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6463 (run_gcc): Likewise.
6465 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6467 * combine.c (simplify_compare_const): Add MODE argument.
6468 Handle mode_width 0 as very large mode_width.
6469 (try_combine, simplify_comparison): Adjust callers.
6471 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6472 type to avoid signed integer overflow.
6473 * explow.c (plus_constant): Likewise.
6475 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6477 * doc/generic.texi: Correct typos.
6479 2014-03-24 Tobias Burnus <burnus@net-b.de>
6481 * doc/invoke.texi (-flto): Expand section about
6482 using static libraries with LTO.
6484 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6486 PR rtl-optimization/60501
6487 * optabs.def (addptr3_optab): New optab.
6488 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6489 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6490 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6492 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6494 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6496 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6498 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6501 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6502 (_mm256_undefined_ps): Define.
6503 (_mm256_undefined_pd): Define.
6504 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6505 (_mm_undefined_pd): Define.
6506 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6507 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6508 (_mm512_undefined_ps): Define.
6509 (_mm512_undefined_pd): Define.
6510 Use _mm*_undefined_*.
6511 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6513 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6515 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6516 (lshr_simd): DI mode added.
6517 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6518 (aarch64_ushr_simddi): Likewise.
6519 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6520 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6521 (vshrd_n_u64): Likewise.
6523 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6525 * Makefile.in (s-macro_list): Depend on cc1.
6527 2014-03-23 Teresa Johnson <tejohnson@google.com>
6529 * ipa-utils.c (ipa_print_order): Use specified dump file.
6531 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6533 PR rtl-optimization/60601
6534 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6536 * gcc.c (eval_spec_function): Initialize save_growing_value.
6538 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6541 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6542 code == MINUS_EXPR, never swap op0 with op1.
6544 * toplev.c (init_local_tick): Avoid signed integer multiplication
6546 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6547 shift by first operand's bitsize.
6549 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6552 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6554 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6555 TARGET_ISA_64BIT_P(x).
6557 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6559 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6560 pattern for vector nor instead of subtract from splat(-1).
6561 (altivec_expand_vec_perm_const_le): Likewise.
6563 2014-03-21 Richard Henderson <rth@twiddle.net>
6566 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6567 related insns after epilogue_completed.
6569 2014-03-21 Martin Jambor <mjambor@suse.cz>
6572 * cgraph.h (symtab_node): New flag body_removed.
6573 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6574 when removing bodies.
6575 * symtab.c (dump_symtab_base): Dump body_removed flag.
6576 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6577 had their bodies removed.
6579 2014-03-21 Martin Jambor <mjambor@suse.cz>
6582 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6585 2014-03-21 Richard Biener <rguenther@suse.de>
6587 PR tree-optimization/60577
6588 * tree-core.h (struct tree_base): Document nothrow_flag use
6589 in VAR_DECL_NONALIASED.
6590 * tree.h (VAR_DECL_NONALIASED): New.
6591 (may_be_aliased): Adjust.
6592 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6594 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6596 * expr.c (expand_expr_real_1): Remove outdated comment.
6598 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6601 * ira.c (adjust_cleared_regs): Call copy_rtx on
6602 *reg_equiv[REGNO (loc)].src_p before passing it to
6603 simplify_replace_fn_rtx.
6606 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6607 into CONST, put pic register as first operand of PLUS. Use
6608 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6610 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6612 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6614 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6616 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6617 around for store forwarding issue in the FPU on the UT699.
6618 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6619 loads and operations if -mfix-ut699 is specified.
6620 (divtf3_hq): Tweak attribute.
6621 (sqrttf2_hq): Likewise.
6623 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6625 * calls.c (store_one_arg): Remove incorrect const qualification on the
6626 type of the temporary.
6627 * cfgexpand.c (expand_return): Likewise.
6628 * expr.c (expand_constructor): Likewise.
6629 (expand_expr_real_1): Likewise.
6631 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6633 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6636 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6639 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6641 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6643 * config/arm/aarch-common-protos.h
6644 (alu_cost_table): Fix spelling of "extend".
6645 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6647 2014-03-19 Richard Biener <rguenther@suse.de>
6650 * tree-core.h (tree_type_common): Re-order pointer members
6651 to reduce recursion depth during GC walks.
6653 2014-03-19 Marek Polacek <polacek@redhat.com>
6656 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6657 before accessing it.
6659 2014-03-19 Richard Biener <rguenther@suse.de>
6662 * lto-streamer-in.c (input_function): In WPA stage do not drop
6665 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6667 PR tree-optimization/60559
6668 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6669 with build_zero_cst assignment.
6671 2014-03-18 Kai Tietz <ktietz@redhat.com>
6673 PR rtl-optimization/56356
6674 * sdbout.c (sdbout_parms): Verify that parms'
6675 incoming argument is valid.
6676 (sdbout_reg_parms): Likewise.
6678 2014-03-18 Richard Henderson <rth@redhat.com>
6681 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6682 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6683 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6685 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6687 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6688 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6689 Italicize plugin event names in description. Explain that
6690 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6691 Remind that no GCC functions should be called after PLUGIN_FINISH.
6692 Explain what pragmas with expansion are.
6694 2014-03-18 Martin Liska <mliska@suse.cz>
6696 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6697 gimple call statement is update.
6698 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6699 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6701 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6704 * ubsan.c (ubsan_instrument_unreachable): Call
6705 initialize_sanitizer_builtins.
6706 (ubsan_pass): Likewise.
6709 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6710 varpool_finalize_decl instead of rest_of_decl_compilation.
6712 2014-03-18 Richard Biener <rguenther@suse.de>
6714 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6715 by using bitmap_and_compl instead of bitmap_and_compl_into.
6716 (df_rd_transfer_function): Likewise.
6718 2014-03-18 Richard Biener <rguenther@suse.de>
6720 * doc/lto.texi (fresolution): Fix typo.
6722 2014-03-18 Richard Biener <rguenther@suse.de>
6724 * doc/invoke.texi (flto): Update for changes in 4.9.
6726 2014-03-18 Richard Biener <rguenther@suse.de>
6728 * doc/loop.texi: Remove section on the removed lambda framework.
6729 Update loop docs with recent changes in preserving loop structure.
6731 2014-03-18 Richard Biener <rguenther@suse.de>
6733 * doc/lto.texi (-fresolution): Document.
6735 2014-03-18 Richard Biener <rguenther@suse.de>
6737 * doc/contrib.texi: Adjust my name.
6739 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6742 * internal-fn.c: Include diagnostic-core.h.
6743 (expand_BUILTIN_EXPECT): New function.
6744 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6745 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6746 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6747 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6749 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6750 Revert 3 argument __builtin_expect code.
6751 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6752 * gimple-fold.c (gimple_fold_call): Likewise.
6753 * tree.h (fold_builtin_expect): New prototype.
6754 * builtins.c (build_builtin_expect_predicate): Add predictor
6755 argument, if non-NULL, create 3 argument __builtin_expect.
6756 (fold_builtin_expect): No longer static. Add ARG2 argument,
6757 pass it through to build_builtin_expect_predicate.
6758 (fold_builtin_2): Adjust caller.
6759 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6760 * internal-fn.def (BUILTIN_EXPECT): New.
6762 2014-03-18 Tobias Burnus <burnus@net-b.de>
6765 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6766 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6767 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6769 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6772 * predict.c (combine_predictions_for_bb): Fix up formatting.
6773 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6774 fill what it points to if non-NULL.
6775 (tree_predict_by_opcode): Adjust caller, use the predictor.
6776 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6778 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6780 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6781 proper constant for the store mode.
6783 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6785 * symtab.c (change_decl_assembler_name): Fix transparent alias
6788 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6790 * config/aarch64/aarch64.c: Correct the comments about the
6791 aarch64 stack layout.
6793 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6795 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6796 check for GF_OMP_FOR_KIND_FOR.
6798 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6800 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6801 ymm and zmm register names.
6803 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6806 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6807 note creation for the 2010-08-31 changes.
6809 2014-03-17 Marek Polacek <polacek@redhat.com>
6812 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6813 as -fno-tree-loop-vectorize.
6814 (expand_omp_simd): Likewise.
6816 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6818 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6819 (eligible_for_call_delay): New prototype.
6820 * config/sparc/sparc.c (tls_call_delay): Rename into...
6821 (eligible_for_call_delay): ...this. Return false if the instruction
6822 cannot be put in the delay slot of a branch.
6823 (eligible_for_restore_insn): Simplify.
6824 (eligible_for_return_delay): Return false if the instruction cannot be
6825 put in the delay slot of a branch and simplify.
6826 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6827 put in the delay slot of a branch.
6828 * config/sparc/sparc.md (fix_ut699): New attribute.
6829 (tls_call_delay): Delete.
6830 (in_call_delay): Reimplement.
6831 (eligible_for_sibcall_delay): Rename into...
6832 (in_sibcall_delay): ...this.
6833 (eligible_for_return_delay): Rename into...
6834 (in_return_delay): ...this.
6835 (in_branch_delay): Reimplement.
6836 (in_uncond_branch_delay): Delete.
6837 (in_annul_branch_delay): Delete.
6839 2014-03-14 Richard Henderson <rth@redhat.com>
6842 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6843 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6844 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6845 (floathi splitters): Remove.
6846 (float<SWI48x>xf2): New pattern.
6847 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6848 code that tried to handle DImode for 32-bit, but which was excluded
6849 by the pattern's condition. Drop allocation of stack temporary.
6850 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6851 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6852 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6853 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6854 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6855 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6856 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6857 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6858 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6859 (*float<SWI48x><X87MODEF>2_i387): Remove.
6860 (all float _with_temp splitters): Remove.
6861 (*float<SWI48x><MODEF>2_i387): New pattern.
6862 (*float<SWI48><MODEF>2_sse): New pattern.
6863 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6864 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6866 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6867 Marek Polacek <polacek@redhat.com>
6870 * common.opt (dump_base_name_prefixed): New Variable.
6871 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6872 if x_dump_base_name_prefixed is already set, set it at the end.
6874 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6876 PR rtl-optimization/60508
6877 * lra-constraints.c (get_reload_reg): Add new parameter
6879 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6880 Pass the new parameter values.
6882 2014-03-14 Richard Biener <rguenther@suse.de>
6884 * common.opt: Revert unintented changes from r205065.
6887 2014-03-14 Richard Biener <rguenther@suse.de>
6890 * cfghooks.c (split_block): Properly adjust all loops the
6891 block was a latch of.
6893 2014-03-14 Martin Jambor <mjambor@suse.cz>
6896 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6899 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6902 * config/avr/avr.c (avr_set_current_function): Pass function name
6903 through default_strip_name_encoding before sanity checking instead
6904 of skipping the first char of the assembler name.
6906 2014-03-13 Richard Henderson <rth@redhat.com>
6909 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6910 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6911 * config/i386/i386-protos.h: Likewise.
6912 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6913 in the expander instead of a splitter.
6914 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6915 any possibility of requiring a memory.
6916 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6917 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6918 (fp branch splitters): Update for ix86_split_fp_branch.
6919 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6920 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6921 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6922 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6923 (*fop_<MODEF>_3_i387): Likewise.
6924 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6925 (splitters for the fop_* register patterns): Remove.
6926 (fscalexf4_i387): Rename from *fscalexf4_i387.
6927 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6929 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6931 PR tree-optimization/59779
6932 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6933 type for bitsize and maxsize instead of HOST_WIDE_INT.
6935 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6937 PR rtl-optimization/57320
6938 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6939 the CFG after thread_prologue_and_epilogue_insns.
6941 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6943 PR rtl-optimization/57189
6944 * lra-constraints.c (process_alt_operands): Disfavor spilling
6947 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6949 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6951 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6953 PR tree-optimization/59025
6955 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6956 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6958 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6961 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6962 calls of avr_out_plus_1.
6964 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6966 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6967 BB's single pred and update the father loop's latch info later.
6969 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6971 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6975 (VEC_base): Likewise.
6976 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6977 registers, we need to swap double words in little endian mode.
6979 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6980 to be a container mode for 128-bit integer operations added in ISA
6981 2.07. Unlike TImode and PTImode, the preferred register set is
6982 the Altivec/VMX registers for the 128-bit operations.
6984 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6986 (rs6000_split_128bit_ok_p): Likewise.
6988 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6989 macros for creating ISA 2.07 normal and overloaded builtin
6990 functions with 3 arguments.
6991 (BU_P8V_OVERLOAD_3): Likewise.
6992 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6993 for use as overloaded functions.
6994 (VPERM_1TI_UNS): Likewise.
6995 (VSEL_1TI): Likewise.
6996 (VSEL_1TI_UNS): Likewise.
6997 (ST_INTERNAL_1ti): Likewise.
6998 (LD_INTERNAL_1ti): Likewise.
6999 (XXSEL_1TI): Likewise.
7000 (XXSEL_1TI_UNS): Likewise.
7001 (VPERM_1TI): Likewise.
7002 (VPERM_1TI_UNS): Likewise.
7003 (XXPERMDI_1TI): Likewise.
7004 (SET_1TI): Likewise.
7005 (LXVD2X_V1TI): Likewise.
7006 (STXVD2X_V1TI): Likewise.
7007 (VEC_INIT_V1TI): Likewise.
7008 (VEC_SET_V1TI): Likewise.
7009 (VEC_EXT_V1TI): Likewise.
7010 (EQV_V1TI): Likewise.
7011 (NAND_V1TI): Likewise.
7012 (ORC_V1TI): Likewise.
7013 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7014 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7016 (VADDUQM): Likewise.
7017 (VSUBCUQ): Likewise.
7018 (VADDEUQM): Likewise.
7019 (VADDECUQ): Likewise.
7020 (VSUBEUQM): Likewise.
7021 (VSUBECUQ): Likewise.
7023 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7024 __int128_t and __uint128_t types.
7025 (__uint128_type): Likewise.
7026 (altivec_categorize_keyword): Add support for vector __int128_t,
7027 vector __uint128_t, vector __int128, and vector unsigned __int128
7028 as a container type for TImode operations that need to be done in
7029 VSX/Altivec registers.
7030 (rs6000_macro_to_expand): Likewise.
7031 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7032 to support 128-bit integer instructions vaddcuq, vadduqm,
7033 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7034 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7036 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7037 for V1TImode, and set up preferences to use VSX/Altivec registers.
7038 Setup VSX reload handlers.
7039 (rs6000_debug_reg_global): Likewise.
7040 (rs6000_init_hard_regno_mode_ok): Likewise.
7041 (rs6000_preferred_simd_mode): Likewise.
7042 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7043 (easy_altivec_constant): Likewise.
7044 (output_vec_const_move): Likewise.
7045 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7047 (rs6000_expand_vector_extract): Likewise.
7048 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7050 (rs6000_const_vec): Add support for V1TImode.
7051 (rs6000_emit_le_vsx_load): Swap double words when loading or
7052 storing TImode/V1TImode.
7053 (rs6000_emit_le_vsx_store): Likewise.
7054 (rs6000_emit_le_vsx_move): Likewise.
7055 (rs6000_emit_move): Add support for V1TImode.
7056 (altivec_expand_ld_builtin): Likewise.
7057 (altivec_expand_st_builtin): Likewise.
7058 (altivec_expand_vec_init_builtin): Likewise.
7059 (altivec_expand_builtin): Likewise.
7060 (rs6000_init_builtins): Add support for V1TImode type. Add
7061 support for ISA 2.07 128-bit integer builtins. Define type names
7062 for the VSX/Altivec vector types.
7063 (altivec_init_builtins): Add support for overloaded vector
7064 functions with V1TImode type.
7065 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7066 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7068 (rs6000_split_128bit_ok_p): Likewise.
7069 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7070 __int128_t and vector __uint128_t.
7072 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7073 and mode attributes.
7080 (VS_scalar): Likewise.
7081 (VS_double): Likewise.
7082 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7084 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7085 we support the ISA 2.07 128-bit integer arithmetic instructions.
7086 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7087 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7088 and TImode types for use with the builtin functions.
7089 (V1TI_type_node): Likewise.
7090 (unsigned_V1TI_type_node): Likewise.
7091 (intTI_type_internal_node): Likewise.
7092 (uintTI_type_internal_node): Likewise.
7094 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7095 128-bit builtin functions.
7096 (UNSPEC_VADDEUQM): Likewise.
7097 (UNSPEC_VADDECUQ): Likewise.
7098 (UNSPEC_VSUBCUQ): Likewise.
7099 (UNSPEC_VSUBEUQM): Likewise.
7100 (UNSPEC_VSUBECUQ): Likewise.
7101 (VM): Add V1TImode to vector mode iterators.
7103 (VI_unit): Likewise.
7104 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7105 (altivec_vaddcuq): Likewise.
7106 (altivec_vsubuqm): Likewise.
7107 (altivec_vsubcuq): Likewise.
7108 (altivec_vaddeuqm): Likewise.
7109 (altivec_vaddecuq): Likewise.
7110 (altivec_vsubeuqm): Likewise.
7111 (altivec_vsubecuq): Likewise.
7113 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7115 (BOOL_128): Likewise.
7116 (BOOL_REGS_OUTPUT): Likewise.
