1 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
3 PR rtl-optimization/61215
4 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
5 simplify_gen_subreg until final substitution.
7 2014-05-23 Alan Modra <amodra@gmail.com>
10 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
11 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
12 Use "Y" constraint rather than "m".
14 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
16 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
18 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
19 New function declaration.
20 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
21 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
22 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
23 (aarch64_init_builtins) : Initialize builtins
24 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
25 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
26 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
27 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
28 and __builtins_aarch64_set_fpsr.
29 (aarch64_atomic_assign_expand_fenv): New function.
30 * config/aarch64/aarch64.md (set_fpcr): New pattern.
31 (get_fpcr) : Likewise.
32 (set_fpsr) : Likewise.
33 (get_fpsr) : Likewise.
34 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
36 * doc/extend.texi (AARCH64 Built-in Functions) : Document
37 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
38 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
40 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
42 PR rtl-optimization/60969
43 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
44 constraints. Set up mem cost for NO_REGS case.
46 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
48 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
50 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
52 * config/darwin.c: Include "lto-section-names.h".
53 (LTO_SEGMENT_NAME): Don't define.
54 * config/i386/winnt.c: Include "lto-section-names.h".
55 * lto-streamer.c: Include "lto-section-names.h".
56 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
57 * lto-wrapper.c: Include "lto-section-names.h".
58 (LTO_SECTION_NAME_PREFIX): Don't define.
59 * lto-section-names.h: New file.
60 * cgraphunit.c: Include "lto-section-names.h".
62 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
64 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
66 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
69 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
71 2014-05-22 Nick Clifton <nickc@redhat.com>
73 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
76 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
78 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
79 -> (T)A transformation to integer types.
81 2014-05-22 Teresa Johnson <tejohnson@google.com>
83 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
84 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
85 (gcov_rewrite): Use gcov_nonruntime_assert.
87 (gcov_write_words): Ditto.
88 (gcov_write_length): Ditto.
89 (gcov_read_words): Use gcov_nonruntime_assert, and remove
90 gcc_assert from IN_LIBGCOV code.
91 (gcov_read_summary): Use gcov_error to flag profile corruption.
92 (gcov_sync): Use gcov_nonruntime_assert.
93 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
94 (gcov_histo_index): Use gcov_nonruntime_assert.
95 (static void gcov_histogram_merge): Ditto.
96 (compute_working_sets): Ditto.
97 * gcov-io.h (gcov_nonruntime_assert): Define.
98 (gcov_error): Define for !IN_LIBGCOV
100 2014-05-22 Richard Biener <rguenther@suse.de>
102 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
103 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
104 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
105 and deallocation site.
106 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
107 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
108 passing through the incoming points-to set.
109 (handle_lhs_call): Use flags argument instead of recomputing it.
110 (find_func_aliases_for_call): Call handle_lhs_call with proper
113 2014-05-22 Jakub Jelinek <jakub@redhat.com>
115 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
116 all padding bits in REAL_VALUE_TYPE are cleared.
118 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
120 Cleanup and improve multipass_dfa_lookahead_guard
121 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
122 (core2i7_first_cycle_multipass_begin,)
123 (core2i7_first_cycle_multipass_issue,)
124 (core2i7_first_cycle_multipass_backtrack): Update signature.
126 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
127 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
128 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
130 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
131 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
133 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
135 * doc/tm.texi: Regenerate.
137 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
138 * haifa-sched.c (ready_try): Make signed to allow negative values.
139 (rebug_ready_list_1): Update.
140 (choose_ready): Simplify.
141 (sched_extend_ready_list): Update.
143 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
145 Remove IA64 speculation tweaking flags
146 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
147 speculation tuning flags.
148 (msched-prefer-non-data-spec-insns,)
149 (msched-prefer-non-control-spec-insns): Obsolete options.
150 * haifa-sched.c (choose_ready): Remove handling of
151 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
152 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
153 and PREFER_NON_DATA_SPEC.
154 * sel-sched.c (process_spec_exprs): Remove handling of
155 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
157 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
159 Improve scheduling debug output
160 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
161 (advance_one_cycle): Update.
162 (schedule_insn, queue_to_ready): Add debug printouts.
163 (debug_ready_list_1): New static function.
164 (debug_ready_list): Update.
165 (max_issue): Add debug printouts.
166 (dump_insn_stream): New static function.
167 (schedule_block): Use it. Also better indent printouts.
169 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
171 Fix sched_insn debug counter
172 * haifa-sched.c (schedule_insn): Update.
173 (struct haifa_saved_data): Add nonscheduled_insns_begin.
174 (save_backtrack_point, restore_backtrack_point): Update.
175 (first_nonscheduled_insn): New static function.
176 (queue_to_ready, choose_ready): Use it.
177 (schedule_block): Init nonscheduled_insns_begin.
178 (sched_emit_insn): Update.
181 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
183 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
185 (aarch64_secondary_reload) : LikeWise.
186 (aarch64_class_max_nregs) : Remove CORE_REGS.
187 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
188 (REG_CLASS_NAMES) : Likewise.
189 (REG_CLASS_CONTENTS) : LikeWise.
190 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
192 2014-05-21 Guozhi Wei <carrot@google.com>
195 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
197 (vqdmulhq_n_s16): Likewise.
199 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
201 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
203 2014-05-21 Marek Polacek <polacek@redhat.com>
206 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
208 2014-05-21 Martin Jambor <mjambor@suse.cz>
210 * doc/invoke.texi (Optimize Options): Document parameters
211 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
212 ipa-cp-array-index-hint-bonus.
214 2014-05-21 Mark Wielaard <mjw@redhat.com>
217 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
218 version >= 3 or not strict DWARF.
219 * langhooks.h (struct lang_hooks_for_types): Add
220 enum_underlying_base_type.
221 * langhooks.c (lhd_enum_underlying_base_type): New function.
222 * gcc/langhooks.h (struct lang_hooks_for_types): Add
223 enum_underlying_base_type.
224 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
225 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
226 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
228 2014-05-21 Richard Biener <rguenther@suse.de>
230 * doc/invoke.texi (-flto-partition=): Document one and
233 2014-05-21 John Marino <gnugcc@marino.st>
235 * config.gcc (*-*-dragonfly*): New target.
236 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
237 * configure: Regenerate.
238 * config/dragonfly-stdint.h: New.
239 * config/dragonfly.h: New.
240 * config/dragonfly.opt: New.
241 * config/i386/dragonfly.h: New.
242 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
244 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
246 * tree.def (VOID_CST): New.
247 * tree-core.h (TI_VOID): New.
248 * tree.h (void_node): New.
249 * tree.c (tree_node_structure_for_code, tree_code_size)
250 (iterative_hash_expr): Handle VOID_CST.
251 (build_common_tree_nodes): Initialize void_node.
253 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
255 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
257 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
259 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
262 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
263 flag_reorder_blocks_and_partition.
264 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
266 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
269 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
271 (*addc_r_t): Add new insn_and_split.
273 2014-05-21 Jakub Jelinek <jakub@redhat.com>
276 * omp-low.c (handle_simd_reference): New function.
277 (lower_rec_input_clauses): Use it. Defer adding reference
278 initialization even for reduction without placeholder if in simd,
279 handle it properly later on.
281 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
283 PR tree-optimization/60899
284 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
285 assume all static symbols will have definition wile parsing and
286 check the do have definition later in compilation; check that
287 variable referring symbol will be output before concluding that
288 reference is safe; be conservative for referring local statics;
289 be more precise about when comdat is output in other partition.
291 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
294 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
295 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
296 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
298 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
300 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
301 set writeonly flag only for vars actually written to.
303 2014-05-20 Dehao Chen <dehao@google.com>
305 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
306 and callee count to get clone count.
307 * tree-inline.c (expand_call_inline): Use callee count instead of bb
310 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
312 PR rtl-optimization/61243
313 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
315 2014-05-20 Xinliang David Li <davidxl@google.com>
317 * cgraphunit.c (walk_polymorphic_call_targets): Add
318 dbgcnt and fopt-info support.
319 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
320 * ipa-devirt.c (ipa_devirt): Ditto.
321 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
323 * ipa.c (walk_polymorphic_call_targets): Ditto.
324 * gimple-fold.c (fold_gimple_assign): Ditto.
325 (gimple_fold_call): Ditto.
326 * dbgcnt.def: New counter.
328 2014-05-20 DJ Delorie <dj@redhat.com>
330 * config/msp430/msp430.md (split): Don't allow subregs when
331 splitting SImode adds.
332 (andneghi): Fix subtraction logic.
333 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
335 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
337 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
339 * except.c (switch_to_exception_section, resolve_unique_section,
340 get_named_text_section, default_function_rodata_section,
341 align_variable, get_block_for_decl, default_section_type_flags):
342 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
343 * symtab.c (symtab_add_to_same_comdat_group,
344 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
345 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
347 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
348 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
349 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
350 (c6x_function_in_section_p): Likewise.
351 * config/darwin.c (machopic_select_section): Likewise.
352 * config/arm/arm.c (arm_function_in_section_p): Likewise.
353 * config/mips/mips.c (mips_function_rodata_section): Likewise.
354 * config/mep/mep.c (mep_select_section): LIkewise.
355 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
357 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
359 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
360 EH region of calls to pure functions that can throw an exception.
361 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
362 (copy_reference_ops_from_call): Also copy the EH region of the call if
363 it can throw an exception.
365 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
367 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
368 nested VEC_SELECTs that are inverses of each other.
370 2014-05-20 Richard Biener <rguenther@suse.de>
372 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
374 (extract_and_process_scc_for_name): not here.
375 (cond_dom_walker::before_dom_children): Only process
376 stmts that end the BB in interesting ways.
377 (run_scc_vn): Mark param uses as visited.
379 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
381 * config/arm/arm.md (arith_shiftsi): Do not predicate for
384 2014-05-20 Nick Clifton <nickc@redhat.com>
386 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
387 (msp430_gimplify_va_arg_expr): New function.
388 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
390 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
391 operand 0 in order to prevent confusion about the number of
394 2014-05-20 Richard Biener <rguenther@suse.de>
396 PR tree-optimization/61221
397 * tree-ssa-pre.c (el_to_update): Remove.
398 (eliminate_dom_walker::before_dom_children): Handle released
399 VDEFs by value-numbering them to the associated VUSE. Update
400 stmt immediately for substituted call address.
401 (eliminate): Remove delayed stmt updating code.
402 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
403 possibly late re-numbered vuses.
404 (vn_reference_lookup_2): Adjust.
405 (vn_reference_lookup_pieces): Likewise.
406 (vn_reference_lookup): Likewise.
408 2014-05-20 Richard Biener <rguenther@suse.de>
410 * config.gcc: Remove need_64bit_hwint.
411 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
412 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
414 * config.in: Regenerate.
415 * configure: Likewise.
417 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
419 * doc/extend.texi: Create Label Attributes section,
420 move all label attributes into it and reference it.
422 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
424 * arm.c (thumb1_reorg): When scanning backwards skip anything
425 that's not a proper insn.
427 2014-05-19 Richard Biener <rguenther@suse.de>
429 PR tree-optimization/61221
430 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
431 Do nothing for unreachable blocks.
432 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
433 Improve unreachability detection.
435 2014-05-19 Richard Biener <rguenther@suse.de>
437 PR tree-optimization/61209
438 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
440 2014-05-19 Nick Clifton <nickc@redhat.com>
442 * except.c (init_eh): Fix computation of builtin setjmp buffer
443 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
445 2014-05-19 Richard Biener <rguenther@suse.de>
447 PR tree-optimization/61184
448 * tree-vrp.c (is_negative_overflow_infinity): Use
449 TREE_OVERFLOW_P and do that check first.
450 (is_positive_overflow_infinity): Likewise.
451 (is_overflow_infinity): Likewise.
452 (vrp_operand_equal_p): Properly treat operands with
453 differing overflow as not equal.
455 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
457 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
458 shift simplification where it was intended.
460 2014-05-19 Christian Bruel <christian.bruel@st.com>
463 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
465 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
468 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
471 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
473 * reg-notes.def (CROSSING_JUMP): Likewise.
474 * rtl.h (rtx_def): Update comment for jump flag.
475 (CROSSING_JUMP_P): Define.
476 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
477 of a REG_CROSSING_JUMP note.
478 * cfghooks.c (tidy_fallthru_edges): Likewise.
479 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
480 * emit-rtl.c (try_split): Likewise.
481 * haifa-sched.c (sched_create_recovery_edges): Likewise.
482 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
483 * jump.c (redirect_jump_2): Likewise.
484 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
485 (relax_delay_slots): Likewise.
486 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
488 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
489 * config/sh/sh.md (jump_compact): Likewise.
490 * bb-reorder.c (rotate_loop): Likewise.
491 (pass_duplicate_computed_gotos::execute): Likewise.
492 (add_reg_crossing_jump_notes): Rename to...
493 (update_crossing_jump_flags): ...this.
494 (pass_partition_blocks::execute): Update accordingly.
496 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
498 * tree.h: Remove extraneous template <>.
500 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
502 * ipa.c (symtab_remove_unreachable_nodes): Remove
503 symbol from comdat group if its body was eliminated.
504 (comdat_can_be_unshared_p_1): Static symbols can always
506 * symtab.c (symtab_remove_from_same_comdat_group): Break out
508 (symtab_unregister_node): ... this one.
509 (verify_symtab_base): More strict checking of comdats.
510 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
512 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
514 * tree-pass.h (make_pass_ipa_comdats): New pass.
515 * timevar.def (TV_IPA_COMDATS): New timevar.
516 * passes.def (pass_ipa_comdats): Add.
517 * Makefile.in (OBJS): Add ipa-comdats.o
518 * ipa-comdats.c: New file.
520 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
522 * ipa.c (update_visibility_by_resolution_info): New function.
523 (function_and_variable_visibility): Use it.
525 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
527 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
529 (FOR_EACH_DEFINED_SYMBOL): New macro.
530 (varpool_first_static_initializer, varpool_next_static_initializer,
531 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
532 (symtab_in_same_comdat_p): Correctly deal with inline functions.
534 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
536 * ggc-page.c (ggc_handle_finalizers): Add comment.
538 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
540 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
541 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
542 (ggc_internal_cleared_alloc): Likewise.
543 * ggc-page.c (finalizer): New class.
544 (vec_finalizer): Likewise.
545 (globals::finalizers): New member.
546 (globals::vec_finalizers): Likewise.
547 (ggc_internal_alloc): Record the finalizer if any for the block being
549 (ggc_handle_finalizers): New function.
550 (ggc_collect): Call ggc_handle_finalizers.
551 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
553 (ggc_internal_cleared_alloc): Likewise.
554 (finalize): New function.
555 (need_finalization_p): Likewise.
556 (ggc_alloc): Install the type's destructor as the finalizer if it
558 (ggc_cleared_alloc): Likewise.