7117 (BOOL_REGS_OP1): Likewise.
7118 (BOOL_REGS_OP2): Likewise.
7119 (BOOL_REGS_UNARY): Likewise.
7120 (BOOL_REGS_AND_CR0): Likewise.
7122 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7123 128-bit integer builtin support.
7124 (vec_vadduqm): Likewise.
7125 (vec_vaddecuq): Likewise.
7126 (vec_vaddeuqm): Likewise.
7127 (vec_vsubecuq): Likewise.
7128 (vec_vsubeuqm): Likewise.
7129 (vec_vsubcuq): Likewise.
7130 (vec_vsubuqm): Likewise.
7132 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7133 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7134 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7135 128-bit integer add/subtract to ISA 2.07.
7137 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7139 * config/arc/arc.c (arc_predicate_delay_insns):
7140 Fix third argument passed to conditionalize_nonjump.
7142 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7144 * config/aarch64/aarch64-builtins.c
7145 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7146 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7147 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7148 instead of __builtin_lfloor.
7149 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7151 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7153 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7154 (tree_ssa_ifcombine_bb_1): New function.
7155 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7156 is an empty forwarder block to then_bb or vice versa and then_bb
7157 and else_bb are effectively swapped.
7159 2014-03-12 Christian Bruel <christian.bruel@st.com>
7162 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7163 REG_CFA_DEF_CFA note.
7164 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7165 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7167 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7169 PR tree-optimization/60454
7170 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7172 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7174 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7175 Do not define target_cpu_default2 to generic.
7176 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7177 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7178 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7180 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7181 Marc Glisse <marc.glisse@inria.fr>
7183 PR tree-optimization/60502
7184 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7185 instead of build_low_bits_mask.
7187 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7190 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7191 if there are multiple uses, but op doesn't live on E edge.
7192 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7193 clobber stmts before __builtin_unreachable.
7195 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7197 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7198 hard_frame_pointer_rtx.
7199 * cse.c (cse_insn): Remove volatile check.
7200 * cselib.c (cselib_process_insn): Likewise.
7201 * dse.c (scan_insn): Likewise.
7203 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7205 * config/arc/arc.c (conditionalize_nonjump): New function,
7207 (arc_ifcvt): ... this.
7208 (arc_predicate_delay_insns): Use it.
7210 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7212 * config/arc/predicates.md (extend_operand): During/after reload,
7213 allow const_int_operand.
7214 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7215 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7216 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7218 (umulsi3_highpart_i): Likewise.
7220 2014-03-11 Richard Biener <rguenther@suse.de>
7222 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7223 Add asserts to guard possible wrong-code bugs.
7225 2014-03-11 Richard Biener <rguenther@suse.de>
7227 PR tree-optimization/60429
7228 PR tree-optimization/60485
7229 * tree-ssa-structalias.c (set_union_with_increment): Properly
7230 take into account all fields that overlap the shifted vars.
7231 (do_sd_constraint): Likewise.
7232 (do_ds_constraint): Likewise.
7233 (get_constraint_for_ptr_offset): Likewise.
7235 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7237 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7238 (nios2_compute_frame_layout):
7239 Add calculation of cfun->machine->fp_save_offset.
7240 (nios2_expand_prologue): Correct setting of frame pointer register
7242 (nios2_expand_epilogue): Update recovery of stack pointer from
7243 frame pointer accordingly.
7244 (nios2_initial_elimination_offset): Update calculation of offset
7245 for eliminating to HARD_FRAME_POINTER_REGNUM.
7247 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7250 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7251 cgraph_get_create_node on VAR_DECLs.
7253 2014-03-10 Richard Biener <rguenther@suse.de>
7256 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7258 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7260 * config/vms/vms.opt (vms_float_format): New variable.
7262 2014-03-08 Tobias Burnus <burnus@net-b.de>
7264 * doc/invoke.texi (-fcilkplus): Update implementation status.
7266 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7267 Richard Biener <rguenther@suse.de>
7269 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7270 consistently accross all TUs.
7271 (run_gcc): Enable -fshort-double automatically at link at link-time
7272 and disallow override.
7274 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7277 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7278 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7279 if they can't be used.
7281 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7283 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7284 for Solaris 11/x86 ld.
7285 * configure: Regenerate.
7287 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7289 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7290 (LIB_TLS_SPEC): Save as ld_tls_libs.
7291 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7292 (HAVE_AS_IX86_TLSLDM): New test.
7293 * configure, config.in: Regenerate.
7294 * config/i386/i386.c (legitimize_tls_address): Fall back to
7295 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7296 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7297 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7298 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7300 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7302 * common.opt (fira-loop-pressure): Mark as optimization.
7304 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7306 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7307 an OpenMP mappable type.
7309 2014-03-06 Matthias Klose <doko@ubuntu.com>
7311 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7312 MULTILIB_OSDIRNAMES is not defined.
7314 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7315 Meador Inge <meadori@codesourcery.com>
7318 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7319 (arm_legitimize_address): Call legitimize_tls_address for any
7320 arm_tls_referenced_p expression, handle constant addend. Call it
7321 before testing for !TARGET_ARM.
7322 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7324 2014-03-06 Richard Biener <rguenther@suse.de>
7330 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7332 * tree-streamer.c (record_common_node): Assert we don't record
7333 nodes with type double.
7334 (preload_common_node): Skip type double, complex double and double
7335 pointer since it is now frontend dependent due to fshort-double option.
7337 2014-03-06 Richard Biener <rguenther@suse.de>
7339 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7340 or -fno-lto is specified and the linker has full plugin support.
7341 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7342 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7343 * lto-wrapper.c (merge_and_complain): Merge compile-time
7344 optimization levels.
7345 (run_gcc): And pass it through to the link options.
7347 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7351 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7353 * cselib.c (remove_useless_values): Skip to avoid quadratic
7354 behavior if the condition moved from...
7355 (cselib_process_insn): ... here holds.
7357 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7360 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7361 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7364 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7365 (TM_H): Add x86-tune.def.
7367 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7369 * config/aarch64/aarch64.c (generic_tunings):
7370 Use cortexa57_extra_costs.
7372 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7375 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7376 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7377 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7380 2014-03-04 Heiher <r@hev.cc>
7382 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7383 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7385 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7387 * config/i386/predicates.md (const2356_operand): Change to ...
7388 (const2367_operand): ... this.
7389 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7391 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7392 (*avx512pf_scatterpf<mode>sf): Ditto.
7393 (avx512pf_scatterpf<mode>df): Ditto.
7394 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7395 (*avx512pf_scatterpf<mode>df): Ditto.
7396 * config/i386/i386.c (ix86_expand_builtin): Update
7397 incorrect hint operand error message.
7399 2014-03-04 Richard Biener <rguenther@suse.de>
7401 * lto-section-in.c (lto_get_section_data): Fix const cast.
7403 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7405 * tree-streamer.c (record_common_node): Assert we don't record
7406 nodes with type double.
7407 (preload_common_node): Skip type double, complex double and double
7408 pointer since it is now frontend dependent due to fshort-double option.
7410 2014-03-04 Richard Biener <rguenther@suse.de>
7413 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7414 (lto_input_toplevel_asms): Likewise.
7415 * lto-section-in.c (lto_get_section_data): Instead do it here
7418 2014-03-04 Richard Biener <rguenther@suse.de>
7420 PR tree-optimization/60382
7421 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7422 dead PHIs a reduction.
7424 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7426 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7428 (_mm_prefetch): Move out of GCC target("sse") pragma.
7429 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7430 GCC target("prfchw") pragma.
7431 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7433 * config/i386/i386.c (ix86_option_override_internal): Enable
7434 -mprfchw with -mprefetchwt1.
7436 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7438 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7441 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7443 * opts.h (CL_PCH_IGNORE): Define.
7444 * targhooks.c (option_affects_pch_p):
7445 Return false for options that have CL_PCH_IGNORE set.
7446 * opt-functions.awk: Process PchIgnore.
7447 * doc/options.texi: Document PchIgnore.
7449 * config/arc/arc.opt (misize): Add PchIgnore property.
7451 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7453 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7454 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7455 constraint on constants to permit them being loaded into
7456 GENERAL_REGS or BASE_REGS.
7458 2014-03-03 Nick Clifton <nickc@redhat.com>
7460 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7461 anti-cacnonical alternatives.
7462 (negandhi3_real): New pattern.
7463 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7465 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7467 * config/avr/avr-mcus.def: Remove atxmega16x1.
7468 * config/avr/avr-tables.opt: Regenerate.
7469 * config/avr/t-multilib: Regenerate.
7470 * doc/avr-mmcu.texi: Regenerate.
7472 2014-03-03 Tobias Grosser <tobias@grosser.es>
7473 Mircea Namolaru <mircea.namolaru@inria.fr>
7475 PR tree-optimization/58028
7476 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7479 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7481 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7482 not handled by recognizers.
7484 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7487 * function.c (expand_function_end): Don't emit
7488 clobber_return_register sequence if clobber_after is a BARRIER.
7489 * cfgexpand.c (construct_exit_block): Append instructions before
7490 return_label to prev_bb.
7492 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7494 * config/rs6000/constraints.md: Document reserved use of "wc".
7496 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7499 * ipa.c (function_and_variable_visibility): When dissolving comdat
7500 group, also set all symbols to local.
7502 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7507 2013-12-14 Jan Hubicka <jh@suse.cz>
7509 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7511 2014-03-02 Jon Beniston <jon@beniston.com>
7517 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7518 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7519 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7520 (simple_return, *simple_return): New patterns
7521 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7522 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7524 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7526 * dwarf2out.c (gen_subprogram_die): Tidy.
7528 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7531 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7532 (*mov_t_msb_neg_negc): ... this new insn.
7534 2014-02-28 Jason Merrill <jason@redhat.com>
7537 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7540 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7543 * dwarf2out.c (decltype_auto_die): New static.
7544 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7545 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7546 (is_cxx_auto): Likewise.
7548 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7550 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7551 we are not using general regs only.
7553 2014-02-28 Richard Biener <rguenther@suse.de>
7556 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7557 previous fix and only allow to remove trivial pre-headers
7558 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7559 (remove_forwarder_block): Properly update the latch of a loop.
7561 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7564 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7565 (cselib_preserved_hash_table): New.
7566 (preserve_constants_and_equivs): Move preserved vals to it.
7567 (cselib_find_slot): Look it up first.
7568 (cselib_init): Initialize it.
7569 (cselib_finish): Release it.
7570 (dump_cselib_table): Dump it.
7572 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7575 * cselib.c (remove_useless_values): Skip to avoid quadratic
7576 behavior if the condition moved from...
7577 (cselib_process_insn): ... here holds.
7579 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7582 * var-tracking.c (vt_initialize): Apply the same condition to
7583 preserve the CFA base value.
7585 2014-02-28 Joey Ye <joey.ye@arm.com>
7588 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7589 if reload in progress or completed.
7591 2014-02-28 Tobias Burnus <burnus@net-b.de>
7594 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7597 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7599 * doc/tm.texi.in (Condition Code Status): Update documention for
7600 relative locations of cc0-setter and cc0-user.
7602 2014-02-27 Jeff Law <law@redhat.com>
7604 PR rtl-optimization/52714
7605 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7606 into two independent simple sets, if I3 is a jump, ensure the
7607 pattern we place into I3 is a (set (pc) ...).
7609 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7610 Jeff Law <law@redhat.com>
7612 PR rtl-optimization/49847
7613 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7614 are in different blocks.
7615 * doc/tm.texi (Condition Code Status): Update documention for
7616 relative locations of cc0-setter and cc0-user.
7618 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7621 * lra.c (lra_emit_add): Check SUBREG too.
7623 2014-02-27 Andreas Schwab <schwab@suse.de>
7625 * config/m68k/m68k.c (m68k_option_override): Disable
7626 -flive-range-shrinkage for classic m68k.
7627 (m68k_override_options_after_change): Likewise.
7629 2014-02-27 Marek Polacek <polacek@redhat.com>
7632 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7633 -Wmaybe-uninitialized.
7635 2014-02-27 Alan Modra <amodra@gmail.com>
7638 * reload1.c (emit_input_reload_insns): When reload_override_in,
7639 set old to rl->in_reg when rl->in_reg is a subreg.
7641 2014-02-26 Richard Biener <rguenther@suse.de>
7644 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7646 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7648 * common/config/i386/predicates.md (const1256_operand): Remove.
7649 (const2356_operand): New.
7650 (const_1_to_2_operand): Remove.
7651 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7652 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7653 (*avx512pf_gatherpf<mode>sf): Ditto.
7654 (avx512pf_gatherpf<mode>df): Ditto.
7655 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7656 (*avx512pf_gatherpf<mode>df): Ditto.
7657 (avx512pf_scatterpf<mode>sf): Ditto.
7658 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7659 (*avx512pf_scatterpf<mode>sf): Ditto.
7660 (avx512pf_scatterpf<mode>df): Ditto.
7661 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7662 (*avx512pf_scatterpf<mode>df): Ditto.
7663 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7665 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7667 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7668 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7669 (_mm512_mask_testn_epi64_mask): Move to ...
7670 * config/i386/avx512cdintrin.h: Here.
7671 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7672 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7673 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7674 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7675 TARGET_AVX512F from TARGET_AVX512CD.
7677 2014-02-26 Richard Biener <rguenther@suse.de>
7680 * ipa.c (walk_polymorphic_call_targets): Properly guard
7681 call to inline_update_overall_summary.
7683 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7686 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7687 and latches only if requested. Fix latch if it is removed.
7688 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7689 LOOPS_HAVE_PREHEADERS.
7691 2014-02-25 Andrew Pinski <apinski@cavium.com>
7693 * builtins.c (expand_builtin_thread_pointer): Create a new target
7694 when the target is NULL.
7696 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7698 PR rtl-optimization/60317
7699 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7700 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7701 * lra-assigns.c: Include params.h.
7702 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7703 other reload pseudos considerations.
7705 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7707 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7708 to use canonical form for nor<mode>3.
7710 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7713 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7716 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7718 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7719 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7720 (ix86_handle_option): Handle OPT_mprefetchwt1.
7721 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7722 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7724 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7725 OPTION_MASK_ISA_PREFETCHWT1.
7726 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7727 (PTA_PREFETCHWT1): New.
7728 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7729 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7730 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7731 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7732 (*prefetch_avx512pf_<mode>_: Change into ...
7733 (*prefetch_prefetchwt1_<mode>: This.
7734 * config/i386/i386.opt (mprefetchwt1): New.
7735 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7736 (_mm_prefetch): Handle intent to write.
7737 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7739 2014-02-25 Richard Biener <rguenther@suse.de>
7742 * emit-rtl.c (mem_attrs_htab): Remove.
7743 (mem_attrs_htab_hash): Likewise.
7744 (mem_attrs_htab_eq): Likewise.
7745 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7746 (init_emit_once): Do not allocate mem_attrs_htab.
7748 2014-02-25 Richard Biener <rguenther@suse.de>
7751 * lto-opts.c (lto_write_options): Output non-explicit conservative
7752 -fwrapv, -fno-trapv and -fno-strict-overflow.
7753 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7754 (run_gcc): And pass them through.
7756 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7758 * sel-sched.c (calculate_new_fences): New parameter ptime.
7759 Calculate it as a maximum over all fence cycles.
7760 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7761 Print the final schedule timing when sched_verbose.
7763 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7765 PR rtl-optimization/60292
7766 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7767 bit fot the fence instruction.
7769 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7771 * calls.h: Fix typo in comment.
7773 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7775 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7776 adjusting offsetable addresses.
7778 2014-02-24 Guozhi Wei <carrot@google.com>
7780 * sparseset.h (sparseset_pop): Fix the wrong index.
7782 2014-02-24 Walter Lee <walt@tilera.com>
7784 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7785 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7787 * common/config/tilegx/tilegx-common.c
7788 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7789 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7791 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7792 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7793 (tilegx_gimplify_va_arg_expr): Handle big endian.
7794 (tilegx_expand_unaligned_load): Ditto.
7795 (tilegx_expand_unaligned_store): Ditto.
7796 (TARGET_RETURN_IN_MSB): New.
7797 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7798 (TARGET_ENDIAN_DEFAULT): New.
7799 (TARGET_BIG_ENDIAN): Handle big endian.
7800 (BYTES_BIG_ENDIAN): Ditto.
7801 (WORDS_BIG_ENDIAN): Ditto.
7802 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7805 * config/tilegx/tilegx.md (extv): Handle big endian.
7807 (insn_st<n>): Ditto.
7808 (insn_st<n>_add<bitsuffix>): Ditto.
7809 (insn_stnt<n>): Ditto.
7810 (insn_stnt<n>_add<bitsuffix>):Ditto.
7811 (vec_interleave_highv8qi): Handle big endian.
7812 (vec_interleave_highv8qi_be): New.
7813 (vec_interleave_highv8qi_le): New.
7814 (insn_v1int_h): Handle big endian.
7815 (vec_interleave_lowv8qi): Handle big endian.
7816 (vec_interleave_lowv8qi_be): New.
7817 (vec_interleave_lowv8qi_le): New.