559 (ggc_vec_alloc): Likewise.
560 (ggc_cleared_vec_alloc): Likewise.
562 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
564 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
566 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
568 * alias.c (record_alias_subset): Adjust.
569 * bitmap.c (bitmap_element_allocate): Likewise.
570 (bitmap_gc_alloc_stat): Likewise.
571 * cfg.c (init_flow): Likewise.
572 (alloc_block): Likewise.
573 (unchecked_make_edge): Likewise.
574 * cfgloop.c (alloc_loop): Likewise.
575 (flow_loops_find): Likewise.
576 (rescan_loop_exit): Likewise.
577 * cfgrtl.c (init_rtl_bb_info): Likewise.
578 * cgraph.c (insert_new_cgraph_node_version): Likewise.
579 (cgraph_allocate_node): Likewise.
580 (cgraph_create_edge_1): Likewise.
581 (cgraph_allocate_init_indirect_info): Likewise.
582 * cgraphclones.c (cgraph_clone_edge): Likewise.
583 * cgraphunit.c (add_asm_node): Likewise.
584 (init_lowered_empty_function): Likewise.
585 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
586 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
587 (alpha_use_linkage): Likewise.
588 * config/arc/arc.c (arc_init_machine_status): Likewise.
589 * config/arm/arm.c (arm_init_machine_status): Likewise.
590 * config/avr/avr.c (avr_init_machine_status): Likewise.
591 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
592 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
593 * config/cris/cris.c (cris_init_machine_status): Likewise.
594 * config/darwin.c (machopic_indirection_name): Likewise.
595 (darwin_build_constant_cfstring): Likewise.
596 (darwin_enter_string_into_cfstring_table): Likewise.
597 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
598 * config/frv/frv.c (frv_init_machine_status): Likewise.
599 * config/i386/i386.c (get_dllimport_decl): Likewise.
600 (ix86_init_machine_status): Likewise.
601 (assign_386_stack_local): Likewise.
602 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
603 (i386_pe_maybe_record_exported_symbol): Likewise.
604 (i386_pe_record_stub): Likewise.
605 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
606 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
607 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
608 (m32c_note_pragma_address): Likewise.
609 * config/mep/mep.c (mep_init_machine_status): Likewise.
610 (mep_note_pragma_flag): Likewise.
611 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
612 (mips16_local_alias): Likewise.
613 (mips_init_machine_status): Likewise.
614 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
615 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
616 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
617 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
618 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
619 * config/pa/pa.c (pa_init_machine_status): Likewise.
620 (pa_get_deferred_plabel): Likewise.
621 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
622 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
623 (rs6000_init_machine_status): Likewise.
624 (output_toc): Likewise.
625 * config/s390/s390.c (s390_init_machine_status): Likewise.
626 * config/score/score.c (score_output_external): Likewise.
627 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
628 * config/spu/spu.c (spu_init_machine_status): Likewise.
629 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
630 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
631 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
632 * coverage.c (coverage_end_function): Likewise.
633 * dbxout.c (dbxout_init): Likewise.
634 * doc/gty.texi: Don't mention variable_size attribute.
635 * dwarf2cfi.c (new_cfi): Adjust.
636 (new_cfi_row): Likewise.
637 (copy_cfi_row): Likewise.
638 (create_cie_data): Likewise.
639 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
640 (new_loc_descr): Likewise.
641 (find_AT_string_in_table): Likewise.
642 (add_addr_table_entry): Likewise.
644 (add_var_loc_to_decl): Likewise.
645 (clone_die): Likewise.
646 (clone_as_declaration): Likewise.
647 (break_out_comdat_types): Likewise.
648 (new_loc_list): Likewise.
649 (add_loc_descr_to_each): Likewise.
650 (add_location_or_const_value_attribute): Likewise.
651 (add_linkage_name): Likewise.
652 (lookup_filename): Likewise.
653 (dwarf2out_var_location): Likewise.
654 (new_line_info_table): Likewise.
655 (dwarf2out_init): Likewise.
656 (mem_loc_descriptor): Likewise.
657 (loc_descriptor): Likewise.
658 (add_const_value_attribute): Likewise.
659 (tree_add_const_value_attribute): Likewise.
660 (comp_dir_string): Likewise.
661 (dwarf2out_vms_debug_main_pointer): Likewise.
662 (string_cst_pool_decl): Likewise.
663 * emit-rtl.c (set_mem_attrs): Likewise.
664 (get_reg_attrs): Likewise.
665 (start_sequence): Likewise.
666 (init_emit): Likewise.
667 (init_emit_regs): Likewise.
668 * except.c (init_eh_for_function): Likewise.
669 (gen_eh_region): Likewise.
670 (gen_eh_region_catch): Likewise.
671 (gen_eh_landing_pad): Likewise.
672 (add_call_site): Likewise.
673 * function.c (add_frame_space): Likewise.
674 (insert_temp_slot_address): Likewise.
675 (assign_stack_temp_for_type): Likewise.
676 (get_hard_reg_initial_val): Likewise.
677 (allocate_struct_function): Likewise.
678 (prepare_function_start): Likewise.
679 (types_used_by_var_decl_insert): Likewise.
680 * gengtype.c (variable_size_p): Remove function.
681 (enum alloc_quantity): Remove enum.
682 (write_typed_alloc_def): Remove function.
683 (write_typed_struct_alloc_def): Likewise.
684 (write_typed_typedef_alloc_def): Likewise.
685 (write_typed_alloc_defns): Likewise.
687 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
688 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
689 * ggc.h (ggc_alloc): new function.
690 (ggc_cleared_alloc): Likewise.
691 (ggc_vec_alloc): Template on type of vector element, and remove
692 element size argument.
693 (ggc_cleared_vec_alloc): Likewise.
694 * gimple.c (gimple_build_omp_for): Adjust.
695 (gimple_copy): Likewise.
696 * ipa-cp.c (get_replacement_map): Likewise.
697 (find_aggregate_values_for_callers_subset): Likewise.
698 (known_aggs_to_agg_replacement_list): Likewise.
699 * ipa-devirt.c (get_odr_type): Likewise.
700 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
701 (read_agg_replacement_chain): Likewise.
702 * loop-iv.c (get_simple_loop_desc): Likewise.
703 * lto-cgraph.c (input_node_opt_summary): Likewise.
704 * lto-section-in.c (lto_new_in_decl_state): Likewise.
705 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
706 (input_eh_region): Likewise.
707 (input_eh_lp): Likewise.
708 (input_cfg): Likewise.
709 * optabs.c (set_optab_libfunc): Likewise.
710 (init_tree_optimization_optabs): Likewise.
711 (set_conv_libfunc): Likewise.
712 * passes.c (do_per_function_toporder): Likewise.
713 * rtl.h: Don't use variable_size gty attribute.
714 * sese.c (if_region_set_false_region): Adjust.
715 * stringpool.c (gt_pch_save_stringpool): Likewise.
716 * target-globals.c (save_target_globals): Likewise.
717 * toplev.c (general_init): Likewise.
718 * trans-mem.c (record_tm_replacement): Likewise.
719 (split_bb_make_tm_edge): Likewise.
720 * tree-cfg.c (move_sese_region_to_fn): Likewise.
721 * tree-data-ref.h (lambda_vector_new): Likewise.
722 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
723 * tree-iterator.c (tsi_link_before): Likewise.
724 (tsi_link_after): Likewise.
725 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
726 * tree-ssa-loop-niter.c (record_estimate): Likewise.
727 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
728 * tree-ssa-operands.h: Don't use variable_size gty attribute.
729 * tree-ssa.c (init_tree_ssa): Adjust.
730 * tree-ssanames.c (set_range_info): Likewise.
731 (get_ptr_info): Likewise.
732 (duplicate_ssa_name_ptr_info): Likewise.
733 (duplicate_ssa_name_range_info): Likewise.
734 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
735 (unpack_ts_fixed_cst_value_fields): Likewise.
736 * tree.c (build_fixed): Likewise.
737 (build_real): Likewise.
738 (build_string): Likewise.
739 (decl_priority_info): Likewise.
740 (decl_debug_expr_insert): Likewise.
741 (decl_value_expr_insert): Likewise.
742 (decl_debug_args_insert): Likewise.
743 (type_hash_add): Likewise.
744 (build_omp_clause): Likewise.
745 * ubsan.c (decl_for_type_insert): Likewise.
746 * varasm.c (get_unnamed_section): Likewise.
747 (get_noswitch_section): Likewise.
748 (get_section): Likewise.
749 (get_block_for_section): Likewise.
750 (create_block_symbol): Likewise.
751 (build_constant_desc): Likewise.
752 (create_constant_pool): Likewise.
753 (force_const_mem): Likewise.
754 (record_tm_clone_pair): Likewise.
755 * varpool.c (varpool_create_empty_node): Likewise.
757 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
759 * dwarf2out.c (tree_add_const_value_attribute): Call
760 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
761 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
762 instead of ggc_internal_<x>alloc_stat.
763 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
764 (ggc_realloc): Likewise.
765 * ggc-none.c (ggc_internal_alloc): Likewise.
766 (ggc_internal_cleared_alloc): Likewise.
767 * ggc-page.c: Likewise.
768 * ggc.h (ggc_internal_alloc_stat): Likewise.
769 (ggc_internal_alloc): Remove macro.
770 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
771 (ggc_internal_cleared_alloc): Remove macro.
772 (GGC_RESIZEVEC): Adjust.
773 (ggc_resizevar): Remove macro.
774 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
775 (ggc_internal_cleared_vec_alloc_stat): Likewise.
776 (ggc_internal_vec_cleared_alloc): Remove macro.
777 (ggc_alloc_atomic_stat): Drop _stat suffix.
778 (ggc_alloc_atomic): Remove macro.
779 (ggc_alloc_cleared_atomic): Remove macro.
780 (ggc_alloc_string_stat): Drop _stat suffix.
781 (ggc_alloc_string): Remove macro.
782 (ggc_alloc_rtx_def_stat): Adjust.
783 (ggc_alloc_tree_node_stat): Likewise.
784 (ggc_alloc_cleared_tree_node_stat): Likewise.
785 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
786 (ggc_alloc_cleared_simd_clone_stat): Likewise.
787 * gimple.c (gimple_build_omp_for): Likewise.
788 (gimple_copy): Likewise.
789 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
790 * toplev.c (realloc_for_line_map): Adjust.
791 * tree-data-ref.h (lambda_vector_new): Likewise.
792 * tree-phinodes.c (allocate_phi_node): Likewise.
793 * tree.c (grow_tree_vec_stat): Likewise.
794 * vec.h (va_gc::reserve): Adjust.
796 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
798 * config/microblaze/microblaze.c (break_handler): New Declaration.
799 (microblaze_break_function_p,microblaze_is_break_handler): New.
800 (compute_frame_size): Use microblaze_break_function_p.
801 Add the test of break_handler.
802 (microblaze_function_prologue) : Add the test of variable
803 break_handler. Check the fnname by BREAK_HANDLER_NAME.
804 (microblaze_function_epilogue) : Add the test of break_handler.
805 (microblaze_globalize_label) : Add the test of break_handler.
806 Check the name by BREAK_HANDLER_NAME.
808 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
810 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
811 microblaze_is_break_handler test.
812 (call_internal1,call_value_intern): Use microblaze_break_function_p.
815 * config/microblaze/microblaze-protos.h
816 (microblaze_break_function_p,microblaze_is_break_handler):
819 * doc/extend.texi (MicroBlaze break_handler Functions): Document
820 new MicroBlaze break_handler functions.
822 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
824 * doc/extend.texi (Size of an asm): Move node text according
825 to its @menu entry position.
827 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
829 PR tree-optimization/61140
830 PR tree-optimization/61150
831 PR tree-optimization/61197
832 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
834 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
836 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
838 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
840 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
841 __SIZEOF_INT128__ is defined.
843 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
845 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
846 (rs6000_delegitimize_address): Use it.
848 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
850 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
851 inplace argument. Store the new address in the original MEM when true.
852 * emit-rtl.c (change_address_1): Likewise.
853 (adjust_address_1, adjust_automodify_address_1, offset_address):
855 * rtl.h (plus_constant): Add an inplace argument.
856 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
857 when true. Avoid generating (plus X (const_int 0)).
858 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
859 in-place. Pass true to plus_constant.
860 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
862 2014-05-16 Dehao Chen <dehao@google.com>
864 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
866 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
869 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
871 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
873 2014-05-16 Dehao Chen <dehao@google.com>
875 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
876 optimize_function_for_size_p.
877 * regs.h (REG_FREQ_FROM_BB): Likewise.
879 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
882 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
883 negt_reg_operand cases.
884 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
886 * config/sh/predicates.md (cbranch_treg_value): Simplify.
888 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
890 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
893 2014-05-16 David Malcolm <dmalcolm@redhat.com>
896 2014-04-29 David Malcolm <dmalcolm@redhat.com>
898 * tree-cfg.c (dump_function_to_file): Dump the return type of
899 functions, in a line to itself before the function body, mimicking
900 the layout of a C function.
902 2014-05-16 Dehao Chen <dehao@google.com>
904 * cfghooks.c (make_forwarder_block): Use direct computation to
905 get fall-through edge's count and frequency.
907 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
909 * config/arc/arc.c (arc_init): Fix typo in error message.
910 * config/i386/i386.c (ix86_expand_builtin): Likewise.
911 (split_stack_prologue_scratch_regno): Likewise.
912 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
913 word from error message.
915 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
917 * ira-costs.c: Fix typo in comment.
919 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
921 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
923 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
925 * varpool.c (dump_varpool_node): Dump write-only flag.
926 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
928 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
929 write-only variables.
930 * ipa.c (process_references): New function.
931 (set_readonly_bit): New function.
932 (set_writeonly_bit): New function.
933 (clear_addressable_bit): New function.
934 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
935 fix handling of aliases.
936 * cgraph.h (struct varpool_node): Add writeonly flag.
938 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
940 PR rtl-optimization/60969
941 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
942 Calculate costs for this case.
944 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
946 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
947 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
949 2014-05-16 Richard Biener <rguenther@suse.de>
951 PR tree-optimization/61194
952 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
953 bool patterns ending in a COND_EXPR.
955 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
957 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
959 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
961 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
962 where we were unable to cost an RTX.
964 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
966 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
969 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
970 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
972 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
974 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
975 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
977 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
978 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
980 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
981 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
983 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
986 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
987 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
989 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
992 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
993 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
995 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
996 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
998 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
999 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1001 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1004 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1005 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1007 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1008 ZERO_EXTEND and SIGN_EXTEND better.
1010 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1011 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1013 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1016 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1017 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1019 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1020 costs when costing loads and stores to memory.
1022 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1023 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1025 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1028 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1030 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1032 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1033 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1035 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1037 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1039 (aarch64_rtx_mult_cost): New.