7818 (insn_v1int_l): Handle big endian.
7819 (vec_interleave_highv4hi): Handle big endian.
7820 (vec_interleave_highv4hi_be): New.
7821 (vec_interleave_highv4hi_le): New.
7822 (insn_v2int_h): Handle big endian.
7823 (vec_interleave_lowv4hi): Handle big endian.
7824 (vec_interleave_lowv4hi_be): New.
7825 (vec_interleave_lowv4hi_le): New.
7826 (insn_v2int_l): Handle big endian.
7827 (vec_interleave_highv2si): Handle big endian.
7828 (vec_interleave_highv2si_be): New.
7829 (vec_interleave_highv2si_le): New.
7830 (insn_v4int_h): Handle big endian.
7831 (vec_interleave_lowv2si): Handle big endian.
7832 (vec_interleave_lowv2si_be): New.
7833 (vec_interleave_lowv2si_le): New.
7834 (insn_v4int_l): Handle big endian.
7835 * config/tilegx/tilegx.opt (mbig-endian): New option.
7836 (mlittle-endian): New option.
7837 * doc/install.texi: Document tilegxbe-linux.
7838 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7840 2014-02-24 Martin Jambor <mjambor@suse.cz>
7843 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7844 there are no parameter descriptors.
7846 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7848 PR rtl-optimization/60268
7849 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7850 initialization to ...
7851 (sched_rgn_init): ... here.
7852 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7854 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7856 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7859 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7861 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7864 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7866 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7867 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7869 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7871 * config/microblaze/predicates.md: Add cmp_op predicate.
7872 * config/microblaze/microblaze.md: Add branch_compare instruction
7873 which uses cmp_op predicate and emits cmp insn before branch.
7874 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7875 to microblaze_expand_conditional_branch and consolidate logic.
7876 (microblaze_expand_conditional_branch): emit branch_compare
7877 insn instead of handling cmp op separate from branch insn.
7879 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7881 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7884 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7886 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7887 define_insn with define_expand and new define_insn
7888 *altivec_lve<VI_char>x_internal.
7889 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7890 and new define_insn *altivec_stve<VI_char>x_internal.
7891 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7893 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7895 (altivec_expand_stvex_be): New function.
7897 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7899 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7900 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7901 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7902 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7904 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7907 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7908 instead of emit_move_insn.
7910 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7912 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7914 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7915 gen_altivec_vsumsws.
7917 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7919 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7920 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7921 (altivec_lvxl_<mode>): New define_expand incorporating
7922 -maltivec=be semantics where needed.
7923 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7924 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7925 semantics where needed.
7926 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7927 (altivec_stvx_<mode>): New define_expand incorporating
7928 -maltivec=be semantics where needed.
7929 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7930 VM2 iterator instead of V4SI.
7931 (altivec_stvxl_<mode>): New define_expand incorporating
7932 -maltivec=be semantics where needed.
7933 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7934 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7935 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7936 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7937 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7938 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7939 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7940 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7941 ALTIVEC_BUILTIN_STVXL.
7942 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7943 (altivec_expand_stvx_be): Likewise.
7944 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7945 (altivec_expand_lvx_be): Likewise.
7946 (altivec_expand_stvx_be): Likewise.
7947 (altivec_expand_builtin): Add cases for
7948 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7949 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7950 (altivec_init_builtins): Add definitions for
7951 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7952 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7954 2014-02-21 Catherine Moore <clm@codesourcery.com>
7956 * doc/invoke.texi (mvirt, mno-virt): Document.
7957 * config/mips/mips.opt (mvirt): New option.
7958 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7960 2014-02-21 Richard Biener <rguenther@suse.de>
7962 PR tree-optimization/60276
7963 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7964 (STMT_VINFO_MIN_NEG_DIST): New macro.
7965 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7966 STMT_VINFO_MIN_NEG_DIST.
7967 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7968 made for negative dependence distances still hold.
7970 2014-02-21 Richard Biener <rguenther@suse.de>
7973 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7974 DECL_INITIAL for globals not in the current function context.
7976 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7978 PR tree-optimization/56490
7979 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7980 * tree-ssa-uninit.c: Include params.h.
7981 (compute_control_dep_chain): Add num_calls argument, return false
7982 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7983 num_calls to recursive call.
7984 (find_predicates): Change dep_chain into normal array,
7985 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7986 variable and adjust compute_control_dep_chain caller.
7987 (find_def_preds): Likewise.
7989 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7991 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7992 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7994 2014-02-21 Nick Clifton <nickc@redhat.com>
7996 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7997 (pushhi1): Likewise.
7998 (popqi1): Add mode to pre_dec.
8001 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8003 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8004 mode for mask of V8SFmode permutation.
8006 2014-02-20 Richard Henderson <rth@redhat.com>
8009 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8010 a new pseudo for OLDVAL.
8012 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8015 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8016 gen_reg_rtx if d->testing_p.
8017 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8018 if d->testing_p and we will certainly return true.
8019 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8022 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8024 * emit-rtl.c (gen_reg_rtx): Assert that
8025 crtl->emit.regno_pointer_align_length is non-zero.
8027 2014-02-20 Richard Henderson <rth@redhat.com>
8030 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8031 on failure the store back into EXPECT.
8033 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8034 Sandra Loosemore <sandra@codesourcery.com>
8036 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8037 * config/nios2/nios2.c (nios2_function_profiler): Add
8038 -fPIC (flag_pic == 2) support.
8039 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8040 (nios2_large_offset_p): New function.
8041 (nios2_unspec_reloc_p): Move up position, update to use
8042 nios2_large_offset_p.
8043 (nios2_unspec_address): Remove function.
8044 (nios2_unspec_offset): New function.
8045 (nios2_large_got_address): New function.
8046 (nios2_got_address): Add large offset support.
8047 (nios2_legitimize_tls_address): Update usage of removed and new
8049 (nios2_symbol_binds_local_p): New function.
8050 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8051 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8052 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8053 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8054 processing for (const (unspec ...)).
8055 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8057 2014-02-20 Richard Biener <rguenther@suse.de>
8059 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8060 doing the substitution.
8061 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8063 2014-02-20 Martin Jambor <mjambor@suse.cz>
8066 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8067 info when checking whether lattices are bottom.
8069 2014-02-20 Richard Biener <rguenther@suse.de>
8072 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8075 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8078 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8079 parameter specifying the scaling.
8080 (inline_call): Update.
8081 (want_inline_recursively): Guard division by zero.
8082 (recursive_inlining): Update.
8083 * ipa-inline.h (clone_inlined_nodes): Update.
8085 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8088 * config/i386/i386.c (classify_argument): Pass structures of size
8089 64 bytes or less in register.
8091 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8092 Kirill Yukhin <kirill.yukhin@intel.com>
8094 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8095 (_mm_rcp28_round_ss): Ditto.
8096 (_mm_rsqrt28_round_sd): Ditto.
8097 (_mm_rsqrt28_round_ss): Ditto.
8098 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8099 (_mm_rcp14_round_ss): Ditto.
8100 (_mm_rsqrt14_round_sd): Ditto.
8101 (_mm_rsqrt14_round_ss): Ditto.
8102 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8103 the first input operand, get rid of match_dup.
8104 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8106 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8108 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8109 operand as the first input operand, set type attribute.
8110 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8112 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8113 operand as the first input operand, set type attribute.
8115 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8117 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8120 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8123 * config/i386/i386.c (construct_container): Remove TFmode check
8124 for X86_64_INTEGER_CLASS.
8126 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8129 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8130 only when -Wpsabi is enabled.
8132 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8135 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8136 passing arrays in registers are the same as for structs, so remove the
8137 special case for them.
8139 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8141 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8142 destination type, extract only the valid bits if the source type is not
8143 integral and has a different mode.
8145 2014-02-19 Richard Biener <rguenther@suse.de>
8148 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8151 2014-02-19 Richard Biener <rguenther@suse.de>
8154 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8155 (ipa_modify_call_arguments): Emit an argument load explicitely and
8156 preserve virtual SSA form there and for the replacement call.
8157 Do not update SSA form nor free dominance info.
8159 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8161 * ipa.c (function_and_variable_visibility): Also clear WEAK
8162 flag when disolving COMDAT_GROUP.
8164 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8166 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8167 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8169 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8170 do more sanity checks.
8171 (detect_type_change): Return true when giving up early.
8172 (compute_complex_assign_jump_func): Fix type parameter of
8173 ipa_set_ancestor_jf.
8174 (compute_complex_ancestor_jump_func): Likewise.
8175 (update_jump_functions_after_inlining): Fix updating of
8177 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8179 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8181 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8182 inline clones when edge disappears.
8184 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8187 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8188 Split 64-bit moves into 2 patterns. Do not allow the use of
8189 direct move for TDmode in little endian, since the decimal value
8190 has little endian bytes within a word, but the 64-bit pieces are
8191 ordered in a big endian fashion, and normal subreg's of TDmode are
8193 (mov<mode>_64bit_dm): Likewise.
8194 (movtd_64bit_nodm): Likewise.
8196 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8198 PR tree-optimization/60174
8199 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8200 statement of an SSA_NAME that occurs in an abnormal PHI node.
8202 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8205 * final.c (SEEN_BB): Remove.
8206 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8207 (final_scan_insn): Don't force_source_line on second
8208 NOTE_INSN_BASIC_BLOCK.
8210 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8213 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8214 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8215 (type_natural_mode): Warn ABI change when %zmm register is not
8216 available for AVX512F vector value passing.
8218 2014-02-18 Kai Tietz <ktietz@redhat.com>
8221 * config/i386/i386.c (ix86_expand_prologue): Use value in
8222 rax register as displacement when restoring %r10 or %rax.
8223 Fix wrong offset when restoring both registers.
8225 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8227 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8228 assertion with conditional return.
8230 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8231 Uros Bizjak <ubizjak@gmail.com>
8234 * config/i386/driver-i386.c (host_detect_local_cpu): If
8235 YMM state is not saved by the OS, also clear has_f16c. Move
8236 CPUID 0x80000001 handling before YMM state saving checking.
8238 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8240 PR rtl-optimization/58960
8241 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8242 factored out from ...
8243 (sched_init): ... here.
8244 (free_global_sched_pressure_data): New, factored out from ...
8245 (sched_finish): ... here.
8246 * sched-int.h (free_global_sched_pressure_data): Declare.
8247 * sched-rgn.c (nr_regions_initial): New static global.
8248 (haifa_find_rgns): Initialize it.
8249 (schedule_region): Disable sched-pressure for the newly
8252 2014-02-17 Richard Biener <rguenther@suse.de>
8254 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8255 release SSA defs of pattern stmts.
8257 2014-02-17 Richard Biener <rguenther@suse.de>
8259 * tree-inline.c (expand_call_inline): Release the virtual
8260 operand defined by the call we are about to inline.
8262 2014-02-17 Richard Biener <rguenther@suse.de>
8264 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8266 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8267 Ilya Tocar <ilya.tocar@intel.com>
8269 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8270 arguments order in builtin.
8271 (_mm512_permutexvar_epi64): Ditto.
8272 (_mm512_mask_permutexvar_epi64): Ditto
8273 (_mm512_maskz_permutexvar_epi32): Ditto
8274 (_mm512_permutexvar_epi32): Ditto
8275 (_mm512_mask_permutexvar_epi32): Ditto
8277 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8279 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8280 (p8_vmrgow): Likewise.
8282 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8284 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8287 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8290 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8291 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8292 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8293 using direct move instructions on ISA 2.07. Also adjust
8294 instruction length for 64-bit.
8295 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8296 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8298 2014-02-15 Alan Modra <amodra@gmail.com>
8302 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8303 find_replacement on parts of insn rtl that might be reloaded.
8305 2014-02-15 Richard Biener <rguenther@suse.de>
8307 PR tree-optimization/60183
8308 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8309 (tree_ssa_phiprop): Calculate and free post-dominators.
8311 2014-02-14 Jeff Law <law@redhat.com>
8313 PR rtl-optimization/60131
8314 * ree.c (get_extended_src_reg): New function.
8315 (combine_reaching_defs): Use it rather than assuming location of REG.
8316 (find_and_remove_re): Verify first operand of extension is
8317 a REG before adding the insns to the copy list.
8319 2014-02-14 Roland McGrath <mcgrathr@google.com>
8321 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8322 * configure: Regenerated.
8323 * config.in: Regenerated.
8324 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8325 instead of ASM_SHORT.
8327 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8328 Richard Earnshaw <rearnsha@arm.com>
8330 PR rtl-optimization/59535
8331 * lra-constraints.c (process_alt_operands): Encourage alternative
8332 when unassigned pseudo class is superset of the alternative class.
8333 (inherit_reload_reg): Don't inherit when optimizing for code size.
8334 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8335 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8336 modes not less than 4 for Thumb1.
8338 2014-02-14 Kyle McMartin <kyle@redhat.com>
8341 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8343 2014-02-14 Richard Biener <rguenther@suse.de>
8345 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8346 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8347 Do not strip INDIRECT_REFs.
8349 2014-02-14 Richard Biener <rguenther@suse.de>
8352 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8353 DECL_FUNCTION_SPECIFIC_TARGET.
8354 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8355 * tree-streamer-out.c (pack_ts_target_option): Remove.
8356 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8357 (write_ts_function_decl_tree_pointers): Do not stream
8358 DECL_FUNCTION_SPECIFIC_TARGET.
8359 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8360 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8361 (lto_input_ts_function_decl_tree_pointers): Do not stream
8362 DECL_FUNCTION_SPECIFIC_TARGET.
8364 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8366 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8367 (get_initial_def_for_induction, vectorizable_induction): Ignore
8368 debug stmts when looking for exit_phi.
8369 (vectorizable_live_operation): Fix up condition.
8371 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8373 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8374 nreverse() because it changes the content of original tree list.
8376 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8378 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8379 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8381 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8383 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8384 GNU coding standards.
8386 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8389 * dwarf2out.c (gen_subprogram_die): Don't call
8390 add_calling_convention_attribute if subr_die is old_die.
8392 2014-02-13 Sharad Singhai <singhai@google.com>
8394 * doc/optinfo.texi: Fix order of nodes.
8396 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8398 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8399 operands[2], not operands[3].
8401 2014-02-13 Richard Biener <rguenther@suse.de>
8404 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8405 mention the possibility of an in-tree build.
8406 (CLooG): Update recommended version to 0.18.1, mention the
8407 possibility of an in-tree build and clarify that the ISL
8408 bundled with CLooG does not work.
8410 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8413 * expr.c (compress_float_constant): If x is a hard register,
8414 extend into a pseudo and then move to x.
8416 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8418 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8419 caused by bad second argument to warning_at() with -mhotpatch and
8420 nested functions (e.g. with gfortran).
8422 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8424 * opts.c (option_name): Remove "enabled by default" rider.
8426 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8428 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8430 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8431 Uros Bizjak <ubizjak@gmail.com>
8434 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8435 * configure: Regenerated.
8437 2014-02-12 Richard Biener <rguenther@suse.de>
8439 * vec.c (vec_prefix::calculate_allocation): Move as
8440 inline variant to vec.h.
8441 (vec_prefix::calculate_allocation_1): New out-of-line version.
8442 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8443 (vec_prefix::m_has_auto_buf): Rename to ...
8444 (vec_prefix::m_using_auto_storage): ... this.
8445 (vec_prefix::calculate_allocation): Inline the easy cases
8446 and dispatch to calculate_allocation_1 which doesn't need the
8448 (va_heap::reserve): Use gcc_checking_assert.
8449 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8450 m_using_auto_storage.
8451 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8453 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8454 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8455 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8457 2014-02-12 Richard Biener <rguenther@suse.de>
8459 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8460 when we found a dependence.
8462 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8464 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8466 (maybe_fold_stmt): ... into this new function.
8467 * omp-low.c (lower_omp): Update comment.
8469 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8472 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8475 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8477 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8478 identifiers in comments.
8479 (cortexa53_extra_costs): Likewise.
8480 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8481 (cortexa7_extra_costs): Likewise.
8482 (cortexa12_extra_costs): Likewise.
8483 (cortexa15_extra_costs): Likewise.
8484 (v7m_extra_costs): Likewise.
8486 2014-02-12 Richard Biener <rguenther@suse.de>
8489 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8490 of posix_memalign being successful.
8491 (lower_stmt): Restrict lowering of posix_memalign to when
8492 -ftree-bit-ccp is enabled.
8494 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8496 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8498 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8500 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8502 PR rtl-optimization/60116
8503 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8504 other_insn once the combination has been validated.
8506 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8509 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8511 * ipa-devirt.c: Include demangle.h
8512 (odr_violation_reported): New static variable.
8513 (add_type_duplicate): Update odr_violations.
8514 (maybe_record_node): Add completep parameter; update it.
8515 (record_target_from_binfo): Add COMPLETEP parameter;
8516 update it as needed.
8517 (possible_polymorphic_call_targets_1): Likewise.
8518 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8519 rename FINAL to COMPLETE.
8520 (record_targets_from_bases): Sanity check we found the binfo;
8521 fix COMPLETEP updating.
8522 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8523 parameter, fix computing of COMPLETEP.
8524 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8525 at LTO time do demangling.
8526 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8527 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8529 (gimple_get_virt_method_for_binfo): Likewise.
8530 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8531 gimple_get_virt_method_for_vtable): Update prototypes.
8533 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8536 * genautomata.c (add_presence_absence): Fix typo with
8537 {final_}presence_list.
8539 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8542 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8543 for VSX/Altivec vectors that land in GPR registers.
8545 2014-02-11 Richard Henderson <rth@redhat.com>
8546 Jakub Jelinek <jakub@redhat.com>
8549 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8550 around drhs if type conversion to lacc->type is not useless.
8552 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8554 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8556 (cortex-a57.cortex-a53): Likewise.
8557 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8559 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8561 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8564 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8566 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8567 add_options_for_arm_vfp3.
8569 2014-02-11 Jeff Law <law@redhat.com>
8572 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8573 object with an undesirable mode.
8575 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8578 * config/i386/sol2-9.h: New file.
8579 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8580 *-*-solaris2.9*): Use it.
8582 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8584 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8585 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8587 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8589 * config/microblaze/microblaze.c: Extend mcpu version format
8591 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8593 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8595 2014-02-10 Richard Henderson <rth@redhat.com>
8598 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8599 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8600 ms-abi vs -mno-accumulate-outgoing-args.
8601 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8602 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8605 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8608 * cfgexpand.c (expand_asm_operands): Attach source location to
8609 ASM_INPUT rtx objects.
8610 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8612 2014-02-10 Nick Clifton <nickc@redhat.com>
8614 * config/mn10300/mn10300.c (popcount): New function.
8615 (mn10300_expand_prologue): Include saved registers in stack usage
8618 2014-02-10 Jeff Law <law@redhat.com>
8621 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8622 when changing the SET_DEST of a prior insn to avoid an input reload.
8624 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8626 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8627 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8628 -mcall-openbsd, or -mcall-linux.
8629 (CC1_ENDIAN_BIG_SPEC): Remove.
8630 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8631 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8632 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8633 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8634 and %cc1_endian_default.
8635 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8637 2014-02-10 Richard Biener <rguenther@suse.de>
8639 PR tree-optimization/60115
8640 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8641 MEM_REF handling. Properly verify that the accesses are not
8642 out of the objects bound.
8644 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8646 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8649 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8651 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8652 proper constants and fix formatting.
8653 (possible_polymorphic_call_targets): Fix formatting.
8655 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8656 Ilya Tocar <ilya.tocar@intel.com>
8658 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8659 (_mm512_loadu_epi32): Renamed into...
8660 (_mm512_loadu_si512): This.
8661 (_mm512_storeu_epi32): Renamed into...
8662 (_mm512_storeu_si512): This.
8663 (_mm512_maskz_ceil_ps): Removed.
8664 (_mm512_maskz_ceil_pd): Ditto.
8665 (_mm512_maskz_floor_ps): Ditto.
8666 (_mm512_maskz_floor_pd): Ditto.
8667 (_mm512_floor_round_ps): Ditto.
8668 (_mm512_floor_round_pd): Ditto.
8669 (_mm512_ceil_round_ps): Ditto.
8670 (_mm512_ceil_round_pd): Ditto.
8671 (_mm512_mask_floor_round_ps): Ditto.
8672 (_mm512_mask_floor_round_pd): Ditto.
8673 (_mm512_mask_ceil_round_ps): Ditto.
8674 (_mm512_mask_ceil_round_pd): Ditto.
8675 (_mm512_maskz_floor_round_ps): Ditto.
8676 (_mm512_maskz_floor_round_pd): Ditto.
8677 (_mm512_maskz_ceil_round_ps): Ditto.
8678 (_mm512_maskz_ceil_round_pd): Ditto.
8679 (_mm512_expand_pd): Ditto.
8680 (_mm512_expand_ps): Ditto.
8681 * config/i386/i386.c (ix86_builtins): Remove
8682 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8683 (bdesc_args): Ditto.
8684 * config/i386/predicates.md (const1256_operand): New.
8685 (const_1_to_2_operand): Ditto.
8686 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8687 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8688 (*avx512pf_gatherpf<mode>sf): Ditto.
8689 (avx512pf_gatherpf<mode>df): Ditto.
8690 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8691 (*avx512pf_gatherpf<mode>df): Ditto.
8692 (avx512pf_scatterpf<mode>sf): Ditto.
8693 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8694 (*avx512pf_scatterpf<mode>sf): Ditto.
8695 (avx512pf_scatterpf<mode>df): Ditto.
8696 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8697 (*avx512pf_scatterpf<mode>df): Ditto.
8698 (avx512f_expand<mode>): Removed.
8699 (<shift_insn><mode>3<mask_name>): Change predicate type.
8701 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8703 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8704 not at the end of datarefs vector use ordered_remove to avoid
8705 reordering datarefs vector.
8708 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8709 mark local addressable non-static vars as GOVD_PRIVATE
8710 instead of GOVD_LOCAL.
8711 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8712 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8716 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8717 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8718 assume_aligned or alloc_align attributes.
8719 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8720 arguments. Handle also assume_aligned and alloc_align attributes.
8721 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8722 calls to functions with assume_aligned or alloc_align attributes.
8723 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8725 2014-02-08 Terry Guo <terry.guo@arm.com>
8727 * doc/invoke.texi: Document ARM -march=armv7e-m.
8729 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8731 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8732 flag on __cilkrts_rethrow builtin.
8735 * ipa-cp.c (determine_versionability): Fail at -O0
8736 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8737 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8740 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8743 * tree-inline.c (copy_forbidden): Fail for
8744 __attribute__((optimize (0))) functions.
8746 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8748 * varpool.c: Include pointer-set.h.
8749 (varpool_remove_unreferenced_decls): Variables in other partitions
8750 will not be output; be however careful to not lose information
8753 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8755 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8756 lookup in the vtable constructor.
8758 2014-02-07 Jeff Law <law@redhat.com>
8761 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8762 define_insn_and_split.
8764 * ipa-inline.c (inline_small_functions): Fix typos.
8766 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8768 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8769 (s390_can_use_return_insn): Declare.
8770 * config/s390/s390.h (EPILOGUE_USES): Define.
8771 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8773 (s390_chunkify_start): Handle return JUMP_LABELs.
8774 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8775 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8776 (s390_can_use_return_insn): New functions.
8777 (s390_fix_long_loop_prediction): Handle conditional returns.
8778 (TARGET_SET_UP_BY_PROLOGUE): Define.
8779 * config/s390/s390.md (ANY_RETURN): New code iterator.
8780 (*creturn, *csimple_return, return, simple_return): New patterns.
8782 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8784 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8785 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8786 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8787 REG_CFA_RESTORE list when deciding not to restore a register.
8789 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8791 * config/s390/s390.c: Include tree-pass.h and context.h.
8792 (s390_early_mach): New function, split out from...
8793 (s390_emit_prologue): ...here.
8794 (pass_data_s390_early_mach): New pass structure.
8795 (pass_s390_early_mach): New class.
8796 (s390_option_override): Create and register early_mach pass.
8797 Move to end of file.
8799 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8801 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8802 to match for the exit block.
8804 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8806 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8807 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8808 Reject misaligned operands.
8810 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8812 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8814 2014-02-07 Richard Biener <rguenther@suse.de>
8817 * gimple-low.c (lower_builtin_posix_memalign): New function.
8818 (lower_stmt): Call it to lower posix_memalign in a way
8819 to make alignment info accessible.
8821 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8824 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8825 __builtin_setjmp_receiver.
8827 2014-02-07 Richard Biener <rguenther@suse.de>
8830 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8831 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8832 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8833 Handle BUILT_IN_POSIX_MEMALIGN.
8834 (find_func_clobbers): Likewise.
8835 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8836 (call_may_clobber_ref_p_1): Likewise.
8838 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8841 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8844 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8847 * lto-cgraph.c (lto_output_node): Use
8848 symtab_get_symbol_partitioning_class.
8849 (lto_output_varpool_node): likewise.
8850 (symtab_get_symbol_partitioning_class): Move here from
8852 * cgraph.h (symbol_partitioning_class): Likewise.
8853 (symtab_get_symbol_partitioning_class): Declare.
8855 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8857 * ggc.h (ggc_internal_cleared_alloc): New macro.
8858 * vec.h (vec_safe_copy): Handle memory stats.
8859 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8860 * target-globals.c (save_target_globals): Likewise.
8862 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8865 * expr.c (emit_move_resolve_push): Export; be bit more selective
8866 on when to clear alias set.
8867 * expr.h (emit_move_resolve_push): Declare.
8868 * function.h (struct function): Add tail_call_marked.
8869 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8870 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8871 * config/i386/i386.md (TImode move expander): De not call
8873 (FP push expanders): Preserve memory attributes.
8874 * config/i386/sse.md (push<mode>1): Remove.
8875 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8876 (ix86_expand_push): Remove.
8877 * config/i386/mmx.md (push<mode>1): Remove.
8879 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8881 PR rtl-optimization/60030
8882 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8883 lopart with paradoxical subreg before shifting it up by hprec.
8885 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8887 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8888 Remove extra newline at end of file.
8889 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8890 (arm_issue_rate): Handle cortexa57.
8891 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8892 (cortex-a57.cortex-a53): Likewise.
8894 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8897 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8898 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8900 (arm_expand_prologue): Adjust all callers.
8901 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8902 info, registers also at the lowest numbered registers side. Use
8903 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8907 * var-tracking.c (adjust_mems): Before adding a SET to
8908 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8910 2014-02-06 Alan Modra <amodra@gmail.com>
8913 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8914 change SDmode to DDmode when lra_in_progress.
8916 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8919 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8920 free_data_ref on the dr first, and before goto again also set dr
8921 to the next dr. For simd_lane_access, free old datarefs[i] before
8922 overwriting it. For get_vectype_for_scalar_type failure, don't
8923 free_data_ref if simd_lane_access.
8925 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8928 * tree.h (opts_for_fn): New inline function.
8929 (opt_for_fn): Define.
8930 * config/i386/i386.c (ix86_function_regparm): Use
8931 opt_for_fn (decl, optimize) instead of optimize.
8933 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8935 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8936 for SYMBOL_REF in large memory model.
8938 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8940 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8942 (cortex-a57): Likewise.
8943 (cortex-a57.cortex-a53): Likewise.
8945 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8946 Kugan Vivekanandarajah <kuganv@linaro.org>
8948 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8950 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8952 2014-02-06 Richard Biener <rguenther@suse.de>
8954 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8955 set_loop_copy and initialize_original_copy_tables.
8957 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8959 * config/aarch64/aarch64-simd.md
8960 (aarch64_ashr_simddi): Change QI to SI.
8962 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8963 Jakub Jelinek <jakub@redhat.com>
8966 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8969 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8971 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8972 CODE_FOR_altivec_vpku[hw]um to
8973 CODE_FOR_altivec_vpku[hw]um_direct.
8974 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8975 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8976 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8977 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8979 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8981 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8982 generation for -maltivec=be.
8983 (altivec_vsumsws): Simplify redundant test.
8985 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8987 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8988 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8989 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8990 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8991 gen_altivec_vpkuwum.
8992 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8994 (altivec_vpks<VI_char>ss): Likewise.
8995 (altivec_vpks<VI_char>us): Likewise.
8996 (altivec_vpku<VI_char>us): Likewise.
8997 (altivec_vpku<VI_char>um): Likewise.
8998 (altivec_vpku<VI_char>um_direct): New (copy of
8999 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9001 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9002 target is little endian and -maltivec=be is not specified.
9003 (*altivec_vupkhs<VU_char>_direct): New (copy of
9004 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9005 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9006 target is little endian and -maltivec=be is not specified.
9007 (*altivec_vupkls<VU_char>_direct): New (copy of
9008 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9009 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9010 little endian and -maltivec=be is not specified.
9011 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9012 little endian and -maltivec=be is not specified.
9014 2014-02-05 Richard Henderson <rth@redhat.com>
9017 * combine-stack-adj.c: Revert r206943.
9018 * sched-int.h (struct deps_desc): Add last_args_size.
9019 * sched-deps.c (init_deps): Initialize it.
9020 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9021 contain REG_ARGS_SIZE notes.
9023 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9025 * lto-cgraph.c (asm_nodes_output): Make global.
9026 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9027 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9028 (driver_handle_option): Handle OPT_fwpa.
9030 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9033 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9034 a comment typo and formatting issue. If odr_hash hasn't been
9035 created, return vNULL and set *completep to false.
9038 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9039 bb with no successors.
9041 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9044 * doc/invoke.texi (-march): Clarify documentation for ARM.
9048 2014-02-05 Richard Biener <rguenther@suse.de>
9050 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9051 when not vectorizing because of too many alias checks.
9052 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9053 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9055 2014-02-05 Nick Clifton <nickc@redhat.com>
9057 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9058 accept extended registers in any mode when compiling for the MN10300.
9060 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9062 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9063 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9064 sanitization attributes.
9065 (can_inline_edge_p): Likewise.
9066 (sanitize_attrs_match_for_inline_p): New function.
9068 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9070 * ipa-prop.c (detect_type_change): Shor circuit testing of
9071 type changes on THIS pointer.
9073 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9076 * config/pa/pa.c (legitimize_tls_address): Return original address
9077 if not passed a SYMBOL_REF rtx.
9078 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9080 (pa_emit_move_sequence): Simplify TLS source operands.
9081 (pa_legitimate_constant_p): Reject all TLS constants.
9082 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9083 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9085 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9087 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9088 groups when we know they are controlled by LTO.
9089 * varasm.c (default_binds_local_p_1): If object is in other partition,
9090 it will be resolved locally.
9092 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9094 * config/host-linux.c (linux_gt_pch_use_address): Don't
9095 use SSIZE_MAX because it is not always defined.
9097 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9100 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9101 threshold for pseudo splitting.
9102 (update_ebb_live_info): Process call argument hard registers and
9103 hard registers from insn definition too.
9104 (max_small_class_regs_num): New constant.
9105 (inherit_in_ebb): Update live hard regs through EBBs. Update
9106 reloads_num only for small register classes. Don't split for
9109 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9112 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9115 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9117 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9120 2014-02-04 Marek Polacek <polacek@redhat.com>
9122 * gdbinit.in (pel): Define.
9124 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9126 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9129 2014-02-04 Richard Biener <rguenther@suse.de>
9132 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9133 in function context local.
9134 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9135 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9136 similar to LTO_imported_decl_ref.
9138 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9140 PR tree-optimization/60002
9141 * cgraphclones.c (build_function_decl_skip_args): Clear
9144 PR tree-optimization/60023
9145 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9146 false to gsi_replace.
9147 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9148 has been in some EH region and vec_stmt could throw, add
9149 vec_stmt into the same EH region.
9150 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9151 has no lhs, ignore it.
9152 * internal-fn.c (expand_MASK_LOAD): Likewise.
9155 * tree-inline.c (copy_forbidden): Fail for
9156 __attribute__((optimize (0))) functions.
9159 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9160 is set, copy one less argument.
9161 (expand_simd_clones): Don't subtract clone_info->inbranch
9162 from simd_clone_struct_alloc argument.
9164 PR rtl-optimization/57915
9165 * recog.c (simplify_while_replacing): If all unary/binary/relational
9166 operation arguments are constant, attempt to simplify those.
9169 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9170 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9172 2014-02-04 Richard Biener <rguenther@suse.de>
9174 PR tree-optimization/60012
9175 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9176 TBAA disambiguation to all DDRs.
9178 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9181 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9182 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9184 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9187 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9189 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9191 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9192 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9194 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9197 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9198 to figure out targets of polymorphic calls with known decl.
9199 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9200 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9201 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9202 (get_polymorphic_call_info): ... here.
9203 (get_polymorphic_call_info_from_invariant): New function.
9205 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9207 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9208 lookup via vtable pointer; check for type consistency
9209 and turn inconsitent facts into UNREACHABLE.
9210 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9211 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9212 type inconsistent querries; return UNREACHABLE instead.
9214 2014-02-03 Richard Henderson <rth@twiddle.net>
9217 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9218 already processed this node.
9219 (normalize_one_pred_1): Pass along mark_set.
9220 (normalize_one_pred): Create and destroy a pointer_set_t.
9221 (normalize_one_pred_chain): Likewise.
9223 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9225 PR gcov-profile/58602
9226 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9228 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9231 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9232 -fno-devirtualize; try to devirtualize by the knowledge of
9233 virtual table pointer given by aggregate propagation.
9234 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9235 (ipa_print_node_jump_functions): Dump also offset that
9236 is relevant for polymorphic calls.
9237 (determine_known_aggregate_parts): Add arg_type parameter; use it
9238 instead of determining the type from pointer type.
9239 (ipa_compute_jump_functions_for_edge): Update call of
9240 determine_known_aggregate_parts.
9241 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9242 (gimple_get_virt_method_for_binfo): ... here; simplify using
9243 vtable_pointer_value_to_vtable.
9244 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9245 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9246 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9247 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9249 (vtable_pointer_value_to_binfo): ... here.
9250 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9252 2014-02-03 Teresa Johnson <tejohnson@google.com>
9254 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9255 redef of outer loop index variable.