1040 (aarch64_rtx_costs): Use it, refactor as appropriate.
1042 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1043 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1045 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1046 emit instructions, return number of instructions which would
1048 (aarch64_add_constant): Update call to aarch64_build_constant.
1049 (aarch64_output_mi_thunk): Likewise.
1050 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1053 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1055 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1056 (TARGET_RTX_COSTS): Call it.
1058 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1060 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1061 (cortexa57_vector_cost): Likewise.
1062 (cortexa57_tunings): Use them.
1064 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1066 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1067 (cpu_addrcost_table): Use it.
1068 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1069 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1072 2014-05-16 Richard Biener <rguenther@suse.de>
1074 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1075 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1076 (visit_phi): Ignore edges marked as not executable.
1077 (class cond_dom_walker): New.
1078 (cond_dom_walker::before_dom_children): Value-number
1079 control statements and mark successor edges as not
1080 executable if possible.
1081 (run_scc_vn): First walk all control statements in
1082 dominator order, marking edges as not executable.
1083 * tree-inline.c (copy_edges_for_bb): Be not confused
1084 about random edge flags.
1086 2014-05-16 Richard Biener <rguenther@suse.de>
1088 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1090 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1093 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1094 (__TM_simple_begin): Use it.
1095 (__TM_begin): Likewise.
1097 2014-05-15 Martin Jambor <mjambor@suse.cz>
1100 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1101 type_preserved flag when the indirect edge is polymorphic.
1103 2014-05-15 Martin Jambor <mjambor@suse.cz>
1105 PR tree-optimization/61090
1106 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1107 build_ref_for_model.
1109 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1111 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1112 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1114 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1116 PR tree-optimization/61158
1117 * fold-const.c (fold_binary_loc): If X is zero-extended and
1118 shiftc >= prec, make sure zerobits is all ones instead of
1119 invoking undefined behavior.
1121 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1123 * regcprop.h: New file.
1124 * regcprop.c (skip_debug_insn_p): New decl.
1125 (replace_oldest_value_reg): Check skip_debug_insn_p.
1126 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1127 * shrink-wrap.c: Include regcprop.h.
1128 (prepare_shrink_wrap): Call
1129 copyprop_hardreg_forward_bb_without_debug_insn.
1131 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1133 * shrink-wrap.h: Update comment.
1134 * shrink-wrap.c: Update comment.
1135 (next_block_for_reg): Rename to live_edge_for_reg.
1136 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1137 (move_insn_for_shrink_wrap): Split live_edge.
1138 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1140 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1142 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1144 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1145 * config/sparc/sparc.md (fptype_ut699): New attribute.
1146 (in_branch_delay): Return false if -mfix-ut699 is specified and
1147 fptype_ut699 is set to single.
1148 (truncdfsf2): Add fptype_ut699 attribute.
1149 (fix_truncdfsi2): Likewise.
1150 (floatsisf2): Change fptype attribute.
1151 (fix_truncsfsi2): Likewise.
1152 (negtf2_notv9): Delete.
1153 (negtf2_v9): Likewise.
1154 (negtf2_hq): New instruction.
1155 (negtf2): New instruction and splitter.
1156 (negdf2_notv9): Rewrite.
1157 (abstf2_notv9): Delete.
1158 (abstf2_hq_v9): Likewise.
1159 (abstf2_v9): Likewise.
1160 (abstf2_hq): New instruction.
1161 (abstf2): New instruction and splitter.
1162 (absdf2_notv9): Rewrite.
1164 2014-05-14 Cary Coutant <ccoutant@google.com>
1167 * opts.c (common_handle_option): Don't special-case "-g".
1168 (set_debug_level): Default to at least level 2 with "-g".
1170 2014-05-14 DJ Delorie <dj@redhat.com>
1172 * config/msp430/msp430.c (msp430_builtin): Add
1173 MSP430_BUILTIN_DELAY_CYCLES.
1174 (msp430_init_builtins): Register void __delay_cycles(long long).
1175 (msp430_builtin_decl): Add it.
1176 (cg_magic_constant): New.
1177 (msp430_expand_delay_cycles): New.
1178 (msp430_expand_builtin): Call it.
1179 (msp430_print_operand_raw): Change integer printing from "int" to
1181 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1182 (delay_cycles_start): New.
1183 (delay_cycles_end): New.
1184 (delay_cycles_32): New.
1185 (delay_cycles_32x): New.
1186 (delay_cycles_16): New.
1187 (delay_cycles_16x): New.
1188 (delay_cycles_2): New.
1189 (delay_cycles_1): New.
1190 * doc/extend.texi: Document __delay_cycles().
1192 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1194 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1195 length attribute computation.
1197 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1200 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1202 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1205 * config/sparc/sparc.md: Fix types of low and high in DI constant
1206 splitter. Use gen_int_mode in some other splitters.
1208 2014-05-14 Martin Jambor <mjambor@suse.cz>
1211 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1213 2014-05-14 James Norris <jnorris@codesourcery.com>
1215 * omp-low.c (expand_parallel_call): Remove shadow variable.
1216 (expand_omp_taskreg): Likewise.
1218 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1220 * common/config/i386/i386-common.c
1221 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1222 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1223 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1224 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1225 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1226 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1227 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1228 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1229 xsavecintrin.h, xsavesintrin.h.
1230 (x86_64-*-*): Ditto.
1231 * config/i386/clflushoptintrin.h: New.
1232 * config/i386/xsavecintrin.h: Ditto.
1233 * config/i386/xsavesintrin.h: Ditto.
1234 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1235 (bit_XSAVES): Ditto.
1236 (bit_XSAVES): Ditto.
1237 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1238 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1240 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1241 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1242 OPTION_MASK_ISA_XSAVES.
1243 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1245 (PTA_CLFLUSHOPT) Define.
1246 (PTA_XSAVEC): Ditto.
1247 (PTA_XSAVES): Ditto.
1248 (ix86_option_override_internal): Handle new options.
1249 (ix86_valid_target_attribute_inner_p): Ditto.
1250 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1251 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1252 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1253 (bdesc_special_args): Add __builtin_ia32_xsaves,
1254 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1255 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1256 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1257 (ix86_expand_builtin): Handle new builtins.
1258 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1259 (TARGET_CLFLUSHOPT_P): Ditto.
1260 (TARGET_XSAVEC): Ditto.
1261 (TARGET_XSAVEC_P): Ditto.
1262 (TARGET_XSAVES): Ditto.
1263 (TARGET_XSAVES_P): Ditto.
1264 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1265 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1266 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1268 (ANY_XRSTOR64): Ditto.
1270 (xrstor): Change into <xrstor>.
1271 (xrstor_rex64): Change into <xrstor>_rex64.
1272 (xrstor64): Change into <xrstor>64
1274 * config/i386/i386.opt (mclflushopt): New.
1277 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1279 * doc/invoke.texi: Document new options.
1281 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1283 PR rtl-optimization/60866
1284 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1285 Default it to -1. Pass it down to init_simplejump_data.
1286 (init_simplejump_data): New parameter old_seqno. Pass it down
1287 to get_seqno_for_a_jump.
1288 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1289 initializing new jump seqno as a last resort. Add comment.
1290 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1291 jump and pass it down to sel_init_new_insn.
1292 (sel_redirect_edge_and_branch_force): Likewise.
1294 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1296 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1297 shifted values to avoid build warning.
1299 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1301 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1302 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1303 (cfg_layout_merge_blocks): Likewise.
1304 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1306 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1308 PR rtl-optimization/60901
1309 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1310 bb predecessor belongs to the same scheduling region. Adjust comment.
1312 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1314 * doc/sourcebuild.texi: (dfp_hw): Document.
1315 (p8vector_hw): Likewise.
1316 (powerpc_eabi_ok): Likewise.
1317 (powerpc_elfv2): Likewise.
1318 (powerpc_htm_ok): Likewise.
1319 (ppc_recip_hw): Likewise.
1322 2014-05-13 Cary Coutant <ccoutant@google.com>
1324 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1326 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1328 * gengtype-parse.c (require3): Eliminate in favor of...
1330 (require_template_declaration): Update to support optional single *
1333 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1334 (create_user_defined_type): Handle a single level of explicit
1335 pointerness within template arguments.
1336 (struct write_types_data): Add field "kind".
1337 (filter_type_name): Handle "*" character.
1338 (write_user_func_for_structure_ptr): Require a write_types_data
1339 rather than just a prefix string, so that we can look up the kind
1340 of the wtd and use it as an index into wrote_user_func_for_ptr,
1341 ensuring that such functions are written at most once. Support
1342 subclasses by invoking the marking function of the ultimate base class.
1343 (write_user_func_for_structure_body): Require a write_types_data
1344 rather than just a prefix string, so that we can pass this to
1345 write_user_func_for_structure_ptr.
1346 (write_func_for_structure): Likewise.
1347 (ggc_wtd): Add initializer of new "kind" field.
1348 (pch_wtd): Likewise.
1350 * gengtype.h (enum write_types_kinds): New.
1351 (struct type): Add field wrote_user_func_for_ptr to the "s"
1354 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1356 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1357 instead of const_binop.
1358 (fold_binary_loc): Likewise.
1360 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1362 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1363 calculation to match get_ref_base_and_extent.
1365 2014-05-13 Catherine Moore <clm@codesourcery.com>
1366 Sandra Loosemore <sandra@codesourcery.com>
1368 * configure.ac: Fix assembly for explicit JALR relocation check.
1369 * configure: Regenerate.
1371 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1373 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1374 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1375 Remove associated type declarations and initialisations.
1376 (arm_expand_neon_builtin): Likewise.
1377 (neon_emit_pair_result_insn): Delete.
1378 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1379 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1380 (neon_vzip<mode>): Likewise.
1381 (neon_vuzp<mode>): Likewise.
1383 2014-05-13 Richard Biener <rguenther@suse.de>
1386 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1387 it needs revisiting whether the call still may be tail-called.
1389 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1391 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1392 * rtl.h (block_symbol): Reduce number of fields to 2.
1393 (rtx_def): Add u2.symbol_ref_flags.
1394 (SYMBOL_REF_FLAGS): Use it.
1395 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1396 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1397 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1398 Lower index of SYMBOL_REF_DATA.
1399 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1400 Print SYMBOL_REF_FLAGS at the same time.
1401 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1403 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1405 * rtl.def (VAR_LOCATION): Remove "i" field.
1406 * rtl.h (rtx_def): Add u2.var_location_status.
1407 (PAT_VAR_LOCATION_STATUS): Use it.
1408 (gen_rtx_VAR_LOCATION): Declare.
1409 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1410 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1411 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1413 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1415 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1416 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1418 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1420 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1421 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1422 * rtl.h (rtx_def): Add insn_uid to u2 field.
1423 (RTX_FLAG_CHECK8): Delete in favor of...
1424 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1425 (INSN_DELETED_P): Update accordingly.
1426 (INSN_UID): Use u2.insn_uid.
1427 (INSN_CHAIN_CODE_P): Define.
1428 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1429 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1430 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1431 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1432 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1433 indices accordingly.
1434 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1435 Update indices for insn-chain rtxes.
1436 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1437 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1438 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1439 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1440 * combine.c (try_combine): Likewise.
1441 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1443 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1445 * rtl.def (REG): Remove middle field.
1446 * rtl.h (rtx_def): Add orignal_regno to u2.
1447 (ORIGINAL_REGNO): Use it instead of field 1.
1448 (REG_ATTRS): Lower field index accordingly.
1449 * gengtype.c (adjust_field_rtx_def): Remove handling of
1450 ORIGINAL_REGNO. Move REG_ATTRS index down.
1451 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1452 code that prints the REGNO.
1454 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1456 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1459 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1461 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1463 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1465 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1466 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1468 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1470 * config/aarch64/aarch64-protos.h
1471 (aarch64_hard_regno_caller_save_mode): New prototype.
1472 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1474 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1476 2014-05-13 Christian Bruel <christian.bruel@st.com>
1478 * target.def (mode_switching): New hook vector.
1479 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1480 (mode_exit, modepriority_to_mode): Likewise.
1481 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1482 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1483 * target.h: Include tm.h and hard-reg-set.h.
1484 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1485 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1486 * doc/tm.texi Regenerate.
1487 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1488 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1489 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1490 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1491 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1492 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1493 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1494 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1495 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1496 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1497 (ix86_emit_mode_set): Hookify.
1498 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1500 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1501 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1502 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1503 (epiphany_mode_priority_to_mode): Remove declaration.
1504 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1505 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1506 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1508 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1509 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1510 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1512 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1515 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1516 is const0_rtx, return immediately. Don't test count == 0 when
1519 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1521 * Makefile.in: add shrink-wrap.o.
1522 * config/i386/i386.c: include "shrink-wrap.h"
1523 * function.c: Likewise.
1524 (requires_stack_frame_p, next_block_for_reg,
1525 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1526 dup_block_and_redirect): Move to shrink-wrap.c
1527 (thread_prologue_and_epilogue_insns): Extract three code segments
1528 as functions in shrink-wrap.c
1529 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1531 * shrink-wrap.c: New file.
1532 * shrink-wrap.h: New file.
1534 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1536 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1537 reference to Solaris.
1539 2014-05-12 Mike Stump <mikestump@comcast.net>
1542 * genattrtab.c (filename): Add.
1543 (convert_set_attr_alternative): Improve error message.
1544 (check_defs): Restore read_md_filename for error messages.
1545 (gen_insn): Save filename.
1547 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1549 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1550 -fno-local-ivars and -fivar-visibility.
1551 * c-family/c.opt: Make -Wshadow also implicitly enable
1554 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1556 * doc/tm.texi: Remove reference to deleted macro.
1557 * doc/tm.texi.in: Likewise.
1559 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1562 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1565 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1567 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1568 * config/arm/aout.h (License): Same.
1569 * config/arm/bpabi.h (License): Same.
1570 * config/arm/elf.h (License): Same.
1571 * config/arm/linux-elf.h (License): Same.
1572 * config/arm/linux-gas.h (License): Same.
1573 * config/arm/netbsd-elf.h (License): Same.
1574 * config/arm/uclinux-eabi.h (License): Same.
1575 * config/arm/uclinux-elf.h (License): Same.
1576 * config/arm/vxworks.h (License): Same.
1578 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1580 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1581 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1582 number of operands to 3.
1583 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1584 * tree-nested.c (convert_nonlocal_omp_clauses,
1585 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1586 * gimplify.c (gimplify_scan_omp_clauses): Handle
1587 OMP_CLAUSE_LINEAR_STMT.
1588 * omp-low.c (lower_rec_input_clauses): Fix typo.
1589 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1590 cast between Fortran boolean_type_node and C _Bool if
1593 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1595 PR tree-optimization/61136
1596 * wide-int.h (multiple_of_p): Define a version that doesn't return
1598 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1599 integer_zerop/const_binop pair.