9257 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9261 * doc/extend.texi (Function Attributes): Typo.
9263 2014-02-03 Cong Hou <congh@google.com>
9265 PR tree-optimization/60000
9266 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9267 if the vectorized statement is a store. A store statement can only
9268 appear at the end of pattern statements.
9270 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9272 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9273 (ix86_option_override_internal): Default long double to 64-bit for
9274 32-bit Bionic and to 128-bit for 64-bit Bionic.
9276 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9277 TARGET_LONG_DOUBLE_128 is true.
9278 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9280 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9281 (mlong-double-64): Negate -mlong-double-128.
9282 (mlong-double-128): New option.
9284 * config/i386/i386-c.c (ix86_target_macros): Define
9285 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9287 * doc/invoke.texi: Document -mlong-double-128.
9289 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9291 PR rtl-optimization/60024
9292 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9294 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9296 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9298 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9300 PR rtl-optimization/57662
9301 * sel-sched.c (code_motion_path_driver): Do not mark already not
9302 existing blocks in the visiting bitmap.
9304 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9306 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9307 on the insn being emitted.
9309 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9310 Will Deacon <will.deacon@arm.com>
9312 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9314 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9316 * config/arm/arm-tables.opt: Regenerate.
9318 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9320 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9321 for vector types other than V16QImode.
9322 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9323 define_expand, and call altivec_expand_vec_perm_le when producing
9324 code with little endian element order.
9325 (*altivec_vperm_<mode>_internal): New insn having previous
9326 behavior of altivec_vperm_<mode>.
9327 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9328 altivec_expand_vec_perm_le when producing code with little endian
9330 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9331 behavior of altivec_vperm_<mode>_uns.
9333 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9335 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9336 (altivec_vsumsws): Add handling for -maltivec=be with a little
9338 (altivec_vsumsws_direct): New.
9339 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9340 gen_altivec_vsumsws.
9342 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9344 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9345 vtable_pointer_value_to_binfo): New functions.
9346 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9347 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9349 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9351 * config/nios2/nios2.md (load_got_register): Initialize GOT
9352 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9353 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9355 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9357 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9358 preserverd by passthrough, do not propagate the type.
9360 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9362 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9363 (mips_atomic_assign_expand_fenv): New function.
9364 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9366 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9368 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9369 (__builtin_mips_set_fcsr): Likewise.
9370 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9371 MIPS_USI_FTYPE_VOID.
9372 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9373 (mips16_expand_set_fcsr): Likewise.
9374 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9375 (mips16_set_fcsr_stub): Likewise.
9376 (mips16_get_fcsr_one_only_stub): New class.
9377 (mips16_set_fcsr_one_only_stub): Likewise.
9378 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9379 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9380 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9381 (hard_float): New availability predicate.
9382 (mips_builtins): Add get_fcsr and set_fcsr.
9383 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9384 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9385 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9386 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9387 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9390 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9392 * config/mips/mips.c (mips_one_only_stub): New class.
9393 (mips_need_mips16_rdhwr_p): Replace with...
9394 (mips16_rdhwr_stub): ...this new variable.
9395 (mips16_stub_call_address): New function.
9396 (mips16_rdhwr_one_only_stub): New class.
9397 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9398 (mips_output_mips16_rdhwr): Delete.
9399 (mips_finish_stub): New function.
9400 (mips_code_end): Use it to handle rdhwr stubs.
9402 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9405 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9406 when calculating size of integer atomic types.
9408 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9410 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9412 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9414 PR tree-optimization/60003
9415 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9416 * profile.c (branch_prob): Use gimple_call_builtin_p
9417 to check for BUILT_IN_SETJMP_RECEIVER.
9418 * tree-inline.c (copy_bb): Call notice_special_calls.
9420 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9423 * lra-constraints.c (process_alt_operands): Update reload_sum only
9426 2014-01-31 Richard Henderson <rth@redhat.com>
9429 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9430 until after else_eh is processed.
9432 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9434 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9435 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9436 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9437 in smmintrin.h, remove them.
9438 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9439 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9440 * config/i386/i386.md (ROUND_SAE): Fix value.
9441 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9442 (const48_operand): New.
9443 * config/i386/subst.md (round), (round_expand): Use
9444 const_4_or_8_to_11_operand.
9445 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9447 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9449 * config/i386/constraints.md (Yk): Swap meaning with k.
9450 * config/i386/i386.md (movhi_internal): Change Yk to k.
9451 (movqi_internal): Ditto.
9452 (*k<logic><mode>): Ditto.
9455 (kandn<mode>): Ditto.
9456 (*<code>hi_1): Ditto.
9457 (*<code>qi_1): Ditto.
9458 (kxnor<mode>): Ditto.
9459 (kortestzhi): Ditto.
9460 (kortestchi): Ditto.
9462 (*one_cmplhi2_1): Ditto.
9463 (*one_cmplqi2_1): Ditto.
9464 * config/i386/sse.md (): Change k to Yk.
9465 (avx512f_load<mode>_mask): Ditto.
9466 (avx512f_blendm<mode>): Ditto.
9467 (avx512f_store<mode>_mask): Ditto.
9468 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9469 (avx512f_storedqu<mode>_mask): Ditto.
9470 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9472 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9473 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9474 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9475 (avx512f_maskcmp<mode>3): Ditto.
9476 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9477 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9478 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9479 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9480 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9481 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9482 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9483 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9484 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9485 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9486 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9487 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9488 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9489 (vec_extract_lo_<mode>_maskm): Ditto.
9490 (vec_extract_hi_<mode>_maskm): Ditto.
9491 (avx512f_vternlog<mode>_mask): Ditto.
9492 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9493 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9494 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9495 (avx512f_<code>v8div16qi2_mask): Ditto.
9496 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9497 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9498 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9499 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9500 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9501 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9502 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9503 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9504 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9505 (avx512cd_maskb_vec_dupv8di): Ditto.
9506 (avx512cd_maskw_vec_dupv16si): Ditto.
9507 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9508 (avx512f_vpermi2var<mode>3_mask): Ditto.
9509 (avx512f_vpermi2var<mode>3_mask): Ditto.
9510 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9511 (*avx512f_gathersi<mode>): Ditto.
9512 (*avx512f_gathersi<mode>_2): Ditto.
9513 (*avx512f_gatherdi<mode>): Ditto.
9514 (*avx512f_gatherdi<mode>_2): Ditto.
9515 (*avx512f_scattersi<mode>): Ditto.
9516 (*avx512f_scatterdi<mode>): Ditto.
9517 (avx512f_compress<mode>_mask): Ditto.
9518 (avx512f_compressstore<mode>_mask): Ditto.
9519 (avx512f_expand<mode>_mask): Ditto.
9520 * config/i386/subst.md (mask): Change k to Yk.
9521 (mask_scalar_merge): Ditto.
9524 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9526 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9528 2014-01-31 Richard Biener <rguenther@suse.de>
9531 * builtins.c (fold_builtin_memory_op): Make sure to not
9532 use a floating-point mode or a boolean or enumeral type for
9535 2014-01-30 DJ Delorie <dj@redhat.com>
9537 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9538 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9539 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9540 whenever main() has an epilogue.
9542 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9544 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9545 unused variable "field".
9546 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9547 (vsx_mergeh_<mode>): Likewise.
9548 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9549 (altivec_vmrghh): Likewise.
9550 (altivec_vmrghw): Likewise.
9551 (altivec_vmrglb): Likewise.
9552 (altivec_vmrglh): Likewise.
9553 (altivec_vmrglw): Likewise.
9554 (altivec_vspltb): Add missing uses.
9555 (altivec_vsplth): Likewise.
9556 (altivec_vspltw): Likewise.
9557 (altivec_vspltsf): Likewise.
9559 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9562 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9563 frame related instructions.
9565 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9567 PR rtl-optimization/59959
9568 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9569 any reload of register whose subreg is invalid.
9571 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9573 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9574 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9575 Add missing return type - void.
9577 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9579 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9580 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9581 remove element index adjustment for endian (now handled in vsx.md
9583 (altivec_expand_vec_perm_const): Use
9584 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9585 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9586 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9587 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9588 define_expand and a new define_insn *altivec_vspltb_internal;
9589 adjust for -maltivec=be on a little endian target.
9590 (altivec_vspltb_direct): New.
9591 (altivec_vsplth): Divide into a define_expand and a new
9592 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9593 little endian target.
9594 (altivec_vsplth_direct): New.
9595 (altivec_vspltw): Divide into a define_expand and a new
9596 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9597 little endian target.
9598 (altivec_vspltw_direct): New.
9599 (altivec_vspltsf): Divide into a define_expand and a new
9600 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9601 a little endian target.
9603 2014-01-30 Richard Biener <rguenther@suse.de>
9605 PR tree-optimization/59993
9606 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9607 can propagate form the earlier stmt and avoid the transform
9608 when the intermediate result is needed.
9610 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9612 * README.Portability: Fix typo.
9614 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9616 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9617 comparison_operator with ordered_comparison_operator.
9619 2014-01-30 Nick Clifton <nickc@redhat.com>
9621 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9622 Rename to mn10300_store_multiple_regs.
9623 * config/mn10300/mn10300.c: Likewise.
9624 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9625 store_multiple_regs.
9626 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9627 Call mn10300_store_multiple_regs.
9629 2014-01-30 Nick Clifton <nickc@redhat.com>
9630 DJ Delorie <dj@redhat.com>
9632 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9633 %fp 2 to keep registers after it properly word-aligned.
9634 (rl78_alloc_physical_registers_umul): Handle the case where both
9635 input operands are the same.
9637 2014-01-30 Richard Biener <rguenther@suse.de>
9639 PR tree-optimization/59903
9640 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9643 2014-01-30 Jason Merrill <jason@redhat.com>
9646 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9649 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9651 2014-01-30 Richard Biener <rguenther@suse.de>
9653 PR tree-optimization/59951
9654 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9656 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9659 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9660 SFmode to DFmode case.
9662 2014-01-29 DJ Delorie <dj@redhat.com>
9664 * config/msp430/msp430.opt (-minrt): New.
9665 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9667 (ENDFILE_SPEC): Likewise.
9669 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9671 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9672 (estimate_function_body_sizes): Use it.
9674 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9677 * dwarf2out.c (is_cxx_auto): New.
9678 (is_base_type): Use it.
9679 (gen_type_die_with_usage): Likewise.
9681 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9683 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9684 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9685 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9686 -maltivec=be with LE targets.
9687 (vsx_mergeh_<mode>): Likewise.
9688 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9689 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9690 (altivec_vmrghb): Replace with define_expand and new
9691 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9692 (altivec_vmrghb_direct): New define_insn.
9693 (altivec_vmrghh): Replace with define_expand and new
9694 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9695 (altivec_vmrghh_direct): New define_insn.
9696 (altivec_vmrghw): Replace with define_expand and new
9697 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9698 (altivec_vmrghw_direct): New define_insn.
9699 (*altivec_vmrghsf): Adjust for endianness.
9700 (altivec_vmrglb): Replace with define_expand and new
9701 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9702 (altivec_vmrglb_direct): New define_insn.
9703 (altivec_vmrglh): Replace with define_expand and new
9704 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9705 (altivec_vmrglh_direct): New define_insn.
9706 (altivec_vmrglw): Replace with define_expand and new
9707 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9708 (altivec_vmrglw_direct): New define_insn.
9709 (*altivec_vmrglsf): Adjust for endianness.
9710 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9711 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9712 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9713 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9714 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9715 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9716 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9717 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9719 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9721 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9722 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9725 2014-01-29 Richard Biener <rguenther@suse.de>
9727 PR tree-optimization/58742
9728 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9729 associate_pointerplus_align.
9730 (associate_pointerplus_diff): New function.
9731 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9732 and associate_pointerplus_diff.
9734 2014-01-29 Richard Biener <rguenther@suse.de>
9736 * lto-streamer.h (LTO_major_version): Bump to 3.
9737 (LTO_minor_version): Reset to 0.
9739 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9741 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9742 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9743 (arm_file_start): Generate correct asm header for armv7ve.
9744 * config/arm/bpabi.h: Add multilib support for armv7ve.
9745 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9746 and cortex-a15 to armv7ve.
9747 * config/arm/t-aprofile: Add multilib support for armv7ve.
9748 * doc/invoke.texi: Document -march=armv7ve.
9750 2014-01-29 Richard Biener <rguenther@suse.de>
9752 PR tree-optimization/58742
9753 * tree-ssa-forwprop.c (associate_plusminus): Return true
9754 if we changed sth, defer EH cleanup to ...
9755 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9756 (simplify_mult): New function.
9758 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9761 PR tree-optimization/59920
9762 * tree.c (build_common_builtin_nodes): Remove
9763 __builtin_setjmp_dispatcher initialization.
9764 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9765 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9766 instead of gsi_after_labels + manually skipping debug stmts.
9767 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9768 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9769 * tree-inline.c (copy_edges_for_bb): Remove
9770 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9771 argument. Ignore computed_goto_p stmts. Don't call
9772 make_abnormal_goto_edges. If a call might need abnormal edges
9773 for non-local gotos, see if it already has an edge to
9774 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9775 with true argument, don't do anything then, otherwise add
9776 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9777 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9779 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9780 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9781 (lower_stmt): Don't set data->calls_builtin_setjmp.
9782 (lower_builtin_setjmp): Adjust comment.
9783 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9784 * tree-cfg.c (found_computed_goto): Remove.
9785 (factor_computed_gotos): Remove.
9786 (make_goto_expr_edges): Return bool, true for computed gotos.
9787 Don't call make_abnormal_goto_edges.
9788 (build_gimple_cfg): Don't set found_computed_goto, don't call
9789 factor_computed_gotos.
9790 (computed_goto_p): No longer static.
9791 (make_blocks): Don't set found_computed_goto.
9792 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9793 (make_edges): If make_goto_expr_edges returns true, push bb
9794 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9795 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9796 vector. Record mapping between bbs and OpenMP regions if there
9797 are any, adjust make_gimple_omp_edges caller. Call
9798 handle_abnormal_edges.
9799 (make_abnormal_goto_edges): Remove.
9800 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9801 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9802 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9803 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9804 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9805 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9806 filling *region also set *region_idx to (*region)->entry->index.
9809 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9810 For REGs set ORIGINAL_REGNO.
9812 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9814 * doc/md.texi: Mention that a target shouldn't implement
9815 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9818 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9820 PR tree-optimization/59594
9821 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9822 a copy of the datarefs vector rather than the vector itself.
9824 2014-01-28 Jason Merrill <jason@redhat.com>
9827 * dwarf2out.c (auto_die): New static.
9828 (gen_type_die_with_usage): Handle C++1y 'auto'.
9829 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9832 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9835 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9836 (SPEC_X32): Likewise.
9837 (SPEC_64): Likewise.
9838 * config/i386/i386.c (ix86_option_override_internal): Turn off
9839 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9841 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9842 * config/i386/i386.h (TARGET_16BIT): New macro.
9843 (TARGET_16BIT_P): Likewise.
9844 * config/i386/i386.opt: Add m16.
9845 * doc/invoke.texi: Document -m16.
9847 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9849 PR preprocessor/59935
9850 * input.c (location_get_source_line): Bail out on when line number
9851 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9853 2014-01-28 Richard Biener <rguenther@suse.de>
9855 PR tree-optimization/58742
9856 * tree-ssa-forwprop.c (associate_plusminus): Handle
9857 pointer subtraction of the form (T)(P + A) - (T)P.
9859 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9861 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9864 2014-01-28 Richard Biener <rguenther@suse.de>
9867 2014-01-28 Richard Biener <rguenther@suse.de>
9869 PR rtl-optimization/45364
9870 PR rtl-optimization/59890
9871 * var-tracking.c (local_get_addr_clear_given_value): Handle
9872 already cleared slot.
9873 (val_reset): Handle not allocated local_get_addr_cache.
9874 (vt_find_locations): Use post-order on the inverted CFG.
9876 2014-01-28 Richard Biener <rguenther@suse.de>
9878 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9880 2014-01-28 Richard Biener <rguenther@suse.de>
9882 PR rtl-optimization/45364
9883 PR rtl-optimization/59890
9884 * var-tracking.c (local_get_addr_clear_given_value): Handle
9885 already cleared slot.
9886 (val_reset): Handle not allocated local_get_addr_cache.
9887 (vt_find_locations): Use post-order on the inverted CFG.
9889 2014-01-28 Alan Modra <amodra@gmail.com>
9891 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9892 * configure.ac <recursive call for build != host>: Define
9893 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9894 and LD_FOR_BUILD too.
9895 * configure: Regenerate.
9897 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9899 * config/i386/i386.c (get_builtin_code_for_version): Separate
9900 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9901 Broadwell from Haswell.
9903 2014-01-27 Steve Ellcey <sellcey@mips.com>
9905 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9906 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9907 * config/mips/mips.c (mips_option_override): Change setting
9909 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9910 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9911 Change from Mask to Var.
9913 2014-01-27 Jeff Law <law@redhat.com>
9915 * ipa-inline.c (inline_small_functions): Fix typo.
9917 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9919 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9920 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9921 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9922 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9923 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9924 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9925 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9926 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9927 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9928 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9929 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9930 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9931 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9932 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9933 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9934 (_mm512_storeu_epi64): Ditto.
9935 (_mm512_cmpge_epi32_mask): Ditto.
9936 (_mm512_cmpge_epu32_mask): Ditto.
9937 (_mm512_cmpge_epi64_mask): Ditto.
9938 (_mm512_cmpge_epu64_mask): Ditto.
9939 (_mm512_cmple_epi32_mask): Ditto.
9940 (_mm512_cmple_epu32_mask): Ditto.
9941 (_mm512_cmple_epi64_mask): Ditto.
9942 (_mm512_cmple_epu64_mask): Ditto.
9943 (_mm512_cmplt_epi32_mask): Ditto.
9944 (_mm512_cmplt_epu32_mask): Ditto.
9945 (_mm512_cmplt_epi64_mask): Ditto.
9946 (_mm512_cmplt_epu64_mask): Ditto.
9947 (_mm512_cmpneq_epi32_mask): Ditto.
9948 (_mm512_cmpneq_epu32_mask): Ditto.
9949 (_mm512_cmpneq_epi64_mask): Ditto.
9950 (_mm512_cmpneq_epu64_mask): Ditto.
9951 (_mm512_expand_pd): Ditto.
9952 (_mm512_expand_ps): Ditto.
9953 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9954 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9955 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9956 * config/i386/i386.c (ix86_builtins): Add
9957 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9958 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9959 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9960 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9961 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9962 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9963 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9964 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9965 IX86_BUILTIN_PMOVUSQW512_MEM.
9966 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9967 __builtin_ia32_pmovsqd512mem_mask,
9968 __builtin_ia32_pmovqd512mem_mask,
9969 __builtin_ia32_pmovusqw512mem_mask,
9970 __builtin_ia32_pmovsqw512mem_mask,
9971 __builtin_ia32_pmovqw512mem_mask,
9972 __builtin_ia32_pmovusdw512mem_mask,
9973 __builtin_ia32_pmovsdw512mem_mask,
9974 __builtin_ia32_pmovdw512mem_mask,
9975 __builtin_ia32_pmovqb512mem_mask,
9976 __builtin_ia32_pmovusqb512mem_mask,
9977 __builtin_ia32_pmovsqb512mem_mask,
9978 __builtin_ia32_pmovusdb512mem_mask,
9979 __builtin_ia32_pmovsdb512mem_mask,
9980 __builtin_ia32_pmovdb512mem_mask.
9981 (bdesc_args): Add __builtin_ia32_expanddf512,
9982 __builtin_ia32_expandsf512.
9983 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9984 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9985 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9986 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9987 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9988 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9989 (avx512f_<code>v8div16qi2_mask_store): This.
9990 (avx512f_expand<mode>): New.
9992 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9994 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9996 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9997 (_mm512_prefetch_i32scatter_pd): Ditto.
9998 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9999 (_mm512_prefetch_i64scatter_pd): Ditto.
10000 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10001 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10002 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10003 (_mm512_prefetch_i32scatter_ps): Ditto.
10004 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10005 (_mm512_prefetch_i64scatter_ps): Ditto.
10006 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10007 * config/i386/i386-builtin-types.def: Define
10008 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10009 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10010 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10011 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10012 IX86_BUILTIN_SCATTERPFQPD.
10013 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10014 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10015 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10016 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10017 __builtin_ia32_scatterpfqps.
10018 (ix86_expand_builtin): Expand new built-ins.
10019 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10020 fix memory access data type.
10021 (*avx512pf_gatherpf<mode>_mask): Ditto.
10022 (*avx512pf_gatherpf<mode>): Ditto.
10023 (avx512pf_scatterpf<mode>): Ditto.
10024 (*avx512pf_scatterpf<mode>_mask): Ditto.
10025 (*avx512pf_scatterpf<mode>): Ditto.
10026 (GATHER_SCATTER_SF_MEM_MODE): New.
10027 (avx512pf_gatherpf<mode>df): Ditto.
10028 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10029 (*avx512pf_scatterpf<mode>df): Ditto.
10031 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10034 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10035 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10038 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10040 * common/config/arm/arm-common.c
10041 (arm_rewrite_mcpu): Handle multiple names.
10043 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10045 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10047 * gimple-builder.h (create_gimple_tmp): Delete.
10049 2014-01-27 Christian Bruel <christian.bruel@st.com>
10051 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10054 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10056 * config/pa/pa.md (call): Generate indirect long calls to non-local
10057 functions when outputing 32-bit code.
10058 (call_value): Likewise except for special call to buggy powf function.
10060 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10061 portable runtime and PIC indirect calls.
10062 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10063 and PIC call sequences. Use ldo instead of blr to set return register
10064 in PIC call sequence.
10066 2014-01-25 Walter Lee <walt@tilera.com>
10068 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10069 avoid clobbering a live register.
10071 2014-01-25 Walter Lee <walt@tilera.com>
10073 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10074 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10075 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10076 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10078 2014-01-25 Walter Lee <walt@tilera.com>
10080 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10081 arguments on even registers.
10082 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10084 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10085 (BIGGEST_ALIGNMENT): Ditto.
10086 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10088 2014-01-25 Walter Lee <walt@tilera.com>
10090 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10091 insns before bundling.
10092 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10094 2014-01-25 Walter Lee <walt@tilera.com>
10096 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10097 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10098 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10100 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10102 * config/mips/constraints.md (kl): Delete.
10103 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10104 define expands, using...
10105 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10106 instructions for MIPS16.
10107 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10108 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10110 2014-01-25 Walter Lee <walt@tilera.com>
10112 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10116 2014-01-25 Walter Lee <walt@tilera.com>
10118 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10119 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10121 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10123 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10126 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10128 * print-rtl.c (in_call_function_usage): New var.
10129 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10130 EXPR_LIST mode as mode and not as reg note name.
10132 PR middle-end/59561
10133 * cfgloopmanip.c (copy_loop_info): If
10134 loop->warned_aggressive_loop_optimizations, make sure
10135 the flag is set in target loop too.
10137 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10139 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10141 * builtins.def: Likewise.
10142 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10143 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10144 * ira.c (ira_setup_eliminable_regset): Likewise.
10145 * omp-low.c (gate_expand_omp): Likewise.
10146 (execute_lower_omp): Likewise.
10147 (diagnose_sb_0): Likewise.
10148 (gate_diagnose_omp_blocks): Likewise.
10149 (simd_clone_clauses_extract): Likewise.
10152 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10154 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10155 correction for little endian...
10156 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10159 2014-01-24 Jeff Law <law@redhat.com>
10161 PR tree-optimization/59919
10162 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10163 for non-returning calls.
10165 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10167 * common/config/aarch64/aarch64-common.c
10168 (aarch64_rewrite_mcpu): Handle multiple names.
10169 * config/aarch64/aarch64.h
10170 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10172 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10174 * input.c (add_file_to_cache_tab): Handle the case where fopen
10177 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10180 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10181 from push operand if code of push isn't PRE_DEC.
10183 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10186 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10187 -mquad-memory-atomic. Update -mquad-memory documentation to say
10188 it is only used for non-atomic loads/stores.
10190 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10191 -mquad-memory or -mquad-memory-atomic switches.
10193 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10194 -mquad-memory-atomic to ISA 2.07 support.
10196 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10197 to separate support of normal quad word memory operations (ldq, stq)
10198 from the atomic quad word memory operations.
10200 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10201 support to separate non-atomic quad word operations from atomic
10202 quad word operations. Disable non-atomic quad word operations in
10203 little endian mode so that we don't have to swap words after the
10204 load and before the store.
10205 (quad_load_store_p): Add comment about atomic quad word support.
10206 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10207 options printed with -mdebug=reg.
10209 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10210 -mquad-memory-atomic as the test for whether we have quad word
10211 atomic instructions.
10212 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10213 or -mp8-vector are used, allow byte/half-word atomic operations.
10215 * config/rs6000/sync.md (load_lockedti): Insure that the address
10216 is a proper indexed or indirect address for the lqarx instruction.
10217 On little endian systems, swap the hi/lo registers after the lqarx
10219 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10220 insure the address is valid for the lqarx instruction.
10221 (store_conditionalti): Insure that the address is a proper indexed
10222 or indirect address for the stqcrx. instruction. On little endian
10223 systems, swap the hi/lo registers before doing the stqcrx.
10225 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10226 insure the address is valid for the stqcrx. instruction.
10228 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10229 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10230 type of quad memory support is available.
10232 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10234 PR regression/59915
10235 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10236 there is a danger of looping.
10238 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10240 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10241 force flag_ira_loop_pressure if set via command line.
10243 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10245 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10246 (ashr_simd): New builtin handling DI mode.
10247 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10248 (aarch64_sshr_simddi): New match pattern.
10249 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10250 (vshrd_n_s64): Likewise.
10251 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10253 2014-01-23 Nick Clifton <nickc@redhat.com>
10255 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10256 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10257 favour of mcu specific scripts.
10258 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10261 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10262 Alex Velenko <Alex.Velenko@arm.com>
10264 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10265 (vaddv_s16): Likewise.
10266 (vaddv_s32): Likewise.
10267 (vaddv_u8): Likewise.
10268 (vaddv_u16): Likewise.
10269 (vaddv_u32): Likewise.
10270 (vaddvq_s8): Likewise.
10271 (vaddvq_s16): Likewise.
10272 (vaddvq_s32): Likewise.
10273 (vaddvq_s64): Likewise.
10274 (vaddvq_u8): Likewise.
10275 (vaddvq_u16): Likewise.
10276 (vaddvq_u32): Likewise.
10277 (vaddvq_u64): Likewise.
10278 (vaddv_f32): Likewise.
10279 (vaddvq_f32): Likewise.
10280 (vaddvq_f64): Likewise.
10281 (vmaxv_f32): Likewise.
10282 (vmaxv_s8): Likewise.
10283 (vmaxv_s16): Likewise.
10284 (vmaxv_s32): Likewise.
10285 (vmaxv_u8): Likewise.
10286 (vmaxv_u16): Likewise.
10287 (vmaxv_u32): Likewise.
10288 (vmaxvq_f32): Likewise.
10289 (vmaxvq_f64): Likewise.
10290 (vmaxvq_s8): Likewise.
10291 (vmaxvq_s16): Likewise.
10292 (vmaxvq_s32): Likewise.
10293 (vmaxvq_u8): Likewise.
10294 (vmaxvq_u16): Likewise.
10295 (vmaxvq_u32): Likewise.
10296 (vmaxnmv_f32): Likewise.
10297 (vmaxnmvq_f32): Likewise.
10298 (vmaxnmvq_f64): Likewise.
10299 (vminv_f32): Likewise.
10300 (vminv_s8): Likewise.
10301 (vminv_s16): Likewise.
10302 (vminv_s32): Likewise.
10303 (vminv_u8): Likewise.
10304 (vminv_u16): Likewise.
10305 (vminv_u32): Likewise.
10306 (vminvq_f32): Likewise.
10307 (vminvq_f64): Likewise.
10308 (vminvq_s8): Likewise.
10309 (vminvq_s16): Likewise.
10310 (vminvq_s32): Likewise.
10311 (vminvq_u8): Likewise.
10312 (vminvq_u16): Likewise.
10313 (vminvq_u32): Likewise.
10314 (vminnmv_f32): Likewise.
10315 (vminnmvq_f32): Likewise.
10316 (vminnmvq_f64): Likewise.
10318 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10320 * config/aarch64/aarch64-simd.md
10321 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10322 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10323 (*aarch64_mul3_elt<mode>): Likewise.
10324 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10325 (*aarch64_mul3_elt_to_64v2df): Likewise.
10326 (*aarch64_mla_elt<mode>): Likewise.
10327 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10328 (*aarch64_mls_elt<mode>): Likewise.
10329 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10330 (*aarch64_fma4_elt<mode>): Likewise.
10331 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10332 (*aarch64_fma4_elt_to_64v2df): Likewise.
10333 (*aarch64_fnma4_elt<mode>): Likewise.
10334 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10335 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10336 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10337 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10338 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10339 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10340 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10341 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10342 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10344 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10346 * config/aarch64/aarch64-simd.md
10347 (aarch64_be_checked_get_lane<mode>): New define_expand.
10348 * config/aarch64/aarch64-simd-builtins.def
10349 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10350 New builtin definition.
10351 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10352 Use new safe be builtin.
10354 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10356 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10358 (aarch64_be_st1<mode>): Likewise.
10359 (aarch_ld1<VALL:mode>): Define_expand modified.
10360 (aarch_st1<VALL:mode>): Likewise.
10361 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10362 (UNSPEC_ST1): Likewise.
10364 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10366 * config/microblaze/microblaze.md: Add trap insn and attribute
10368 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10370 PR preprocessor/58580
10371 * input.h (location_get_source_line): Take an additional line_size
10373 (void diagnostics_file_cache_fini): Declare new function.
10374 * input.c (struct fcache): New type.
10375 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10376 New static constants.
10377 (diagnostic_file_cache_init, total_lines_num)
10378 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10379 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10380 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10381 (get_next_line, read_next_line, goto_next_line, read_line_num):
10382 New static function definitions.
10383 (diagnostic_file_cache_fini): New function.
10384 (location_get_source_line): Take an additional output line_len
10385 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10387 * diagnostic.c (diagnostic_finish): Call
10388 diagnostic_file_cache_fini.
10389 (adjust_line): Take an additional input parameter for the length
10390 of the line, rather than calculating it with strlen.
10391 (diagnostic_show_locus): Adjust the use of
10392 location_get_source_line and adjust_line with respect to their new
10393 signature. While displaying a line now, do not stop at the first
10394 null byte. Rather, display the zero byte as a space and keep
10395 going until we reach the size of the line.
10396 * Makefile.in: Add vec.o to OBJS-libcommon
10398 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10399 Ilya Tocar <ilya.tocar@intel.com>
10401 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10402 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10403 (__builtin_ia32_kmov16): Ditto.
10404 * config/i386/i386.md (UNSPEC_KMOV): New.
10407 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10409 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10410 (_mm512_storeu_si512): Ditto.
10412 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10415 * rtl.h (get_referenced_operands): Declare.
10416 * recog.c (get_referenced_operands): New function.
10417 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10418 operands have been referenced when recording LO_SUM references.
10420 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10422 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10424 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10426 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10427 Enable for generic and recent AMD targets.
10429 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10431 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10432 ARG_SIZE note when adjustment was eliminated.
10434 2014-01-22 Jeff Law <law@redhat.com>
10436 PR tree-optimization/59597
10437 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10438 in file. Accept new argument REGISTERING and use it to modify
10439 dump output appropriately.
10440 (register_jump_thread): Corresponding changes.
10441 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10442 thread paths involving joiner blocks. Add code to dump cancelled
10443 jump threading paths.
10445 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10447 PR rtl-optimization/59477
10448 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10449 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10451 2014-01-22 Tom Tromey <tromey@redhat.com>
10453 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10455 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10457 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10459 PR rtl-optimization/59896
10460 * lra-constraints.c (process_alt_operands): Check unused note for
10461 matched operands of insn with no output reloads.
10463 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10465 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10466 (mips_move_from_gpr_cost): Likewise.
10468 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10470 PR rtl-optimization/59858
10471 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10472 ira_class_hard_regs_num.
10473 (process_alt_operands): Increase reject for dying matched operand.
10475 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10478 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10479 smaller than size, perform several stores or loads and stores
10480 at dst + count - size to store or copy all of size bytes, rather
10481 than just last modesize bytes.
10483 2014-01-20 DJ Delorie <dj@redhat.com>
10485 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10486 that CLOBBERs are REGs before propogating their values.
10488 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10490 PR middle-end/59789
10491 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10492 (cgraph_inline_failed_type): New function.
10493 * cgraph.h (DEFCIFCODE): Add type.
10494 (cgraph_inline_failed_type_t): New enum.
10495 (cgraph_inline_failed_type): New prototype.
10496 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10497 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10498 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10499 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10500 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10501 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10502 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10503 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10504 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10505 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10506 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10507 OPTIMIZATION_MISMATCH.
10508 * tree-inline.c (expand_call_inline): Emit errors during
10509 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10511 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10514 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10515 mode attribute in insn output.
10517 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10519 * output.h (output_constant): Delete.
10520 * varasm.c (output_constant): Make private.
10522 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10524 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10526 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10528 PR middle-end/59860
10529 * tree.h (fold_builtin_strcat): New prototype.
10530 * builtins.c (fold_builtin_strcat): No longer static. Add len
10531 argument, if non-NULL, don't call c_strlen. Optimize
10532 directly into __builtin_memcpy instead of __builtin_strcpy.
10533 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10534 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10536 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10538 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10539 for SImode_address_operand operands, having only a REG argument.
10541 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10543 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10544 loader name using mbig-endian.
10545 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10547 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10549 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10550 (-mtune): Likewise.
10553 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10555 * config/aarch64/aarch64-protos.h
10556 (aarch64_cannot_change_mode_class_ptr): Declare.
10557 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10558 aarch64_cannot_change_mode_class_ptr): New.