1600 (multiple_of_p): Likewise, converting both operands to widest_int
1603 2014-05-09 Teresa Johnson <tejohnson@google.com>
1605 * cgraphunit.c (analyze_functions): Use correct dump file.
1607 2014-05-09 Florian Weimer <fweimer@redhat.com>
1609 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1611 (stack_protect_return_slot_p): New function.
1612 (expand_used_vars): Call stack_protect_decl_p and
1613 stack_protect_return_slot_p for -fstack-protector-strong.
1615 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1618 * fold-const.c (fold_binary_loc): Changed width of mask.
1620 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1622 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1623 unsigned int initializers for regno_in, regno_out.
1625 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1628 * config/avr/avr.md (cc): Add new attribute set_vzn.
1629 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1630 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1631 with INC, DEC or NEG.
1632 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1633 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1634 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1636 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1639 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1641 * wide-int.cc (UTItype): Define.
1642 (UDWtype): Define for appropriate W_TYPE_SIZE.
1644 2014-05-09 Richard Biener <rguenther@suse.de>
1646 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1647 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1648 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1649 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1650 ssa_propagate): Adjust.
1652 2014-05-08 Jeff Law <law@redhat.com>
1654 PR tree-optimization/61009
1655 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1656 tri-state rather than a boolean. When a block is too big to
1657 thread through, inform caller via negative return value.
1658 (thread_across_edge): If a block was too big for normal threading,
1659 then it's too big for a joiner too, so remove temporary equivalences
1660 and return immediately.
1662 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1663 Matthias Klose <doko@ubuntu.com>
1666 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1668 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1671 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1673 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1676 * config/alpha/alpha.c: Include gimple-iterator.h.
1677 (alpha_gimple_fold_builtin): New function. Move
1678 ALPHA_BUILTIN_UMULH folding from ...
1679 (alpha_fold_builtin): ... here.
1680 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1682 2014-05-08 Wei Mi <wmi@google.com>
1685 * config/i386/i386.c (ix86_compute_frame_layout): Update
1686 preferred_stack_boundary for call, expanded from tls descriptor.
1687 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1688 to depend on SP register.
1689 (*tls_local_dynamic_base_32_gnu): Ditto.
1690 (*tls_local_dynamic_32_once): Ditto.
1691 (tls_global_dynamic_64_<mode>): Set
1692 ix86_tls_descriptor_calls_expanded_in_cfun.
1693 (tls_local_dynamic_base_64_<mode>): Ditto.
1694 (tls_global_dynamic_32): Set
1695 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1696 to depend on SP register.
1697 (tls_local_dynamic_base_32): Ditto.
1699 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1701 * config/arm/arm_neon.h: Update comment.
1702 * config/arm/neon-docgen.ml: Delete.
1703 * config/arm/neon-gen.ml: Delete.
1704 * doc/arm-neon-intrinsics.texi: Update comment.
1706 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1708 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1710 (vand, vorr, veor, vorn, vbic): Remove.
1711 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1713 (neon_vsub_unspec): Likewise.
1714 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1716 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1718 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1719 (vadd_s16): Likewise.
1720 (vadd_s32): Likewise.
1721 (vadd_f32): Likewise.
1722 (vadd_u8): Likewise.
1723 (vadd_u16): Likewise.
1724 (vadd_u32): Likewise.
1725 (vadd_s64): Likewise.
1726 (vadd_u64): Likewise.
1727 (vaddq_s8): Likewise.
1728 (vaddq_s16): Likewise.
1729 (vaddq_s32): Likewise.
1730 (vaddq_s64): Likewise.
1731 (vaddq_f32): Likewise.
1732 (vaddq_u8): Likewise.
1733 (vaddq_u16): Likewise.
1734 (vaddq_u32): Likewise.
1735 (vaddq_u64): Likewise.
1736 (vmul_s8): Likewise.
1737 (vmul_s16): Likewise.
1738 (vmul_s32): Likewise.
1739 (vmul_f32): Likewise.
1740 (vmul_u8): Likewise.
1741 (vmul_u16): Likewise.
1742 (vmul_u32): Likewise.
1743 (vmul_p8): Likewise.
1744 (vmulq_s8): Likewise.
1745 (vmulq_s16): Likewise.
1746 (vmulq_s32): Likewise.
1747 (vmulq_f32): Likewise.
1748 (vmulq_u8): Likewise.
1749 (vmulq_u16): Likewise.
1750 (vmulq_u32): Likewise.
1751 (vsub_s8): Likewise.
1752 (vsub_s16): Likewise.
1753 (vsub_s32): Likewise.
1754 (vsub_f32): Likewise.
1755 (vsub_u8): Likewise.
1756 (vsub_u16): Likewise.
1757 (vsub_u32): Likewise.
1758 (vsub_s64): Likewise.
1759 (vsub_u64): Likewise.
1760 (vsubq_s8): Likewise.
1761 (vsubq_s16): Likewise.
1762 (vsubq_s32): Likewise.
1763 (vsubq_s64): Likewise.
1764 (vsubq_f32): Likewise.
1765 (vsubq_u8): Likewise.
1766 (vsubq_u16): Likewise.
1767 (vsubq_u32): Likewise.
1768 (vsubq_u64): Likewise.
1769 (vand_s8): Likewise.
1770 (vand_s16): Likewise.
1771 (vand_s32): Likewise.
1772 (vand_u8): Likewise.
1773 (vand_u16): Likewise.
1774 (vand_u32): Likewise.
1775 (vand_s64): Likewise.
1776 (vand_u64): Likewise.
1777 (vandq_s8): Likewise.
1778 (vandq_s16): Likewise.
1779 (vandq_s32): Likewise.
1780 (vandq_s64): Likewise.
1781 (vandq_u8): Likewise.
1782 (vandq_u16): Likewise.
1783 (vandq_u32): Likewise.
1784 (vandq_u64): Likewise.
1785 (vorr_s8): Likewise.
1786 (vorr_s16): Likewise.
1787 (vorr_s32): Likewise.
1788 (vorr_u8): Likewise.
1789 (vorr_u16): Likewise.
1790 (vorr_u32): Likewise.
1791 (vorr_s64): Likewise.
1792 (vorr_u64): Likewise.
1793 (vorrq_s8): Likewise.
1794 (vorrq_s16): Likewise.
1795 (vorrq_s32): Likewise.
1796 (vorrq_s64): Likewise.
1797 (vorrq_u8): Likewise.
1798 (vorrq_u16): Likewise.
1799 (vorrq_u32): Likewise.
1800 (vorrq_u64): Likewise.
1801 (veor_s8): Likewise.
1802 (veor_s16): Likewise.
1803 (veor_s32): Likewise.
1804 (veor_u8): Likewise.
1805 (veor_u16): Likewise.
1806 (veor_u32): Likewise.
1807 (veor_s64): Likewise.
1808 (veor_u64): Likewise.
1809 (veorq_s8): Likewise.
1810 (veorq_s16): Likewise.
1811 (veorq_s32): Likewise.
1812 (veorq_s64): Likewise.
1813 (veorq_u8): Likewise.
1814 (veorq_u16): Likewise.
1815 (veorq_u32): Likewise.
1816 (veorq_u64): Likewise.
1817 (vbic_s8): Likewise.
1818 (vbic_s16): Likewise.
1819 (vbic_s32): Likewise.
1820 (vbic_u8): Likewise.
1821 (vbic_u16): Likewise.
1822 (vbic_u32): Likewise.
1823 (vbic_s64): Likewise.
1824 (vbic_u64): Likewise.
1825 (vbicq_s8): Likewise.
1826 (vbicq_s16): Likewise.
1827 (vbicq_s32): Likewise.
1828 (vbicq_s64): Likewise.
1829 (vbicq_u8): Likewise.
1830 (vbicq_u16): Likewise.
1831 (vbicq_u32): Likewise.
1832 (vbicq_u64): Likewise.
1833 (vorn_s8): Likewise.
1834 (vorn_s16): Likewise.
1835 (vorn_s32): Likewise.
1836 (vorn_u8): Likewise.
1837 (vorn_u16): Likewise.
1838 (vorn_u32): Likewise.
1839 (vorn_s64): Likewise.
1840 (vorn_u64): Likewise.
1841 (vornq_s8): Likewise.
1842 (vornq_s16): Likewise.
1843 (vornq_s32): Likewise.
1844 (vornq_s64): Likewise.
1845 (vornq_u8): Likewise.
1846 (vornq_u16): Likewise.
1847 (vornq_u32): Likewise.
1848 (vornq_u64): Likewise.
1850 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1852 * wide-int.cc (UTItype): Define.
1853 (UDWtype): Define for appropriate W_TYPE_SIZE.
1855 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1857 PR tree-optimization/59100
1858 * tree-ssa-phiopt.c: Include tree-inline.h.
1859 (neutral_element_p, absorbing_element_p): New functions.
1860 (value_replacement): Handle conditional binary operations with a
1861 neutral or absorbing element.
1863 2014-05-08 Richard Biener <rguenther@suse.de>
1865 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1866 folding the expression.
1867 (valueize_expr): Remove.
1868 (visit_reference_op_load): Do not valueize the result of
1870 (simplify_binary_expression): Likewise.
1871 (simplify_unary_expression): Likewise.
1873 2014-05-08 Richard Biener <rguenther@suse.de>
1875 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1876 looking at TYPE_ARG_TYPES.
1878 2014-05-08 Richard Biener <rguenther@suse.de>
1880 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1881 pointer propagation special-case.
1883 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1885 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1886 core part of address expressions.
1888 2014-05-08 Alan Modra <amodra@gmail.com>
1891 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1892 loads and stores when -mno-strict-align at any alignment.
1893 (expand_block_clear): Similarly. Also correct calculation of
1896 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1899 * tree-complex.c (expand_complex_move): Keep line info when expanding
1901 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1902 of complex expression. Use new argument to display correct location
1903 for values coming from phi statement.
1904 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1905 (warn_uninitialized_phi): Pass location of phi argument to
1907 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1908 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1910 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1912 * config/rs6000/predicates.md (indexed_address_mem): New.
1913 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1914 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1915 fpstore_ux, fpstore_u.
1916 (sign_extend, indexed, update): New.
1917 (cell_micro): Adjust.
1918 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1919 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1920 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1921 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1922 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1923 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1924 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1925 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1926 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1927 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1928 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1929 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1930 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1931 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1932 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1934 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1935 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1936 *vsx_extract_<mode>_store): Adjust.
1937 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1938 is_cracked_insn, insn_must_be_first_in_group,
1939 insn_must_be_last_in_group): Adjust.
1941 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1943 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1944 ppc440-fpstore): Adjust.
1945 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1946 ppc476-fpstore): Adjust.
1947 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1948 ppc601-fpstore): Adjust.
1949 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1951 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1953 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1954 ppc7450-fpstore): Adjust.
1955 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1956 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1957 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1959 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1960 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1961 cell-fpstore, cell-fpstore-update): Adjust.
1962 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1963 ppce300c3_store, ppce300c3_fpstore): Adjust.
1964 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1965 e500mc_fpstore): Adjust.
1966 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1967 e500mc64_store, e500mc64_fpstore): Adjust.
1968 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1969 e5500_fpstore): Adjust.
1970 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1971 e6500_fpstore): Adjust.
1972 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1974 * config/rs6000/power4.md (power4-load, power4-load-ext,
1975 power4-load-ext-update, power4-load-ext-update-indexed,
1976 power4-load-update-indexed, power4-load-update, power4-fpload,
1977 power4-fpload-update, power4-store, power4-store-update,
1978 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1980 * config/rs6000/power5.md (power5-load, power5-load-ext,
1981 power5-load-ext-update, power5-load-ext-update-indexed,
1982 power5-load-update-indexed, power5-load-update, power5-fpload,
1983 power5-fpload-update, power5-store, power5-store-update,
1984 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1986 * config/rs6000/power6.md (power6-load, power6-load-ext,
1987 power6-load-update, power6-load-update-indexed,
1988 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1989 power6-fpload-update, power6-store, power6-store-update,
1990 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1992 * config/rs6000/power7.md (power7-load, power7-load-ext,
1993 power7-load-update, power7-load-update-indexed,
1994 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1995 power7-fpload-update, power7-store, power7-store-update,
1996 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1998 * config/rs6000/power8.md (power8-load, power8-load-update,
1999 power8-load-ext, power8-load-ext-update, power8-fpload,
2000 power8-fpload-update, power8-store, power8-store-update-indexed,
2001 power8-fpstore, power8-fpstore-update): Adjust.
2002 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2004 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2005 titan_lsu_store, titan_lsu_fpstore): Adjust.
2006 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2008 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2011 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2012 unrolled byte insns. Emit address increments after move insns.
2014 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2016 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2017 const_gimple, rather than a gimple.
2018 (gimple_call_builtin_p): Likewise, for the three variants.
2020 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2021 (gimple_call_builtin_p): Likewise, for the three variants.
2023 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2025 PR tree-optimization/61095
2026 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2028 2014-05-07 Richard Biener <rguenther@suse.de>
2030 PR tree-optimization/61034
2031 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2032 (maybe_skip_until): Use translate to take into account
2033 lattices when trying to do disambiguations.
2034 (get_continuation_for_phi_1): Likewise.
2035 (get_continuation_for_phi): Adjust for added translate arguments.
2036 (walk_non_aliased_vuses): Likewise.
2037 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2038 (walk_non_aliased_vuses): Likewise.
2039 (call_may_clobber_ref_p_1): Declare.
2040 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2041 calls. Stop early if we are only supposed to disambiguate.
2042 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2044 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2046 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2047 Emit an error when the function has arguments.
2049 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2051 * cfgloop.h (unswitch_loops): Remove.
2052 * doc/passes.texi: Remove references to loop-unswitch.c
2053 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2055 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2057 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2058 check for loads group of length 3.
2059 (vect_permute_load_chain): New permutations for loads group of
2061 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2062 of vec_perm_shuffle for the new permutations.
2064 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2066 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2067 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2068 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2069 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2070 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2071 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2072 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2073 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2075 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2077 * loop-unswitch.c: Delete.
2079 2014-05-07 Richard Biener <rguenther@suse.de>
2081 * config.gcc: Always set need_64bit_hwint to yes.
2083 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2085 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2086 of using optimize_size.
2088 2014-05-06 Mike Stump <mikestump@comcast.net>
2090 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2092 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2094 * config/i386/sse.md (*mov<mode>_internal)
2095 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2096 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2097 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2098 (*<code><mode>3, *andnot<mode>3<mask_name>)
2099 (<mask_codefor><code><mode>3<mask_name>): Only consider
2100 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2102 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2105 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2107 * lra-constraints.c (valid_address_p): Move earlier in file.
2108 Add a constraint argument to the address_info version.
2109 (satisfies_memory_constraint_p): New function.
2110 (satisfies_address_constraint_p): Likewise.
2111 (process_alt_operands, curr_insn_transform): Use them.