10559 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10560 backend hook aarch64_cannot_change_mode_class.
10562 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10564 * common/config/aarch64/aarch64-common.c
10565 (aarch64_handle_option): Don't handle any option order logic here.
10566 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10567 selected_cpu, warn on architecture version mismatch.
10568 (aarch64_override_options): Fix parsing order for option strings.
10570 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10571 Iain Sandoe <iain@codesourcery.com>
10574 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10575 warning. Amend comment to reflect current functionality.
10577 2014-01-20 Richard Biener <rguenther@suse.de>
10579 PR middle-end/59860
10580 * builtins.c (fold_builtin_strcat): Remove case better handled
10581 by tree-ssa-strlen.c.
10583 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10585 * config/aarch64/aarch64.opt
10586 (mcpu, march, mtune): Make case-insensitive.
10588 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10591 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10592 if operands[1] is a REG or ZERO_EXTEND of a REG.
10594 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10596 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10598 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10600 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10601 long non-pic millicode calls.
10603 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10605 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10607 2014-01-19 Kito Cheng <kito@0xlab.org>
10609 * builtins.c (expand_movstr): Check movstr expand done or fail.
10611 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10612 H.J. Lu <hongjiu.lu@intel.com>
10615 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10616 to DImode for zero-extended addresses.
10618 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10620 PR rtl-optimization/57763
10621 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10622 on the new indirect jump_insn and increment LABEL_NUSES (label).
10624 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10628 * config.gcc (x86_archs): New variable.
10629 (x86_64_archs): Likewise.
10630 (x86_cpus): Likewise.
10631 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10632 --with-arch/--with-cpu= options.
10633 Support --with-arch=/--with-cpu={nehalem,westmere,
10634 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10636 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10638 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10639 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10641 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10643 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10645 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10648 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10649 clear cpp_get_options (parse_in)->warn_unused_macros for
10650 ix86_target_macros_internal with cpp_define.
10652 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10654 * jump.c (delete_related_insns): Keep (use (insn))s.
10655 * reorg.c (redundant_insn): Check for barriers too.
10657 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10659 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10661 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10663 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10664 call to $$dyncall when TARGET_LONG_CALLS is true.
10666 2014-01-17 Jeff Law <law@redhat.com>
10668 * ree.c (combine_set_extension): Temporarily disable test for
10669 changing number of hard registers.
10671 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10673 PR middle-end/58125
10674 * ipa-inline-analysis.c (inline_free_summary):
10675 Do not free summary of aliases.
10677 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10679 PR middle-end/59706
10680 * gimplify.c (gimplify_expr): Use create_tmp_var
10681 instead of create_tmp_var_raw. If cond doesn't have
10682 integral type, don't add the IFN_ANNOTATE builtin at all.
10684 2014-01-17 Martin Jambor <mjambor@suse.cz>
10687 * ipa-cp.c (prev_edge_clone): New variable.
10688 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10689 Also resize prev_edge_clone vector.
10690 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10691 (ipcp_edge_removal_hook): New function.
10692 (ipcp_driver): Register ipcp_edge_removal_hook.
10694 2014-01-17 Andrew Pinski <apinski@cavium.com>
10695 Steve Ellcey <sellcey@mips.com>
10698 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10701 2014-01-17 Jeff Law <law@redhat.com>
10703 PR middle-end/57904
10704 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10705 so that pass_ccp runs first.
10707 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10709 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10710 (ix86_adjust_cost): Use !TARGET_XXX.
10711 (do_reorder_for_imul): Likewise.
10712 (swap_top_of_ready_list): Likewise.
10713 (ix86_sched_reorder): Likewise.
10715 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10717 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10718 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10719 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10720 (intel_memset): New. Duplicate slm_memset.
10721 (intel_cost): New. Duplicate slm_cost.
10722 (m_INTEL): New macro.
10723 (processor_target_table): Add "intel".
10724 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10725 with PROCESSOR_INTEL for "intel".
10726 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10727 PROCESSOR_SILVERMONT.
10728 (ix86_issue_rate): Likewise.
10729 (ix86_adjust_cost): Likewise.
10730 (ia32_multipass_dfa_lookahead): Likewise.
10731 (swap_top_of_ready_list): Likewise.
10732 (ix86_sched_reorder): Likewise.
10733 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10734 instead of TARGET_OPT_AGU.
10735 * config/i386/i386.h (TARGET_INTEL): New.
10736 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10737 (processor_type): Add PROCESSOR_INTEL.
10738 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10739 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10741 2014-01-17 Marek Polacek <polacek@redhat.com>
10744 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10747 2014-01-17 Richard Biener <rguenther@suse.de>
10749 PR tree-optimization/46590
10750 * opts.c (default_options_table): Add entries for
10751 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10752 all enabled at -O1 but not for -Og.
10753 * common.opt (fbranch-count-reg): Remove Init(1).
10754 (fmove-loop-invariants): Likewise.
10755 (ftree-pta): Likewise.
10757 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10759 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10760 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10761 decls to at least the GCC 4.8 used alignments.
10764 * tree-nested.c (convert_nonlocal_reference_stmt,
10765 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10766 of GIMPLE_BIND stmts, adjust associated decls.
10768 2014-01-17 Richard Biener <rguenther@suse.de>
10770 PR tree-optimization/46590
10771 * vec.h (vec<>::bseach): New member function implementing
10772 binary search according to C89 bsearch.
10773 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10774 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10775 bitmap pointer again. Make accesses_in_loop a flat array.
10776 (mem_ref_obstack): New global.
10777 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10778 (mark_ref_stored): Likewise.
10779 (ref_indep_loop_p_2): Likewise.
10780 (set_ref_stored_in_loop): New helper function.
10781 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10782 (memref_free): Adjust.
10783 (record_mem_ref_loc): Simplify.
10784 (gather_mem_refs_stmt): Adjust.
10785 (sort_locs_in_loop_postorder_cmp): New function.
10786 (analyze_memory_references): Sort accesses_in_loop after
10787 loop postorder number.
10788 (find_ref_loc_in_loop_cmp): New function.
10789 (for_all_locs_in_loop): Find relevant cluster of locs in
10790 accesses_in_loop and iterate without recursion.
10791 (execute_sm): Avoid uninit warning.
10792 (struct ref_always_accessed): Simplify.
10793 (ref_always_accessed::operator ()): Likewise.
10794 (ref_always_accessed_p): Likewise.
10795 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10796 loop postorder numbers here.
10797 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10800 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10803 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10804 on decls for which assemble_alias has been called.
10806 2014-01-17 Nick Clifton <nickc@redhat.com>
10808 * config/msp430/msp430.opt: (mcpu): New option.
10809 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10810 (msp430_option_override): Parse target_cpu. If the MCU name
10811 matches a generic string, clear target_mcu.
10812 (msp430_attr): Allow numeric interrupt values up to 63.
10813 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10814 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10816 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10818 * config/msp430/msp430.md (popm): Use %J rather than %I.
10819 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10820 (addhi_cy_i): Use immediate_operand for operand 2.
10821 * doc/invoke.texi: Document -mcpu option.
10823 2014-01-17 Richard Biener <rguenther@suse.de>
10825 PR rtl-optimization/38518
10826 * df.h (df_analyze_loop): Declare.
10827 * df-core.c: Include cfgloop.h.
10828 (df_analyze_1): Split out main part of df_analyze.
10829 (df_analyze): Adjust.
10830 (loop_inverted_post_order_compute): New function.
10831 (loop_post_order_compute): Likewise.
10832 (df_analyze_loop): New function avoiding whole-function
10833 postorder computes.
10834 * loop-invariant.c (find_defs): Use df_analyze_loop.
10835 (find_invariants): Adjust.
10836 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10838 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10840 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10841 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10843 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10845 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10846 traversal when removing references.
10848 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10851 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10853 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10855 PR middle-end/56791
10856 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10857 pushing a reload for an autoinc when we had previously reloaded an
10858 inner part of the address.
10860 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10862 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10864 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10865 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10866 when not giving up or versioning for alias only because of
10868 (vect_analyze_data_ref_dependences): Set to true.
10869 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10871 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10872 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10875 PR middle-end/58344
10876 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10879 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10880 operand 0 predicate for gathers, use a new pseudo as subtarget.
10882 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10884 PR middle-end/59609
10885 * lra-constraints.c (process_alt_operands): Add printing debug info.
10886 Check absence of input/output reloads for matched operands too.
10888 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10890 PR rtl-optimization/59835
10891 * ira.c (ira_init_register_move_cost): Increase cost for
10894 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10896 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10898 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10901 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10902 non-register objects. Use gen_(high/low)part more consistently.
10905 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10908 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10909 endian support, remove tests for WORDS_BIG_ENDIAN.
10910 (p8_mfvsrd_3_<mode>): Likewise.
10911 (reload_gpr_from_vsx<mode>): Likewise.
10912 (reload_gpr_from_vsxsf): Likewise.
10913 (p8_mfvsrd_4_disf): Likewise.
10915 2014-01-16 Richard Biener <rguenther@suse.de>
10917 PR rtl-optimization/46590
10918 * lcm.c (compute_antinout_edge): Use postorder iteration.
10919 (compute_laterin): Use inverted postorder iteration.
10921 2014-01-16 Nick Clifton <nickc@redhat.com>
10923 PR middle-end/28865
10924 * varasm.c (output_constant): Return the number of bytes actually
10926 (output_constructor_array_range): Update the field size with the
10927 number of bytes emitted by output_constant.
10928 (output_constructor_regular_field): Likewise. Also do not
10929 complain if the total number of bytes emitted is now greater
10930 than the expected fieldpos.
10931 * output.h (output_constant): Update prototype and descriptive comment.
10933 2014-01-16 Marek Polacek <polacek@redhat.com>
10935 PR middle-end/59827
10936 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10937 it is error_mark_node.
10939 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10941 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10942 VALID_AVX256_REG_OR_OI_MODE.
10944 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10946 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10947 current procedure should be profiled.
10949 2014-01-15 Andrew Pinski <apinski@cavium.com>
10951 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10952 of moving from/to the STACK_REG register class.
10954 2014-01-15 Richard Henderson <rth@redhat.com>
10957 * reginfo.c (global_regs_decl): Globalize.
10958 * rtl.h (global_regs_decl): Declare.
10959 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10960 reserved via global_regs.
10962 2014-01-15 Teresa Johnson <tejohnson@google.com>
10964 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10966 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10968 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10969 and vmulosh rather than call gen_vec_widen_smult_*.
10970 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10971 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10972 (vec_widen_smult_even_v16qi): Likewise.
10973 (vec_widen_umult_even_v8hi): Likewise.
10974 (vec_widen_smult_even_v8hi): Likewise.
10975 (vec_widen_umult_odd_v16qi): Likewise.
10976 (vec_widen_smult_odd_v16qi): Likewise.
10977 (vec_widen_umult_odd_v8hi): Likewise.
10978 (vec_widen_smult_odd_v8hi): Likewise.
10979 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10980 vmuloub rather than call gen_vec_widen_umult_*.
10981 (vec_widen_umult_lo_v16qi): Likewise.
10982 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10983 vmulosb rather than call gen_vec_widen_smult_*.
10984 (vec_widen_smult_lo_v16qi): Likewise.
10985 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10986 rather than call gen_vec_widen_umult_*.
10987 (vec_widen_umult_lo_v8hi): Likewise.
10988 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10989 rather than call gen_vec_widen_smult_*.
10990 (vec_widen_smult_lo_v8hi): Likewise.
10992 2014-01-15 Jeff Law <law@redhat.com>
10994 PR tree-optimization/59747
10995 * ree.c (find_and_remove_re): Properly handle case where a second
10996 eliminated extension requires widening a copy created for elimination
10997 of a prior extension.
10998 (combine_set_extension): Ensure that the number of hard regs needed
10999 for a destination register does not change when we widen it.
11001 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11003 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11004 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11005 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11006 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11007 (avr-*-rtems*): Likewise.
11008 (bfin*-rtems*): Likewise.
11009 (moxie-*-rtems*): Likewise.
11010 (h8300-*-rtems*): Likewise.
11011 (i[34567]86-*-rtems*): Likewise.
11012 (lm32-*-rtems*): Likewise.
11013 (m32r-*-rtems*): Likewise.
11014 (m68k-*-rtems*): Likewise.
11015 (microblaze*-*-rtems*): Likewise.
11016 (mips*-*-rtems*): Likewise.
11017 (powerpc-*-rtems*): Likewise.
11018 (sh-*-rtems*): Likewise.
11019 (sparc-*-rtems*): Likewise.
11020 (sparc64-*-rtems*): Likewise.
11021 (v850-*-rtems*): Likewise.
11022 (m32c-*-rtems*): Likewise.
11024 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11026 PR rtl-optimization/59511
11027 * ira.c (ira_init_register_move_cost): Use memory costs for some
11028 cases of register move cost calculations.
11029 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11030 instead of BB frequency.
11031 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11032 * lra-assigns.c (find_hard_regno_for): Ditto.
11034 2014-01-15 Richard Biener <rguenther@suse.de>
11036 PR tree-optimization/59822
11037 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11038 (vectorizable_load): Use it to hoist defs of uses of invariant
11039 loads out of the loop.
11041 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11042 Kugan Vivekanandarajah <kuganv@linaro.org>
11045 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11048 2014-01-15 Richard Biener <rguenther@suse.de>
11050 PR rtl-optimization/59802
11051 * lcm.c (compute_available): Use inverted postorder to seed
11052 the initial worklist.
11054 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11057 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11058 ADDR_REGS for invalid symrefs in non-PIC code.
11060 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11063 * builtins.c (determine_block_size): Initialize *probable_max_size
11064 even if len_rtx is CONST_INT.
11066 2014-01-14 Andrew Pinski <apinski@cavium.com>
11068 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11069 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11070 (cortexa53_tunings): Likewise.
11071 (aarch64_sched_issue_rate): New function.
11072 (TARGET_SCHED_ISSUE_RATE): Define.
11074 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11076 * ira-costs.c (find_costs_and_classes): Add missed
11077 ira_init_register_move_cost_if_necessary.
11079 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11082 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11084 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11087 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11088 to indicate if type is used for function return value. Warn ABI
11089 change if the vector mode isn't available for function return value.
11090 (ix86_function_arg_advance): Pass false to type_natural_mode.
11091 (ix86_function_arg): Likewise.
11092 (ix86_gimplify_va_arg): Likewise.
11093 (function_arg_32): Don't warn ABI change.
11094 (ix86_function_value): Pass true to type_natural_mode.
11095 (ix86_return_in_memory): Likewise.
11096 (ix86_struct_value_rtx): Removed.
11097 (TARGET_STRUCT_VALUE_RTX): Likewise.
11099 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11101 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11102 converting a conditional jump into a conditional return.
11104 2014-01-14 Richard Biener <rguenther@suse.de>
11106 PR tree-optimization/58921
11107 PR tree-optimization/59006
11108 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11109 hoisting invariant stmts.
11110 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11111 invariant loads on the preheader edge if possible.
11113 2014-01-14 Joey Ye <joey.ye@arm.com>
11115 * doc/plugin.texi (Building GCC plugins): Update to C++.
11117 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11119 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11120 (_mm_rcp28_round_ss): Ditto.
11121 (_mm_rsqrt28_round_sd): Ditto.
11122 (_mm_rsqrt28_round_ss): Ditto.
11123 (_mm_rcp28_sd): Ditto.
11124 (_mm_rcp28_ss): Ditto.
11125 (_mm_rsqrt28_sd): Ditto.
11126 (_mm_rsqrt28_ss): Ditto.
11127 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11128 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11129 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11130 (IX86_BUILTIN_RCP28SD): Ditto.
11131 (IX86_BUILTIN_RCP28SS): Ditto.
11132 (IX86_BUILTIN_RSQRT28SD): Ditto.
11133 (IX86_BUILTIN_RSQRT28SS): Ditto.
11134 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11135 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11136 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11137 (ix86_expand_special_args_builtin): Expand new FTYPE.
11138 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11139 (srcp14<mode>): Make insn unary.
11140 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11141 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11142 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11143 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11144 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11145 Fix rounding: make it SAE only.
11146 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11148 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11150 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11151 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11152 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11153 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11154 (round_saeonly_mask_scalar_operand4): Ditto.
11155 (round_saeonly_mask_scalar_op3): Ditto.
11156 (round_saeonly_mask_scalar_op4): Ditto.
11158 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11160 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11161 Implement -maltivec=be for vec_insert and vec_extract.
11163 2014-01-10 DJ Delorie <dj@redhat.com>
11165 * config/msp430/msp430.md (call_internal): Don't allow memory
11166 references with SP as the base register.
11167 (call_value_internal): Likewise.
11168 * config/msp430/constraints.md (Yc): New. For memory references
11169 that don't use SP as a base register.
11171 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11172 "an integer without a # prefix"
11173 * config/msp430/msp430.md (epilogue_helper): Use it.
11175 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11178 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11179 AVX512F gather builtins.
11180 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11181 on gather decls with INTEGER_TYPE masktype.
11182 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11183 directly into the builtin rather than hoisting it before loop.