2112 (process_address): Pass the constraint to valid_address_p when
2113 checking address operands.
2115 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2117 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2118 to their respective blocks. Fix inadvertent use of "node".
2120 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2122 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2124 (init_emit_once): ...here.
2125 * rtl.h (init_derived_machine_modes): Declare.
2126 * toplev.c (do_compile): Call it even if no_backend.
2128 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2129 Mike Stump <mikestump@comcast.net>
2130 Richard Sandiford <rdsandiford@googlemail.com>
2131 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2133 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2134 (rtx_equal_for_memref_p): Update comment.
2135 (adjust_offset_for_component_ref): Use wide-int interfaces.
2136 * builtins.c (get_object_alignment_2): Likewise.
2137 (c_readstr): Likewise.
2138 (target_char_cast): Add comment.
2139 (determine_block_size): Use wide-int interfaces.
2140 (expand_builtin_signbit): Likewise.
2141 (fold_builtin_int_roundingfn): Likewise.
2142 (fold_builtin_bitop): Likewise.
2143 (fold_builtin_bswap): Likewise.
2144 (fold_builtin_logarithm): Use signop.
2145 (fold_builtin_pow): Likewise.
2146 (fold_builtin_memory_op): Use wide-int interfaces.
2147 (fold_builtin_object_size): Likewise.
2148 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2149 nb_iterations_estimate.
2150 (record_niter_bound): Use wide-int interfaces.
2151 (get_estimated_loop_iterations_int): Likewise.
2152 (get_estimated_loop_iterations): Likewise.
2153 (get_max_loop_iterations): Likewise.
2154 * cfgloop.h: Include wide-int.h.
2155 (struct nb_iter_bound): Change bound to widest_int.
2156 (struct loop): Change nb_iterations_upper_bound and
2157 nb_iterations_estimate to widest_int.
2158 (record_niter_bound): Switch to use widest_int.
2159 (get_estimated_loop_iterations): Likewise.
2160 (get_max_loop_iterations): Likewise.
2161 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2162 update for wide-int.
2163 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2164 * combine.c (try_combine): Likewise.
2165 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2166 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2168 (aarch64_float_const_representable_p): Likewise.
2169 * config/arc/arc.c: Include wide-int.h.
2170 (arc_can_use_doloop_p): Use wide-int interfaces.
2171 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2172 (vfp3_const_double_index): Likewise.
2173 * config/avr/avr.c (avr_out_round): Likewise.
2174 (avr_fold_builtin): Likewise.
2175 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2176 (bfin_can_use_doloop_p): Likewise.
2177 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2178 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2179 * config/i386/i386.c: Include wide-int.h.
2180 (ix86_data_alignment): Use wide-int interfaces.
2181 (ix86_local_alignment): Likewise.
2182 (ix86_emit_swsqrtsf): Update real_from_integer.
2183 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2184 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2185 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2186 (zero_constant): Likewise.
2187 (input_operand): Likewise.
2188 (splat_input_operand): Likewise.
2189 (non_logical_cint_operand): Change const_double to const_wide_int.
2190 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2191 (easy_altivec_constant): Remove comment.
2192 (paired_expand_vector_init): Use CONSTANT_P.
2193 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2194 (rs6000_emit_move): Update checks.
2195 (rs6000_aggregate_candidate): Use wide-int interfaces.
2196 (rs6000_expand_ternop_builtin): Likewise.
2197 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2198 (rs6000_assemble_integer): Likewise.
2199 (rs6000_hash_constant): Likewise.
2200 (output_toc): Likewise.
2201 (rs6000_rtx_costs): Likewise.
2202 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2203 * config/rs6000/rs6000-c.c: Include wide-int.h.
2204 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2205 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2206 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2207 Handle CONST_WIDE_INT.
2208 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2209 Use tree_fits_uhwi_p.
2210 * config/sparc/sparc.c: Include wide-int.h.
2211 (sparc_fold_builtin): Use wide-int interfaces.
2212 * config/vax/vax.c: Include wide-int.h.
2213 (vax_float_literal): Use real_from_integer.
2214 * coretypes.h (struct hwivec_def): New.
2216 (const_hwivec): New.
2217 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2218 (equiv_constant): Handle CONST_WIDE_INT.
2219 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2220 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2221 * dbxout.c (stabstr_U): Use wide-int interfaces.
2222 (dbxout_type): Update to use cst_fits_shwi_p.
2223 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2224 (TARGET_SUPPORTS_WIDE_INT): Add default.
2225 * dfp.c: Include wide-int.h.
2226 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2227 decimal_real_to_integer.
2228 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2229 decimal_real_to_integer.
2230 * doc/generic.texi (Constant expressions): Update for wide_int.
2231 * doc/rtl.texi (const_double): Likewise.
2232 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2233 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2234 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2235 (REAL_VALUE_FROM_INT): Remove.
2236 (TARGET_SUPPORTS_WIDE_INT): New.
2237 * doc/tm.texi: Regenerate.
2238 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2239 * double-int.h: Include wide-int.h.
2240 (struct wi::int_traits): New.
2241 * dwarf2out.c (get_full_len): New.
2242 (dw_val_equal_p): Add case dw_val_class_wide_int.
2243 (size_of_loc_descr): Likewise.
2244 (output_loc_operands): Likewise.
2245 (insert_double): Remove.
2246 (insert_wide_int): New.
2248 (print_die): Add case dw_val_class_wide_int.
2249 (attr_checksum): Likewise.
2250 (attr_checksum_ordered): Likewise.
2251 (same_dw_val_p): Likewise.
2252 (size_of_die): Likewise.
2253 (value_format): Likewise.
2254 (output_die): Likewise.
2255 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2257 (clz_loc_descriptor): Use wide-int interfaces.
2258 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2259 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2260 (round_up_to_align): Use wide-int interfaces.
2261 (field_byte_offset): Likewise.
2262 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2263 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2264 CONST_DOUBLE handling. Use wide-int interfaces.
2265 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2266 (gen_enumeration_type_die): Use add_AT_wide.
2267 (hash_loc_operands): Add case dw_val_class_wide_int.
2268 (compare_loc_operands): Likewise.
2269 * dwarf2out.h: Include wide-int.h.
2270 (wide_int_ptr): New.
2271 (enum dw_val_class): Add dw_val_class_wide_int.
2272 (struct dw_val_struct): Add val_wide.
2273 * emit-rtl.c (const_wide_int_htab): New.
2274 (const_wide_int_htab_hash): New.
2275 (const_wide_int_htab_eq): New.
2276 (lookup_const_wide_int): New.
2277 (const_double_htab_hash): Use wide-int interfaces.
2278 (const_double_htab_eq): Likewise.
2279 (rtx_to_double_int): Conditionally compile for wide-int.
2280 (immed_double_int_const): Rename to immed_wide_int_const and
2281 update for wide-int.
2282 (immed_double_const): Conditionally compile for wide-int.
2283 (init_emit_once): Use wide-int interfaces.
2284 * explow.c (plus_constant): Likewise.
2285 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2286 (lshift_value): Use wide-int interfaces.
2287 (expand_mult): Likewise.
2288 (choose_multiplier): Likewise.
2289 (expand_smod_pow2): Likewise.
2290 (make_tree): Likewise.
2291 * expr.c (convert_modes): Consolidate handling of constants.
2292 Use wide-int interfaces.
2293 (emit_group_load_1): Add note.
2294 (store_expr): Update comment.
2295 (get_inner_reference): Use wide-int interfaces.
2296 (expand_constructor): Update comment.
2297 (expand_expr_real_2): Use wide-int interfaces.
2298 (expand_expr_real_1): Likewise.
2299 (reduce_to_bit_field_precision): Likewise.
2300 (const_vector_from_tree): Likewise.
2301 * final.c: Include wide-int-print.h.
2302 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2303 * fixed-value.c: Include wide-int.h.
2304 (fixed_from_string): Use wide-int interfaces.
2305 (fixed_to_decimal): Likewise.
2306 (fixed_convert_from_real): Likewise.
2307 (real_convert_from_fixed): Likewise.
2308 * fold-const.h (mem_ref_offset): Return an offset_int.
2309 (div_if_zero_remainder): Remove code parameter.
2310 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2311 Use wide-int interfaces.
2312 (may_negate_without_overflow_p): Use wide-int interfaces.
2313 (negate_expr_p): Likewise.
2314 (fold_negate_expr): Likewise.
2315 (int_const_binop_1): Likewise.
2316 (const_binop): Likewise.
2317 (fold_convert_const_int_from_int): Likewise.
2318 (fold_convert_const_int_from_real): Likewise.
2319 (fold_convert_const_int_from_fixed): Likewise.
2320 (fold_convert_const_fixed_from_int): Likewise.
2321 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2322 (sign_bit_p): Use wide-int interfaces.
2323 (make_range_step): Likewise.
2324 (build_range_check): Likewise. Pass an integer of the correct type
2325 instead of using integer_one_node.
2326 (range_predecessor): Pass an integer of the correct type instead
2327 of using integer_one_node.
2328 (range_successor): Likewise.
2329 (merge_ranges): Likewise.
2330 (unextend): Use wide-int interfaces.
2331 (extract_muldiv_1): Likewise.
2332 (fold_div_compare): Likewise.
2333 (fold_single_bit_test): Likewise.
2334 (fold_sign_changed_comparison): Likewise.
2335 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2336 (fold_plusminus_mult_expr): Use wide-int interfaces.
2337 (native_encode_int): Likewise.
2338 (native_interpret_int): Likewise.
2339 (fold_unary_loc): Likewise.
2340 (pointer_may_wrap_p): Likewise.
2341 (size_low_cst): Likewise.
2342 (mask_with_tz): Likewise.
2343 (fold_binary_loc): Likewise.
2344 (fold_ternary_loc): Likewise.
2345 (multiple_of_p): Likewise.
2346 (tree_call_nonnegative_warnv_p): Update calls to
2347 tree_int_cst_min_precision and real_from_integer.
2348 (fold_negate_const): Use wide-int interfaces.
2349 (fold_abs_const): Likewise.
2350 (fold_relational_const): Use tree_int_cst_lt.
2351 (round_up_loc): Use wide-int interfaces.
2352 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2353 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2354 * gengtype.c: Remove include of double-int.h.
2355 (do_typedef): Use wide-int interfaces.
2356 (open_base_files): Add wide-int.h.
2357 (main): Add offset_int and widest_int typedefs.
2358 * gengtype-lex.l: Handle "^".
2359 (CXX_KEYWORD): Add "static".
2360 * gengtype-parse.c (require3): New.
2361 (require_template_declaration): Handle constant template arguments
2362 and nested templates.
2363 * gengtype-state.c: Don't include "double-int.h".
2364 * genpreds.c (write_one_predicate_function): Update comment.
2365 (write_tm_constrs_h): Add check for hval and lval use in
2367 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2368 (add_to_sequence): Likewise.
2369 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2370 and const_double_operand.
2371 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2373 * gimple-fold.c (get_base_constructor): Likewise.
2374 (fold_array_ctor_reference): Likewise.
2375 (fold_nonarray_ctor_reference): Likewise.
2376 (fold_const_aggregate_ref_1): Likewise.
2377 (gimple_val_nonnegative_real_p): Likewise.
2378 (gimple_fold_indirect_ref): Likewise.
2379 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2380 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2381 (struct slsr_cand_d): Change index to be widest_int.
2382 (struct incr_info_d): Change incr to be widest_int.
2383 (alloc_cand_and_find_basis): Use wide-int interfaces.
2384 (slsr_process_phi): Likewise.
2385 (backtrace_base_for_ref): Likewise. Return a widest_int.
2386 (restructure_reference): Take a widest_int instead of a double_int.
2387 (slsr_process_ref): Use wide-int interfaces.
2388 (create_mul_ssa_cand): Likewise.
2389 (create_mul_imm_cand): Likewise.
2390 (create_add_ssa_cand): Likewise.
2391 (create_add_imm_cand): Take a widest_int instead of a double_int.
2392 (slsr_process_add): Use wide-int interfaces.
2393 (slsr_process_cast): Likewise.
2394 (slsr_process_copy): Likewise.
2395 (dump_candidate): Likewise.
2396 (dump_incr_vec): Likewise.
2397 (replace_ref): Likewise.
2398 (cand_increment): Likewise. Return a widest_int.
2399 (cand_abs_increment): Likewise.
2400 (replace_mult_candidate): Take a widest_int instead of a double_int.
2401 (replace_unconditional_candidate): Use wide-int interfaces.
2402 (incr_vec_index): Take a widest_int instead of a double_int.
2403 (create_add_on_incoming_edge): Likewise.
2404 (create_phi_basis): Use wide-int interfaces.
2405 (replace_conditional_candidate): Likewise.
2406 (record_increment): Take a widest_int instead of a double_int.
2407 (record_phi_increments): Use wide-int interfaces.
2408 (phi_incr_cost): Take a widest_int instead of a double_int.
2409 (lowest_cost_path): Likewise.
2410 (total_savings): Likewise.
2411 (analyze_increments): Use wide-int interfaces.
2412 (ncd_with_phi): Take a widest_int instead of a double_int.
2413 (ncd_of_cand_and_phis): Likewise.
2414 (nearest_common_dominator_for_cands): Likewise.
2415 (insert_initializers): Use wide-int interfaces.
2416 (all_phi_incrs_profitable): Likewise.
2417 (replace_one_candidate): Likewise.
2418 (replace_profitable_candidates): Likewise.
2419 * godump.c: Include wide-int-print.h.
2420 (go_output_typedef): Use wide-int interfaces.
2421 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2422 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2423 (build_loop_iteration_domains): Likewise.
2424 * hooks.h: Include wide-int.h rather than double-int.h.
2425 (hook_bool_dint_dint_uint_bool_true): Delete.
2426 (hook_bool_wint_wint_uint_bool_true): Declare.
2427 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2428 (hook_bool_wint_wint_uint_bool_true): New.
2429 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2431 (ubsan_expand_si_overflow_mul_check): Likewise.
2432 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2433 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2434 (get_ancestor_addr_info): Likewise.
2435 (ipa_modify_call_arguments): Likewise.
2436 * loop-doloop.c (doloop_modify): Likewise.
2437 (doloop_optimize): Likewise.
2438 * loop-iv.c (iv_number_of_iterations): Likewise.
2439 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2440 (unroll_loop_constant_iterations): Likewise.
2441 (decide_unroll_runtime_iterations): Likewise.
2442 (unroll_loop_runtime_iterations): Likewise.
2443 (decide_peel_simple): Likewise.
2444 (decide_unroll_stupid): Likewise.
2445 * lto-streamer-in.c (streamer_read_wi): Add.
2446 (input_cfg): Use wide-int interfaces.
2447 (lto_input_tree_1): Likewise.
2448 * lto-streamer-out.c (streamer_write_wi): Add.
2449 (hash_tree): Use wide-int interfaces.
2450 (output_cfg): Likewise.
2451 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2452 (GTFILES): Add wide-int.h and signop.h.
2453 (TAGS): Look for .cc files too.
2454 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2455 * optabs.c (expand_subword_shift): Likewise.
2456 (expand_doubleword_shift): Likewise.
2457 (expand_absneg_bit): Likewise.
2458 (expand_copysign_absneg): Likewise.
2459 (expand_copysign_bit): Likewise.
2460 * postreload.c (reload_cse_simplify_set): Likewise.
2461 * predict.c (predict_iv_comparison): Likewise.
2462 * pretty-print.h: Include wide-int-print.h.
2464 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2465 * print-tree.c: Include wide-int-print.h.
2466 (print_node_brief): Use wide-int interfaces.
2467 (print_node): Likewise.
2468 * read-rtl.c (validate_const_wide_int): New.
2469 (read_rtx_code): Add CONST_WIDE_INT case.
2470 * real.c: Include wide-int.h.
2471 (real_to_integer2): Delete.
2472 (real_to_integer): New function, returning a wide_int.
2473 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2474 (ten_to_ptwo): Update call to real_from_integer.
2475 (real_digit): Likewise.
2476 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2477 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2478 (REAL_VALUE_TO_INT): Delete.
2479 (real_to_integer): Declare a wide-int form.
2480 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2481 * recog.c (const_int_operand): Improve comment.
2482 (const_scalar_int_operand): New.
2483 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2484 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2485 (split_double): Likewise.
2486 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2487 (rtx_size): Likewise.
2488 (rtx_alloc_stat_v): New.
2489 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2490 (cwi_output_hex): New.
2491 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2492 (cwi_check_failed_bounds): New.
2493 * rtl.def (CONST_WIDE_INT): New.
2494 * rtl.h: Include <utility> and wide-int.h.
2495 (struct hwivec_def): New.
2496 (CWI_GET_NUM_ELEM): New.
2497 (CWI_PUT_NUM_ELEM): New.
2498 (struct rtx_def): Add num_elem and hwiv.
2499 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2500 (CASE_CONST_UNIQUE): Likewise.
2501 (CASE_CONST_ANY): Likewise.
2502 (CONST_SCALAR_INT_P): Likewise.
2503 (CONST_WIDE_INT_P): New.
2505 (HWIVEC_CHECK): New.
2506 (cwi_check_failed_bounds): New.
2508 (HWIVEC_CHECK): New.
2509 (CONST_WIDE_INT_VEC) New.
2510 (CONST_WIDE_INT_NUNITS) New.
2511 (CONST_WIDE_INT_ELT) New.
2512 (rtx_mode_t): New type.
2513 (wi::int_traits <rtx_mode_t>): New.
2515 (wi::min_value): New.
2516 (wi::max_value): New.
2518 (const_wide_int_alloc): New.
2519 (immed_wide_int_const): New.
2520 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2521 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2522 * signop.h: New file.
2523 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2524 (simplify_const_unary_operation): Use wide-int interfaces.
2525 (simplify_binary_operation_1): Likewise.
2526 (simplify_const_binary_operation): Likewise.
2527 (simplify_const_relational_operation): Likewise.
2528 (simplify_immed_subreg): Likewise.
2529 * stmt.c (expand_case): Likewise.
2530 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2531 signop rather than a bool.
2532 * stor-layout.c (layout_type): Use wide-int interfaces.
2533 (initialize_sizetypes): Update calls to
2534 set_min_and_max_values_for_integral_type.
2535 (set_min_and_max_values_for_integral_type): Take a signop rather
2536 than a bool. Use wide-int interfaces.
2537 (fixup_signed_type): Update accordingly. Remove
2538 HOST_BITS_PER_DOUBLE_INT limit.
2539 (fixup_unsigned_type): Likewise.
2540 * system.h (STATIC_CONSTANT_P): New.
2541 (STATIC_ASSERT): New.
2542 * target.def (can_use_doloop_p): Take widest_ints rather than
2544 * target.h: Include wide-int.h rather than double-int.h.
2545 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2547 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2548 rather than INT_CST_LT_UNSIGNED.
2549 (can_use_doloop_if_innermost): Take widest_ints rather than
2551 * tree-affine.c: Include wide-int-print.h.
2552 (double_int_ext_for_comb): Delete.
2553 (wide_int_ext_for_comb): New.
2554 (aff_combination_zero): Use wide-int interfaces.
2555 (aff_combination_const): Take a widest_int instead of a double_int.
2556 (aff_combination_elt): Use wide-int interfaces.
2557 (aff_combination_scale): Take a widest_int instead of a double_int.
2558 (aff_combination_add_elt): Likewise.
2559 (aff_combination_add_cst): Likewise.
2560 (aff_combination_add): Use wide-int interfaces.
2561 (aff_combination_convert): Likewise.
2562 (tree_to_aff_combination): Likewise.
2563 (add_elt_to_tree): Take a widest_int instead of a double_int.
2564 (aff_combination_to_tree): Use wide-int interfaces.
2565 (aff_combination_remove_elt): Likewise.
2566 (aff_combination_add_product): Take a widest_int instead of
2568 (aff_combination_mult): Use wide-int interfaces.
2569 (aff_combination_expand): Likewise.
2570 (double_int_constant_multiple_p): Delete.
2571 (wide_int_constant_multiple_p): New.
2572 (aff_combination_constant_multiple_p): Take a widest_int pointer
2573 instead of a double_int pointer.
2574 (print_aff): Use wide-int interfaces.
2575 (get_inner_reference_aff): Take a widest_int pointer
2576 instead of a double_int pointer.
2577 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2578 * tree-affine.h: Include wide-int.h.
2579 (struct aff_comb_elt): Change type of coef to widest_int.
2580 (struct affine_tree_combination): Change type of offset to widest_int.
2581 (double_int_ext_for_comb): Delete.
2582 (wide_int_ext_for_comb): New.
2583 (aff_combination_const): Use widest_int instead of double_int.
2584 (aff_combination_scale): Likewise.
2585 (aff_combination_add_elt): Likewise.
2586 (aff_combination_constant_multiple_p): Likewise.
2587 (get_inner_reference_aff): Likewise.
2588 (aff_comb_cannot_overlap_p): Likewise.
2589 (aff_combination_zero_p): Use wide-int interfaces.
2590 * tree.c: Include tree.h.
2591 (init_ttree): Use make_int_cst.
2592 (tree_code_size): Removed code for INTEGER_CST case.
2593 (tree_size): Add INTEGER_CST case.
2594 (make_node_stat): Update comment.
2595 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2596 (build_int_cst_type): Use wide-int interfaces.
2597 (double_int_to_tree): Likewise.
2598 (double_int_fits_to_tree_p): Delete.
2599 (force_fit_type_double): Delete.
2600 (force_fit_type): New.
2601 (int_cst_hash_hash): Use wide-int interfaces.
2602 (int_cst_hash_eq): Likewise.
2603 (build_int_cst_wide): Delete.
2604 (wide_int_to_tree): New.
2605 (cache_integer_cst): Use wide-int interfaces.
2606 (build_low_bits_mask): Likewise.
2607 (cst_and_fits_in_hwi): Likewise.
2608 (real_value_from_int_cst): Likewise.
2609 (make_int_cst_stat): New.
2610 (integer_zerop): Use wide_int interfaces.
2611 (integer_onep): Likewise.
2612 (integer_all_onesp): Likewise.
2613 (integer_pow2p): Likewise.
2614 (integer_nonzerop): Likewise.
2615 (tree_log2): Likewise.
2616 (tree_floor_log2): Likewise.
2617 (tree_ctz): Likewise.
2618 (int_size_in_bytes): Likewise.
2619 (mem_ref_offset): Return an offset_int rather than a double_int.
2620 (build_type_attribute_qual_variant): Use wide_int interfaces.
2621 (type_hash_eq): Likewise
2622 (tree_int_cst_equal): Likewise.
2623 (tree_int_cst_lt): Delete.
2624 (tree_int_cst_compare): Likewise.
2625 (tree_fits_shwi_p): Use wide_int interfaces.
2626 (tree_fits_uhwi_p): Likewise.
2627 (tree_int_cst_sign_bit): Likewise.
2628 (tree_int_cst_sgn): Likewise.
2629 (tree_int_cst_min_precision): Take a signop rather than a bool.
2630 (simple_cst_equal): Use wide_int interfaces.
2631 (compare_tree_int): Likewise.
2632 (iterative_hash_expr): Likewise.
2633 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2635 (get_type_static_bounds): Use wide_int interfaces.
2636 (tree_int_cst_elt_check_failed): New.
2637 (build_common_tree_nodes): Reordered to set prec before filling in
2639 (int_cst_value): Check cst_and_fits_in_hwi.
2640 (widest_int_cst_value): Use wide_int interfaces.
2641 (upper_bound_in_type): Likewise.
2642 (lower_bound_in_type): Likewise.
2643 (num_ending_zeros): Likewise.
2644 (drop_tree_overflow): Likewise.
2645 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2646 (gen_conditions_for_pow_cst_base): Likewise.
2647 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2648 (group_case_labels_stmt): Use wide-int interfaces.
2649 (verify_gimple_assign_binary): Likewise.
2650 (print_loop): Likewise.
2651 * tree-chrec.c (tree_fold_binomial): Likewise.
2652 * tree-core.h (struct tree_base): Add int_length.
2653 (struct tree_int_cst): Change rep of value.
2654 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2655 (dr_may_alias_p): Likewise.
2656 (max_stmt_executions_tree): Likewise.
2657 * tree.def (INTEGER_CST): Update comment.
2658 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2659 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2660 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2661 (dequeue_and_dump): Use wide-int interfaces.
2662 * tree.h: Include wide-int.h.
2663 (NULL_TREE): Moved to earlier loc in file.
2664 (TREE_INT_CST_ELT_CHECK): New.
2665 (tree_int_cst_elt_check_failed): New.
2667 (TREE_INT_CST): Delete.
2668 (TREE_INT_CST_LOW): Use wide-int interfaces.
2669 (TREE_INT_CST_HIGH): Delete.
2670 (TREE_INT_CST_NUNITS): New.
2671 (TREE_INT_CST_EXT_NUNITS): Likewise.
2672 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2673 (TREE_INT_CST_ELT): Likewise.
2674 (INT_CST_LT): Delete.
2675 (tree_int_cst_elt_check): New (two forms).
2676 (type_code_size): Update comment.
2677 (make_int_cst_stat, make_int_cst): New.
2678 (tree_to_double_int): Delete.
2679 (double_int_fits_to_tree_p): Delete.
2680 (force_fit_type_double): Delete.
2681 (build_int_cstu): Replace with out-of-line function.
2682 (build_int_cst_wide): Delete.
2683 (tree_int_cst_lt): Define inline.
2684 (tree_int_cst_le): New.
2685 (tree_int_cst_compare): Define inline.
2686 (tree_int_cst_min_precision): Take a signop rather than a bool.
2687 (wi::int_traits <const_tree>): New.
2688 (wi::int_traits <tree>): New.
2689 (wi::extended_tree): New.
2690 (wi::int_traits <wi::extended_tree>): New.
2691 (wi::to_widest): New.
2692 (wi::to_offset): New.
2693 (wi::fits_to_tree_p): New.
2694 (wi::min_value): New.
2695 (wi::max_value): New.
2696 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2697 (copy_tree_body_r): Likewise.
2698 * tree-object-size.c (compute_object_offset): Likewise.
2699 (addr_object_size): Likewise.
2700 * tree-predcom.c: Include wide-int-print.h.
2701 (struct dref_d): Change type of offset to widest_int.
2702 (dump_dref): Call wide-int printer.
2703 (aff_combination_dr_offset): Use wide-int interfaces.
2704 (determine_offset): Take a widest_int pointer rather than a
2706 (split_data_refs_to_components): Use wide-int interfaces.
2707 (suitable_component_p): Likewise.
2708 (order_drefs): Likewise.
2709 (add_ref_to_chain): Likewise.
2710 (valid_initializer_p): Likewise.
2711 (determine_roots_comp): Likewise.
2712 * tree-pretty-print.c: Include wide-int-print.h.
2713 (dump_generic_node): Use wide-int interfaces.
2714 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2715 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2716 (move_fixed_address_to_symbol): Likewise.
2717 (move_hint_to_base): Likewise.
2718 (move_pointer_to_base): Likewise.
2719 (move_variant_to_index): Likewise.
2720 (most_expensive_mult_to_index): Likewise.
2721 (addr_to_parts): Likewise.
2722 (copy_ref_info): Likewise.
2723 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2724 (indirect_refs_may_alias_p): Likewise.
2725 (stmt_kills_ref_p_1): Likewise.
2726 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2727 * tree-ssa-ccp.c: Update comment at top of file. Include
2729 (struct prop_value_d): Change type of mask to widest_int.
2730 (extend_mask): New function.
2731 (dump_lattice_value): Use wide-int interfaces.
2732 (get_default_value): Likewise.
2733 (set_constant_value): Likewise.
2734 (set_value_varying): Likewise.
2735 (valid_lattice_transition): Likewise.
2736 (set_lattice_value): Likewise.
2737 (value_to_double_int): Delete.
2738 (value_to_wide_int): New.
2739 (get_value_from_alignment): Use wide-int interfaces.
2740 (get_value_for_expr): Likewise.
2741 (do_dbg_cnt): Likewise.
2742 (ccp_finalize): Likewise.
2743 (ccp_lattice_meet): Likewise.
2744 (bit_value_unop_1): Use widest_ints rather than double_ints.
2745 (bit_value_binop_1): Likewise.
2746 (bit_value_unop): Use wide-int interfaces.
2747 (bit_value_binop): Likewise.
2748 (bit_value_assume_aligned): Likewise.
2749 (evaluate_stmt): Likewise.
2750 (ccp_fold_stmt): Likewise.
2751 (visit_cond_stmt): Likewise.
2752 (ccp_visit_stmt): Likewise.
2753 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2754 (constant_pointer_difference): Likewise.
2755 (associate_pointerplus): Likewise.
2756 (combine_conversions): Likewise.
2757 * tree-ssa-loop.h: Include wide-int.h.
2758 (struct tree_niter_desc): Change type of max to widest_int.
2759 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2760 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2761 (remove_redundant_iv_tests): Likewise.
2762 (canonicalize_loop_induction_variables): Likewise.
2763 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2764 (constant_multiple_of): Take a widest_int pointer instead of
2765 a double_int pointer.
2766 (get_computation_aff): Use wide-int interfaces.
2767 (ptr_difference_cost): Likewise.
2768 (difference_cost): Likewise.
2769 (get_loop_invariant_expr_id): Likewise.
2770 (get_computation_cost_at): Likewise.
2771 (iv_elimination_compare_lt): Likewise.
2772 (may_eliminate_iv): Likewise.