11185 PR tree-optimization/59387
11186 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11187 (scev_const_prop): If folded_casts and type has undefined overflow,
11188 use force_gimple_operand instead of force_gimple_operand_gsi and
11189 for each added stmt if it is assign with
11190 arith_code_with_undefined_signed_overflow, call
11191 rewrite_to_defined_overflow.
11192 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11193 gimple-fold.h instead.
11194 (arith_code_with_undefined_signed_overflow,
11195 rewrite_to_defined_overflow): Moved to ...
11196 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11197 rewrite_to_defined_overflow): ... here. No longer static.
11198 Include gimplify-me.h.
11199 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11200 rewrite_to_defined_overflow): New prototypes.
11202 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11204 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11206 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11208 * builtins.c (get_object_alignment_2): Minor tweak.
11209 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11211 2014-01-13 Christian Bruel <christian.bruel@st.com>
11213 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11214 optimized non constant lengths.
11216 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11219 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11220 load as __atomic_load_N if possible.
11222 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11224 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11226 (rs6000_expand_builtin): Adjust call.
11228 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11231 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11232 * config/rs6000/rs6000.c: Include target-globals.h.
11233 (rs6000_set_current_function): Instead of doing target_reinit
11234 unconditionally, use save_target_globals_default_opts and
11235 restore_target_globals.
11237 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11239 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11240 (rs6000_expand_builtin): Handle mffs and mtfsf.
11241 (rs6000_init_builtins): Define mffs and mtfsf.
11242 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11243 (rs6000_mffs): New pattern.
11244 (rs6000_mtfsf): New pattern.
11246 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11248 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11249 Start narrowing with START. Apply candidate-use pair
11250 and check overall cost in narrowing.
11251 (iv_ca_prune): Pass new argument.
11253 2014-01-10 Jeff Law <law@redhat.com>
11255 PR middle-end/59743
11256 * ree.c (combine_reaching_defs): Ensure the defining statement
11257 occurs before the extension when optimizing extensions with
11258 different source and destination hard registers.
11260 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11263 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11264 vtables into the type inheritance graph.
11266 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11268 PR rtl-optimization/59754
11269 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11270 modes in the REGNO != REGNO case.
11272 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11274 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11276 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11278 PR tree-optimization/59745
11279 * tree-predcom.c (tree_predictive_commoning_loop): Call
11280 free_affine_expand_cache if giving up because components is NULL.
11282 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11283 GC in payload of target_globals struct instead of allocating them on
11284 the heap and the larger structs separately using GC.
11285 * target-globals.h (struct target_globals): Make regs, hard_regs,
11286 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11287 of GTY((skip)) and change type to void *.
11288 (reset_target_globals): Cast loads from those fields to corresponding
11291 2014-01-10 Steve Ellcey <sellcey@mips.com>
11294 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11295 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11296 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11298 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11301 * aarch64-modes.def (CC_Zmode): New flags mode.
11302 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11303 represents an equality.
11304 (aarch64_get_condition_code): Handle CC_Zmode.
11305 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11307 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11309 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11310 extraction in good case.
11312 2014-01-10 Richard Biener <rguenther@suse.de>
11314 PR tree-optimization/59374
11315 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11316 checking after SLP discovery. Mark stmts not participating
11317 in any SLP instance properly.
11319 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11321 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11322 when handling a SET rtx.
11324 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11326 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11327 (cortex-a57): Likewise.
11328 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11330 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11332 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11333 non-iwmmxt builtins.
11335 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11339 * ipa-devirt.c record_target_from_binfo): Take as argument
11340 stack of binfos and lookup matching one for virtual inheritance.
11341 (possible_polymorphic_call_targets_1): Update.
11343 2014-01-10 Huacai Chen <chenhc@lemote.com>
11345 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11346 kernel strings for Loongson-2E/2F/3A.
11348 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11350 PR middle-end/59670
11351 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11352 is_gimple_call before calling gimple_call_internal_p.
11354 2014-01-09 Steve Ellcey <sellcey@mips.com>
11356 * Makefile.in (TREE_FLOW_H): Remove.
11357 (TREE_SSA_H): Add file names from tree-flow.h.
11358 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11359 * tree.h: Remove tree-flow.h reference.
11360 * hash-table.h: Remove tree-flow.h reference.
11361 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11362 reference with tree-ssa-loop.h.
11364 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11366 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11367 default element-order behavior for -maltivec.
11368 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11369 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11370 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11371 when targeting big endian, at least for now.
11372 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11374 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11376 PR middle-end/47735
11377 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11378 var satisfies use_register_for_decl, just take into account type
11379 alignment, rather than decl alignment.
11381 PR tree-optimization/59622
11382 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11383 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11384 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11385 Don't devirtualize for inplace at all. For targets.length () == 1,
11386 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11388 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11390 * config/i386/i386.md (cpu): Remove the unused btver1.
11392 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11394 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11396 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11399 * tree-core.h (struct target_globals): New forward declaration.
11400 (struct tree_target_option): Add globals field.
11401 * tree.h (TREE_TARGET_GLOBALS): Define.
11402 (prepare_target_option_nodes_for_pch): New prototype.
11403 * target-globals.h (struct target_globals): Define even if
11404 !SWITCHABLE_TARGET.
11405 * tree.c (prepare_target_option_node_for_pch,
11406 prepare_target_option_nodes_for_pch): New functions.
11407 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11408 * config/i386/i386.c: Include target-globals.h.
11409 (ix86_set_current_function): Instead of doing target_reinit
11410 unconditionally, use save_target_globals_default_opts and
11411 restore_target_globals.
11413 2014-01-09 Richard Biener <rguenther@suse.de>
11415 PR tree-optimization/59715
11416 * tree-cfg.h (split_critical_edges): Declare.
11417 * tree-cfg.c (split_critical_edges): Export.
11418 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11420 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11422 * cfgexpand.c (expand_stack_vars): Optionally disable
11423 asan stack protection.
11424 (expand_used_vars): Likewise.
11425 (partition_stack_vars): Likewise.
11426 * asan.c (asan_emit_stack_protection): Optionally disable
11427 after return stack usage.
11428 (instrument_derefs): Optionally disable memory access instrumentation.
11429 (instrument_builtin_call): Likewise.
11430 (instrument_strlen_call): Likewise.
11431 (asan_protect_global): Optionally disable global variables protection.
11432 * doc/invoke.texi: Added doc for new options.
11433 * params.def: Added new options.
11434 * params.h: Likewise.
11436 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11438 PR rtl-optimization/59724
11439 * ifcvt.c (cond_exec_process_if_block): Don't call
11440 flow_find_head_matching_sequence with 0 longest_match.
11441 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11442 non-active insns if !stop_after.
11443 (try_head_merge_bb): Revert 2014-01-07 changes.
11445 2014-01-08 Jeff Law <law@redhat.com>
11447 * ree.c (get_sub_rtx): New function, extracted from...
11448 (merge_def_and_ext): Here.
11449 (combine_reaching_defs): Use get_sub_rtx.
11451 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11453 * cgraph.h (varpool_variable_node): Do not choke on null node.
11455 2014-01-08 Catherine Moore <clm@codesourcery.com>
11457 * config/mips/mips.md (simple_return): Attempt to use JRC
11459 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11461 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11463 PR rtl-optimization/59137
11464 * reorg.c (steal_delay_list_from_target): Call update_block for
11466 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11468 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11470 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11471 two duplicate entries.
11473 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11476 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11478 * config/mips/mips.c (mips_truncated_op_cost): New function.
11479 (mips_rtx_costs): Adjust test for BADDU.
11480 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11482 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11484 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11485 (*baddu_si): ...this new pattern.
11487 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11490 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11492 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11494 PR middle-end/57748
11495 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11497 (expand_expr, expand_normal): Adjust.
11498 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11499 inner_reference_p. Use inner_reference_p to expand inner references.
11500 (store_expr): Adjust.
11501 * cfgexpand.c (expand_call_stmt): Adjust.
11503 2014-01-08 Rong Xu <xur@google.com>
11505 * gcov-io.c (gcov_var): Move from gcov-io.h.
11506 (gcov_position): Ditto.
11507 (gcov_is_error): Ditto.
11508 (gcov_rewrite): Ditto.
11509 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11510 only part to libgcc/libgcov.h.
11512 2014-01-08 Marek Polacek <polacek@redhat.com>
11514 PR middle-end/59669
11515 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11517 2014-01-08 Marek Polacek <polacek@redhat.com>
11520 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11522 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11524 PR rtl-optimization/59649
11525 * stor-layout.c (get_mode_bounds): For BImode return
11526 0 and STORE_FLAG_VALUE.
11528 2014-01-08 Richard Biener <rguenther@suse.de>
11530 PR middle-end/59630
11531 * gimple.h (is_gimple_builtin_call): Remove.
11532 (gimple_builtin_call_types_compatible_p): New.
11533 (gimple_call_builtin_p): New overload.
11534 * gimple.c (is_gimple_builtin_call): Remove.
11535 (validate_call): Rename to ...
11536 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11537 check return types.
11538 (validate_type): New static function.
11539 (gimple_call_builtin_p): New overload and adjust.
11540 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11541 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11542 (gimple_fold_stmt_to_constant_1): Likewise.
11543 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11545 2014-01-08 Richard Biener <rguenther@suse.de>
11547 PR middle-end/59471
11548 * gimplify.c (gimplify_expr): Gimplify register-register type
11549 VIEW_CONVERT_EXPRs to separate stmts.
11551 2014-01-07 Jeff Law <law@redhat.com>
11553 PR middle-end/53623
11554 * ree.c (combine_set_extension): Handle case where source
11555 and destination registers in an extension insn are different.
11556 (combine_reaching_defs): Allow source and destination registers
11557 in extension to be different under limited circumstances.
11558 (add_removable_extension): Remove restriction that the
11559 source and destination registers in the extension are the same.
11560 (find_and_remove_re): Emit a copy from the extension's
11561 destination to its source after the defining insn if
11562 the source and destination registers are different.
11564 PR middle-end/59285
11565 * ifcvt.c (merge_if_block): If we are merging a block with more than
11566 one successor with a block with no successors, remove any BARRIER
11567 after the second block.
11569 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11571 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11573 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11576 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11577 for 14-bit register offsets when INT14_OK_STRICT is false.
11579 2014-01-07 Roland Stigge <stigge@antcom.de>
11580 Michael Meissner <meissner@linux.vnet.ibm.com>
11583 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11584 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11586 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11588 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11591 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11593 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11594 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11595 rtx is const0_rtx or not.
11597 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11600 * target-globals.c (save_target_globals): Remove this_fn_optab
11602 * toplev.c: Include optabs.h.
11603 (target_reinit): Temporarily restore the global options if another
11604 set of options are in force.
11606 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11608 PR rtl-optimization/58668
11609 * cfgcleanup.c (flow_find_cross_jump): Don't count
11610 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11611 to determine what is counted.
11612 (flow_find_head_matching_sequence): Use active_insn_p to determine
11614 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11616 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11617 determine what is counted.
11619 PR tree-optimization/59643
11620 * tree-predcom.c (split_data_refs_to_components): If one dr is
11621 read and one write, determine_offset fails and the write isn't
11622 in the bad component, just put the read into the bad component.
11624 2014-01-07 Mike Stump <mikestump@comcast.net>
11625 Jakub Jelinek <jakub@redhat.com>
11628 * tree-core.h (struct tree_optimization_option): Change optabs
11629 type from unsigned char * to void *.
11630 * optabs.c (init_tree_optimization_optabs): Adjust
11631 TREE_OPTIMIZATION_OPTABS initialization.
11633 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11636 * config/i386/i386.h (struct machine_function): Add
11637 no_drap_save_restore field.
11638 * config/i386/i386.c (ix86_save_reg): Use
11639 !cfun->machine->no_drap_save_restore instead of
11640 crtl->stack_realign_needed.
11641 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11642 this function clears frame_pointer_needed. Set
11643 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11644 and DRAP reg is needed.
11646 2014-01-06 Marek Polacek <polacek@redhat.com>
11649 * doc/implement-c.texi: Mention that other integer types are
11650 permitted as bit-field types in strictly conforming mode.
11652 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11654 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11655 is newly allocated.
11657 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11659 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11661 2014-01-06 Martin Jambor <mjambor@suse.cz>
11664 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11666 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11668 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11672 * var-tracking.c (add_stores): Preserve the value of the source even if
11673 we don't record the store.
11675 2014-01-06 Terry Guo <terry.guo@arm.com>
11677 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11679 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11682 * config/darwin.c (darwin_function_section): Adjust return values to
11683 correspond to optimisation changes made in r206070.
11685 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11687 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11688 from prefetch_block tune setting.
11689 (nocona_cost): Correct size of prefetch block to 64.
11691 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11693 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11694 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11695 used to save the static chain register in the computation of the offset
11696 from which the FP registers need to be restored.
11698 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11700 PR tree-optimization/59519
11701 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11702 ICE if get_current_def (current_new_name) is already non-NULL, as long
11703 as it is a phi result of some other phi in *new_exit_bb that has
11706 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11707 or vmovdqu* for misaligned_operand.
11708 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11709 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11710 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11711 aligned_mem for AVX512F masked aligned load and store builtins and for
11712 non-temporal moves.
11714 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11716 PR tree-optimization/59651
11717 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11718 Address range for negative step should be added by TYPE_SIZE_UNIT.
11720 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11722 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11723 overlapping registers also for registers other than the stack pointer.
11725 2014-01-03 Marek Polacek <polacek@redhat.com>
11728 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11731 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11734 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11737 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11738 (push splitter): Use <P:MODE_SIZE> instead of
11739 GET_MODE_SIZE (<P:MODE>mode).
11740 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11741 (mov -1, reg peephole2): Likewise.
11742 * config/i386/sse.md (*mov<mode>_internal,
11743 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11744 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11745 *<code><mode>3, *andnot<mode>3<mask_name>,
11746 <mask_codefor><code><mode>3<mask_name>): Likewise.
11747 * config/i386/subst.md (mask_mode512bit_condition,
11748 sd_mask_mode512bit_condition): Likewise.
11750 2014-01-02 Xinliang David Li <davidxl@google.com>
11752 PR tree-optimization/59303
11753 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11754 (dump_predicates): Better output format.
11755 (pred_equal_p): New function.
11756 (is_neq_relop_p): Ditto.
11757 (is_neq_zero_form_p): Ditto.
11758 (pred_expr_equal_p): Ditto.
11759 (pred_neg_p): Ditto.
11760 (simplify_pred): Ditto.
11761 (simplify_preds_2): Ditto.
11762 (simplify_preds_3): Ditto.
11763 (simplify_preds_4): Ditto.
11764 (simplify_preds): Ditto.
11765 (push_pred): Ditto.
11766 (push_to_worklist): Ditto.
11767 (get_pred_info_from_cmp): Ditto.
11768 (is_degenerated_phi): Ditto.
11769 (normalize_one_pred_1): Ditto.
11770 (normalize_one_pred): Ditto.
11771 (normalize_one_pred_chain): Ditto.
11772 (normalize_preds): Ditto.
11773 (normalize_cond_1): Remove function.
11774 (normalize_cond): Ditto.
11775 (is_gcond_subset_of): Ditto.
11776 (is_subset_of_any): Ditto.
11777 (is_or_set_subset_of): Ditto.
11778 (is_and_set_subset_of): Ditto.
11779 (is_norm_cond_subset_of): Ditto.
11780 (pred_chain_length_cmp): Ditto.
11781 (convert_control_dep_chain_into_preds): Type change.
11782 (find_predicates): Ditto.
11783 (find_def_preds): Ditto.
11784 (destroy_predicates_vecs): Ditto.
11785 (find_matching_predicates_in_rest_chains): Ditto.
11786 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11787 (is_pred_expr_subset): Ditto.
11788 (is_pred_chain_subset_of): Ditto.
11789 (is_included_in): Ditto.
11790 (is_superset_of): Ditto.
11792 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11794 Update copyright years.
11796 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11798 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11799 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11800 config/arc/arc.md, config/arc/arc.opt,
11801 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11802 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11803 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11804 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11805 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11806 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11807 vtable-verify.c, vtable-verify.h: Use the standard form for the
11810 2014-01-02 Tobias Burnus <burnus@net-b.de>
11812 * gcc.c (process_command): Update copyright notice dates.
11813 * gcov-dump.c: Ditto.
11815 * doc/cpp.texi: Bump @copying's copyright year.
11816 * doc/cppinternals.texi: Ditto.
11817 * doc/gcc.texi: Ditto.
11818 * doc/gccint.texi: Ditto.
11819 * doc/gcov.texi: Ditto.
11820 * doc/install.texi: Ditto.
11821 * doc/invoke.texi: Ditto.
11823 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11825 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11827 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11829 * config/i386/sse.md (*mov<mode>_internal): Guard
11830 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11832 PR rtl-optimization/59647
11833 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11834 new_rtx into UNSIGNED_FLOAT rtxes.
11836 Copyright (C) 2014 Free Software Foundation, Inc.
11838 Copying and distribution of this file, with or without modification,
11839 are permitted in any medium without royalty provided the copyright
11840 notice and this notice are preserved.