2773 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2774 instead of double_int.
2775 (max_loop_iterations): Likewise.
2776 (max_stmt_executions): Likewise.
2777 (estimated_stmt_executions): Likewise.
2778 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2779 (split_to_var_and_offset): Use wide-int interfaces.
2780 (determine_value_range): Likewise.
2781 (bound_difference_of_offsetted_base): Likewise.
2782 (bounds_add): Take a widest_int instead of a double_int.
2783 (number_of_iterations_ne_max): Use wide-int interfaces.
2784 (number_of_iterations_ne): Likewise.
2785 (number_of_iterations_lt_to_ne): Likewise.
2786 (assert_loop_rolls_lt): Likewise.
2787 (number_of_iterations_lt): Likewise.
2788 (number_of_iterations_le): Likewise.
2789 (number_of_iterations_cond): Likewise.
2790 (number_of_iterations_exit): Likewise.
2791 (finite_loop_p): Likewise.
2792 (derive_constant_upper_bound_assign): Likewise.
2793 (derive_constant_upper_bound): Return a widest_int.
2794 (derive_constant_upper_bound_ops): Likewise.
2795 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2796 (record_estimate): Take a widest_int rather than a double_int.
2797 (record_nonwrapping_iv): Use wide-int interfaces.
2798 (double_int_cmp): Delete.
2799 (wide_int_cmp): New.
2800 (bound_index): Take a widest_int rather than a double_int.
2801 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2802 (maybe_lower_iteration_bound): Likewise.
2803 (estimate_numbers_of_iterations_loop): Likewise.
2804 (estimated_loop_iterations): Take a widest_int pointer than than
2805 a double_int pointer.
2806 (estimated_loop_iterations_int): Use wide-int interfaces.
2807 (max_loop_iterations): Take a widest_int pointer than than
2808 a double_int pointer.
2809 (max_loop_iterations_int): Use wide-int interfaces.
2810 (max_stmt_executions): Take a widest_int pointer than than
2811 a double_int pointer.
2812 (estimated_stmt_executions): Likewise.
2813 (n_of_executions_at_most): Use wide-int interfaces.
2814 (scev_probably_wraps_p): Likewise.
2815 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2817 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2819 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2820 double_ints. Adjust for trailing_wide_ints <3> representation.
2821 (set_nonzero_bits): Likewise.
2822 (get_range_info): Return wide_ints rather than double_ints.
2823 Adjust for trailing_wide_ints <3> representation.
2824 (get_nonzero_bits): Likewise.
2825 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2827 * tree-ssanames.h (struct range_info_def): Replace min, max and
2828 nonzero_bits with a trailing_wide_ints <3>.
2829 (set_range_info): Use wide_int_refs rather than double_ints.
2830 (set_nonzero_bits): Likewise.
2831 (get_range_info): Return wide_ints rather than double_ints.
2832 (get_nonzero_bits): Likewise.
2833 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2834 * tree-ssa-pre.c (phi_translate_1): Likewise.
2835 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2836 (acceptable_pow_call): Likewise.
2837 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2839 (vn_reference_fold_indirect): Likewise.
2840 (vn_reference_maybe_forwprop_address): Likewise.
2841 (valueize_refs_1): Likewise.
2842 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2843 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2844 tree_int_cst_lt and tree_int_cst_le.
2845 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2847 (streamer_alloc_tree): Likewise.
2848 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2849 (streamer_write_tree_header): Likewise.
2850 (streamer_write_integer_cst): Likewise.
2851 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2852 (build_constructors): Likewise.
2853 (array_value_type): Likewise.
2854 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2855 (vect_check_gather): Likewise.
2856 * tree-vect-generic.c (build_replicated_const): Likewise.
2857 (expand_vector_divmod): Likewise.
2858 * tree-vect-loop.c (vect_transform_loop): Likewise.
2859 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2860 (vect_do_peeling_for_alignment): Likewise.
2861 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2862 * tree-vrp.c: Include wide-int.h.
2863 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2864 (extract_range_from_assert): Use wide-int interfaces.
2865 (vrp_int_const_binop): Likewise.
2866 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2867 double_int pointers.
2868 (ranges_from_anti_range): Use wide-int interfaces.
2869 (quad_int_cmp): Delete.
2870 (quad_int_pair_sort): Likewise.
2871 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2872 (extract_range_from_unary_expr_1): Likewise.
2873 (adjust_range_with_scev): Likewise.
2874 (masked_increment): Take and return wide_ints rather than double_ints.
2875 (register_edge_assert_for_2): Use wide-int interfaces.
2876 (check_array_ref): Likewise.
2877 (search_for_addr_array): Likewise.
2878 (maybe_set_nonzero_bits): Likewise.
2879 (union_ranges): Pass an integer of the correct type instead of
2880 using integer_one_node.
2881 (intersect_ranges): Likewise.
2882 (simplify_truth_ops_using_ranges): Likewise.
2883 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2884 (range_fits_type_p): Likewise.
2885 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2887 (simplify_conversion_using_ranges): Use wide-int interfaces.
2888 (simplify_float_conversion_using_ranges): Likewise.
2889 (vrp_finalize): Likewise.
2890 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2891 (gimple_stringops_transform): Likewise.
2892 * varasm.c (decode_addr_const): Likewise.
2893 (const_hash_1): Likewise.
2894 (const_rtx_hash_1): Likewise
2895 (output_constant): Likewise.
2896 (array_size_for_constructor): Likewise.
2897 (output_constructor_regular_field): Likewise.
2898 (output_constructor_bitfield): Likewise.
2899 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2900 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2902 * gencheck.c: Define BITS_PER_UNIT.
2905 * wide-int-print.cc: New.
2906 * wide-int-print.h: New.
2908 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2910 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2912 2014-05-06 Richard Biener <rguenther@suse.de>
2914 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2915 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2916 (TODO_verify_all): Adjust.
2917 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2918 TODO_verify_stmts and TODO_verify_rtl_sharing.
2919 * bb-reorder.c: Likewise.
2920 * cfgexpand.c: Likewise.
2921 * cprop.c: Likewise.
2923 * function.c: Likewise.
2924 * fwprop.c: Likewise.
2926 * gimple-ssa-isolate-paths.c: Likewise.
2927 * gimple-ssa-strength-reduction.c: Likewise.
2928 * ipa-split.c: Likewise.
2929 * loop-init.c: Likewise.
2930 * loop-unroll.c: Likewise.
2931 * lower-subreg.c: Likewise.
2932 * modulo-sched.c: Likewise.
2933 * postreload-gcse.c: Likewise.
2934 * predict.c: Likewise.
2935 * recog.c: Likewise.
2936 * sched-rgn.c: Likewise.
2937 * store-motion.c: Likewise.
2938 * tracer.c: Likewise.
2939 * trans-mem.c: Likewise.
2940 * tree-call-cdce.c: Likewise.
2941 * tree-cfg.c: Likewise.
2942 * tree-cfgcleanup.c: Likewise.
2943 * tree-complex.c: Likewise.
2944 * tree-eh.c: Likewise.
2945 * tree-emutls.c: Likewise.
2946 * tree-if-conv.c: Likewise.
2947 * tree-into-ssa.c: Likewise.
2948 * tree-loop-distribution.c: Likewise.
2949 * tree-object-size.c: Likewise.
2950 * tree-parloops.c: Likewise.
2951 * tree-pass.h: Likewise.
2952 * tree-sra.c: Likewise.
2953 * tree-ssa-ccp.c: Likewise.
2954 * tree-ssa-copy.c: Likewise.
2955 * tree-ssa-copyrename.c: Likewise.
2956 * tree-ssa-dce.c: Likewise.
2957 * tree-ssa-dom.c: Likewise.
2958 * tree-ssa-dse.c: Likewise.
2959 * tree-ssa-forwprop.c: Likewise.
2960 * tree-ssa-ifcombine.c: Likewise.
2961 * tree-ssa-loop-ch.c: Likewise.
2962 * tree-ssa-loop-ivcanon.c: Likewise.
2963 * tree-ssa-loop.c: Likewise.
2964 * tree-ssa-math-opts.c: Likewise.
2965 * tree-ssa-phiopt.c: Likewise.
2966 * tree-ssa-phiprop.c: Likewise.
2967 * tree-ssa-pre.c: Likewise.
2968 * tree-ssa-reassoc.c: Likewise.
2969 * tree-ssa-sink.c: Likewise.
2970 * tree-ssa-strlen.c: Likewise.
2971 * tree-ssa-tail-merge.c: Likewise.
2972 * tree-ssa-uncprop.c: Likewise.
2973 * tree-switch-conversion.c: Likewise.
2974 * tree-tailcall.c: Likewise.
2975 * tree-vect-generic.c: Likewise.
2976 * tree-vectorizer.c: Likewise.
2977 * tree-vrp.c: Likewise.
2979 * var-tracking.c: Likewise.
2980 * bt-load.c: Likewise.
2981 * cfgcleanup.c: Likewise.
2982 * combine-stack-adj.c: Likewise.
2983 * combine.c: Likewise.
2984 * compare-elim.c: Likewise.
2985 * config/epiphany/resolve-sw-modes.c: Likewise.
2986 * config/i386/i386.c: Likewise.
2987 * config/mips/mips.c: Likewise.
2988 * config/s390/s390.c: Likewise.
2989 * config/sh/sh_treg_combine.cc: Likewise.
2990 * config/sparc/sparc.c: Likewise.
2993 * final.c: Likewise.
2994 * ifcvt.c: Likewise.
2995 * mode-switching.c: Likewise.
2996 * passes.c: Likewise.
2997 * postreload.c: Likewise.
2999 * reg-stack.c: Likewise.
3000 * regcprop.c: Likewise.
3001 * regrename.c: Likewise.
3004 2014-05-06 Richard Biener <rguenther@suse.de>
3007 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3008 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3010 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3013 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3015 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3016 Tom de Vries <tom@codesourcery.com>
3018 * target.def (call_fusage_contains_non_callee_clobbers): New
3020 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3022 (@node Miscellaneous Register Hooks): New node.
3023 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3024 * doc/tm.texi: Regenerate.
3026 2014-05-05 Marek Polacek <polacek@redhat.com>
3029 * opts.c (common_handle_option): Call error_at instead of warning_at.
3031 2014-05-05 Richard Biener <rguenther@suse.de>
3033 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3034 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3035 under the TODO_verify_il umbrella.
3037 2014-05-05 Richard Biener <rguenther@suse.de>
3039 * passes.c (execute_function_todo): Move TODO_verify_flow under
3040 the TODO_verify_ul umbrella.
3042 2014-05-05 Richard Biener <rguenther@suse.de>
3045 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3046 X & CST away from a CST that is the mask of a mode.
3048 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3050 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3051 int argument to enum machine_mode.
3052 (picochip_class_max_nregs): Ditto.
3053 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3054 (picochip_class_max_nregs): Ditto.
3056 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3058 * target.def: Add new target hook.
3059 * doc/tm.texi: Regenerate.
3060 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3061 * targhooks.c (default_keep_leaf_when_profiled): New function.
3063 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3064 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3066 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3068 PR tree-optimization/60363
3069 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3070 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3071 (update_destination_phis): New parameter.
3072 (create_edge_and_update_destination_phis): Ditto.
3073 (ssa_fix_duplicate_block_edges): Pass new arguments.
3074 (thread_single_edge): Ditto.
3076 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3078 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3079 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3080 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3081 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3082 Use RS6000_BTM_HARD_FLOAT.
3083 (BU_MISC_2): Likewise.
3084 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3085 RS6000_BTM_HARD_FLOAT.
3086 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3088 (rs6000_invalid_builtin): Add hard floating builtin support.
3089 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3090 hard float builtins.
3091 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3093 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3095 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3096 Add missing function* argument.
3098 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3100 * lra-constraints.c (valid_address_p): Move earlier in file.
3101 Add a constraint argument to the address_info version.
3102 (satisfies_memory_constraint_p): New function.
3103 (satisfies_address_constraint_p): Likewise.
3104 (process_alt_operands, curr_insn_transform): Use them.
3105 (process_address): Pass the constraint to valid_address_p when
3106 checking address operands.
3108 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3110 * config/mips/mips.c (mips_isa_rev): New variable.
3111 (mips_set_architecture): Set it.
3112 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3114 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3115 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3116 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3117 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3118 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3119 conditions in terms of mips_isa_rev.
3120 (mips_isa_rev): Declare.
3122 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3124 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3125 (prob_unlikely, prob_likely): Make variables const.
3127 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3129 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3131 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3133 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3135 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3137 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3138 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3139 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3141 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3142 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3144 Replace usage of ROUND_REG with sh_round_reg.
3145 Use CEIL instead of ROUND_ADVANCE.
3147 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3150 * config/sh/sh.c: Include stdlib headers before everything else.
3152 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3154 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3155 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3156 (gimplify_adjust_omp_clauses): Simd region is never
3157 directly nested in combined parallel. Instead, for linear
3158 with copyin/copyout, if in combined for simd loop, make decl
3159 firstprivate/lastprivate on OMP_FOR.
3160 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3161 expand_omp_for_static_chunk): When setting endvar, also set
3162 fd->loop.v to the same value.
3164 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3166 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3168 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3170 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3173 2014-05-02 Marek Polacek <polacek@redhat.com>
3175 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3177 2014-05-02 Kito Cheng <kito@0xlab.org>
3179 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3180 to a C expression marco.
3181 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3182 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3183 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3184 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3185 HONOR_REG_ALLOC_ORDER.
3186 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3188 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3190 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3192 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3194 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3196 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3198 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3199 (convert_scalar_cond_reduction): Likewise.
3200 (predicate_scalar_phi): Add recognition and transformation
3201 of simple conditioanl reduction to be vectorizable.
3203 2014-05-01 Marek Polacek <polacek@redhat.com>
3206 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3208 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3210 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3211 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3212 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3213 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3214 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3215 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3216 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3217 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3219 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3221 * config/arc/arc.opt (mlra): Move comment above option name
3222 to avoid mis-parsing as language options.
3224 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3226 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3227 * config/sol2.h: ... here.
3228 * config/sol2-10.h: Remove.
3230 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3231 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3232 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3233 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3234 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3235 * config/sol2.h: ... here.
3236 (SECTION_NAME_FORMAT): Don't redefine.
3237 (STARTFILE_ARCH32_SPEC): Rename to ...
3238 (STARTFILE_ARCH_SPEC): ... this.
3239 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3240 * config/sparc/sol2.h: ... here.
3241 (SECTION_NAME_FORMAT): Don't undef.
3242 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3243 (SUBTARGET_EXTRA_SPECS): Remove.
3244 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3246 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3247 (MD_STARTFILE_PREFIX): Remove.
3248 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3249 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3250 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3251 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3252 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3253 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3254 * config/i386/sol2.h: ... here.
3255 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3256 * config/i386/sol2-bi.h: Remove.
3257 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3258 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3260 * config/i386/t-sol2-64: Rename to ...
3261 * config/i386/t-sol2: ... this.
3262 * config/sparc/t-sol2-64: Rename to ...
3263 * config/sparc/t-sol2: ... this.
3265 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3266 sol2_tm_file_head, sol2_tm_file_tail.
3267 Include ${cpu_type}/sol2.h before sol2.h.
3269 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3270 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3271 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3272 Reflect i386/t-sol2-64 renaming.
3273 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3274 Reflect sparc/t-sol2-64 renaming.
3276 2014-04-30 Richard Biener <rguenther@suse.de>
3278 * passes.c (execute_function_todo): Move TODO_verify_stmts
3279 and TODO_verify_ssa under the TODO_verify_il umbrella.
3280 * tree-ssa.h (verify_ssa): Adjust prototype.
3281 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3282 we should verify SSA operands.
3283 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3284 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3285 whether we should verify whether not throwing stmts have EH info.
3286 * graphite-scop-detection.c (create_sese_edges): Adjust.
3287 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3288 * tree-eh.c (lower_try_finally_switch): Do not add the
3289 default case label twice.
3291 2014-04-30 Marek Polacek <polacek@redhat.com>
3293 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3294 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3295 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3296 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3298 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3300 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3301 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3302 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3303 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3304 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3305 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3306 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3307 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3309 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3311 * tree-cfg.c (dump_function_to_file): Dump the return type of
3312 functions, in a line to itself before the function body, mimicking
3313 the layout of a C function.
3315 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3317 PR tree-optimization/60971
3318 * tree-tailcall.c (process_assignment): Reject conversions which
3321 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3323 * calls.c (initialize_argument_information): Always treat
3324 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3325 (expand_call): Likewise.
3326 (emit_library_call_calue_1): Likewise.
3327 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3328 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3331 2014-04-29 Nick Clifton <nickc@redhat.com>
3333 * config/msp430/msp430.md (umulsidi): Fix typo.
3334 (mulhisi3): Enable even inside interrupt handlers.
3335 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3336 bigger return address pushed in large mode.
3338 2014-04-29 Nick Clifton <nickc@redhat.com>
3340 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3341 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3343 * config/m32r/m32r.c (init_reg_tables): Likewise.
3344 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3345 enum to hold the modes.
3347 2014-04-29 Richard Biener <rguenther@suse.de>
3349 * dominance.c (free_dominance_info): Add overload with
3351 (dom_info_state): Likewise.
3352 (dom_info_available_p): Likewise.
3353 * basic-block.h (free_dominance_info, dom_info_state,
3354 dom_info_available_p): Declare overloads.
3355 * passes.c (execute_function_todo): Verify that verifiers
3356 don't change dominator info state. Drop dominator info
3357 for IPA pass invocations.
3358 * cgraph.c (release_function_body): Restore asserts that
3359 dominator information is released.
3361 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3363 * doc/invoke.texi: Fix typo.
3364 * tree-vrp.c: Fix typos.
3365 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3367 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3369 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3371 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3373 * config/aarch64/aarch64-builtins.c
3374 (aarch64_types_storestruct_lane_qualifiers): New.
3375 (TYPES_STORESTRUCT_LANE): Likewise.
3376 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3377 (st3_lane): Likewise.
3378 (st4_lane): Likewise.
3379 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3380 (vec_store_lanesci_lane<mode>): Likewise.
3381 (vec_store_lanesxi_lane<mode>): Likewise.
3382 (aarch64_st2_lane<VQ:mode>): Likewise.
3383 (aarch64_st3_lane<VQ:mode>): Likewise.
3384 (aarch64_st4_lane<VQ:mode>): Likewise.
3385 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3386 * config/aarch64/arm_neon.h
3387 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3388 use new macro arguments.
3389 (__ST3_LANE_FUNC): Likewise.
3390 (__ST4_LANE_FUNC): Likewise.
3391 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3392 (V_THREE_ELEM): Likewise.
3393 (V_FOUR_ELEM): Likewise.
3395 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3397 * doc/gimple.texi: Replace the description of the now-defunct
3398 union gimple_statement_d with a diagram showing the
3399 gimple_statement_base class hierarchy and its relationships to
3400 the GSS_ and GIMPLE_ enums.
3402 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3404 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3405 * config/aarch64/aarch64.c
3406 (aarch64_cannot_change_mode_class): Weaken conditions.
3407 (aarch64_modes_tieable_p): New.
3408 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3410 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3412 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3413 (loadsync_<mode>): Change mode.
3414 (load_quadpti, store_quadpti): New.
3415 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3416 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3418 2014-04-28 Martin Jambor <mjambor@suse.cz>
3420 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3421 same alias type as the original statement.
3422 (subreplacement_assignment_data): New type.
3423 (handle_unscalarized_data_in_subtree): New type of parameter,
3424 generate new memory accesses with same alias type as the original
3426 (load_assign_lhs_subreplacements): Likewise.
3427 (sra_modify_constructor_assign): Generate new memory accesses with
3428 same alias type as the original statement.
3430 2014-04-28 Richard Biener <rguenther@suse.de>
3432 * tree-pass.h (TODO_verify_il): Define.
3433 (TODO_verify_all): Complete properly.
3434 * passes.c (execute_function_todo): Move existing loop-closed
3435 SSA verification under TODO_verify_il.
3436 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3437 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3438 Fix tree sharing issue.
3440 2014-04-28 Richard Biener <rguenther@suse.de>
3443 * builtins.def (DEF_C11_BUILTIN): Add.
3444 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3445 * coretypes.h (enum function_class): Add function_c11_misc.
3446 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3447 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3448 (call_may_clobber_ref_p_1): Likewise.
3449 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3450 (mark_all_reaching_defs_necessary_1): Likewise.
3451 (propagate_necessity): Likewise.
3452 (eliminate_unnecessary_stmts): Likewise.
3453 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3455 2014-04-28 Richard Biener <rguenther@suse.de>
3457 * tree-vrp.c (vrp_var_may_overflow): Remove.
3458 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3459 with overflow immediately bump to one before that value and
3460 let iteration figure out overflow status.
3462 2014-04-28 Richard Biener <rguenther@suse.de>
3464 * configure.ac: Do valgrind header checks unconditionally.
3465 Add --enable-valgrind-annotations.
3466 * system.h: Guard valgrind header inclusion with
3467 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3468 * alloc-pool.c (pool_alloc, pool_free): Use
3469 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3470 to guard possibly dead code.
3471 * config.in: Regenerated.
3472 * configure: Likewise.
3474 2014-04-28 Jeff Law <law@redhat.com>
3476 PR tree-optimization/60902
3477 * tree-ssa-threadedge.c
3478 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3479 over real defs when invalidating outputs from statements that do not
3480 produce useful outputs for threading.
3482 2014-04-28 Richard Biener <rguenther@suse.de>
3484 PR tree-optimization/60979
3485 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3486 SCOPs that end in a block with a successor with abnormal
3489 2014-04-28 Richard Biener <rguenther@suse.de>
3491 * tree-pass.h (execute_pass_list): Adjust prototype.
3492 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3493 (do_per_function): Change callback signature, push all actual
3494 work to the callbals.
3495 (do_per_function_toporder): Likewise.
3496 (execute_function_dump): Adjust.
3497 (execute_function_todo): Likewise.
3498 (clear_last_verified): Likewise.
3499 (verify_curr_properties): Likewise.
3500 (update_properties_after_pass): Likewise.
3501 (execute_pass_list_1): Split out from ...
3502 (execute_pass_list): ... here. Adjust.
3503 (execute_ipa_pass_list): Likewise.
3504 * cgraphunit.c (cgraph_add_new_function): Adjust.
3505 (analyze_function): Likewise.
3506 (expand_function): Likewise.
3507 * cgraph.c (release_function_body): Free dominance info
3508 here instead of asserting it was magically freed elsewhere.
3510 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3512 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3513 * configure: Regenerate.
3514 * config/sparc/sparc.opt (muser-mode): New option.
3515 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3517 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3518 * doc/invoke.texi (SPARC options): Document -muser-mode.
3520 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3522 * cselib.c (find_slot_memmode): Delete.
3523 (cselib_hasher): Change compare_type to a struct.
3524 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3526 (preserve_constants_and_equivs): Adjust for new compare_type.
3527 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3528 (wrap_constant): Delete.
3529 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3531 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3533 * doc/install.texi (Building with profile feedback): Remove
3536 2014-04-26 Tom de Vries <tom@codesourcery.com>
3538 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3541 2014-04-25 Cary Coutant <ccoutant@google.com>
3544 * dwarf2out.c (should_move_die_to_comdat): A type definition
3545 can contain a subprogram definition, but don't move it to a
3547 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3548 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3550 (clone_tree_hash): Rename to...
3551 (clone_tree_partial): ...this; change callers. Copy
3552 DW_TAG_subprogram DIEs as declarations.
3553 (copy_decls_walk): Don't copy children of a declaration into a
3556 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3559 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3562 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3564 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3565 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3567 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3570 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3572 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3574 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3576 PR tree-optimization/60930
3577 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3578 creating a multiply candidate by folding two constant
3579 multiplicands when the result overflows.
3581 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3583 PR tree-optimization/60960
3584 * tree-vect-generic.c (expand_vector_operation): Only call
3585 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3587 2014-04-25 Tom de Vries <tom@codesourcery.com>
3589 * expr.c (clobber_reg_mode): New function.
3590 * expr.h (clobber_reg): New function.
3592 2014-04-25 Tom de Vries <tom@codesourcery.com>
3594 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3597 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3598 Tom de Vries <tom@codesourcery.com>
3600 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3602 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3603 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3604 new argument to find_all_hard_reg_sets call.
3606 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3608 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3609 Use HOST_WIDE_INT_C for mask literal.
3610 (aarch_rev16_shleft_mask_imm_p): Likewise.
3612 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3615 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3617 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3619 PR preprocessor/56540
3620 * config/i386/i386-c.c (ix86_target_macros): Define
3621 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3623 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3625 * configure.ac (tga_func): Remove.
3626 (LIB_TLS_SPEC): Remove.
3627 * configure: Regenerate.
3628 * config.in: Regenerate.
3629 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3631 2014-04-25 Richard Biener <rguenther@suse.de>
3634 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3635 call stmt use/clobber sets during stmt walk instead of
3636 walking the possibly incomplete set of caller edges.
3638 2014-04-25 Richard Biener <rguenther@suse.de>
3641 * passes.c (apply_ipa_transforms): Inline into only caller ...
3642 (execute_one_pass): ... here. Properly bring in function
3643 bodies for nodes we want to apply IPA transforms to.
3645 2014-04-24 Cong Hou <congh@google.com>
3647 PR tree-optimization/60896
3648 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3649 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3650 (vect_mark_pattern_stmts): Set the def type of all statements in
3651 PATTERN_DEF_SEQ as vect_internal_def.
3653 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3655 * doc/extend.texi (PowerPC Built-in Functions): Document new
3656 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3657 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3659 * config/rs6000/predicates.md (const_0_to_3_operand): New
3660 predicate to match 0..3 integer constants.
3662 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3663 to support adding miscellaneous builtin functions.
3664 (BU_DFP_MISC_2): Likewise.
3665 (BU_P7_MISC_1): Likewise.
3666 (BU_P7_MISC_2): Likewise.
3667 (BU_P8V_MISC_3): Likewise.
3668 (BU_MISC_1): Likewise.
3669 (BU_MISC_2): Likewise.
3670 (DIVWE): Add extended divide builtin functions.
3673 (DIVWEUO): Likewise.
3677 (DIVDEUO): Likewise.
3678 (DXEX): Add decimal floating-point builtin functions.
3681 (DDEDPDQ): Likewise.
3683 (DENBCDQ): Likewise.
3690 (CDTBCD): Add new BCD builtin functions.
3694 (BCDADD_LT): Likewise.
3695 (BCDADD_EQ): Likewise.
3696 (BCDADD_GT): Likewise.
3697 (BCDADD_OV): Likewise.
3699 (BCDSUB_LT): Likewise.
3700 (BCDSUB_EQ): Likewise.
3701 (BCDSUB_GT): Likewise.
3702 (BCDSUB_OV): Likewise.
3703 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3704 (UNPACK_TD): Likewise.
3705 (PACK_TF): Likewise.
3706 (UNPACK_TF): Likewise.
3707 (UNPACK_TF_0): Likewise.
3708 (UNPACK_TF_1): Likewise.
3709 (PACK_V1TI): Likewise.
3710 (UNPACK_V1TI): Likewise.
3712 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3713 support for decimal floating point builtin functions.
3714 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3715 functions that take constant arguments.
3716 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3717 (rs6000_init_builtins): Setup long double, _Decimal64, and
3718 _Decimal128 types for new builtin functions.
3719 (builtin_function_type): Set the unsigned flags appropriately for
3720 the new builtin functions.
3721 (rs6000_opt_masks): Add support for decimal floating point builtin
3724 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3725 floating point builtin functions.
3726 (RS6000_BTM_COMMON): Likewise.
3727 (RS6000_BTI_long_double): Likewise.
3728 (RS6000_BTI_dfloat64): Likewise.
3729 (RS6000_BTI_dfloat128): Likewise.
3730 (long_double_type_internal_node): Likewise.
3731 (dfloat64_type_internal_node): Likewise.
3732 (dfloat128_type_internal_node): Likewise.
3734 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3735 2.07 bcd arithmetic instructions.
3736 (UNSPEC_BCDSUB): Likewise.
3737 (UNSPEC_BCD_OVERFLOW): Likewise.
3738 (UNSPEC_BCD_ADD_SUB): Likewise.
3739 (bcd_add_sub): Likewise.
3740 (BCD_TEST): Likewise.
3741 (bcd<bcd_add_sub>): Likewise.
3742 (bcd<bcd_add_sub>_test): Likewise.
3743 (bcd<bcd_add_sub>_test2): Likewise.
3744 (bcd<bcd_add_sub>_<code>): Likewise.
3745 (peephole2 for combined bcd ops): Likewise.
3747 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3748 decimal floating point builtin functions.
3749 (UNSPEC_DENBCD): Likewise.
3750 (UNSPEC_DXEX): Likewise.
3751 (UNSPEC_DIEX): Likewise.
3752 (UNSPEC_DSCLI): Likewise.
3753 (UNSPEC_DSCRI): Likewise.
3754 (D64_D128): Likewise.
3755 (dfp_suffix): Likewise.
3756 (dfp_ddedpd_<mode>): Likewise.
3757 (dfp_denbcd_<mode>): Likewise.
3758 (dfp_dxex_<mode>): Likewise.
3759 (dfp_diex_<mode>): Likewise.
3760 (dfp_dscli_<mode>): Likewise.
3761 (dfp_dscri_<mode>): Likewise.
3763 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3765 (UNSPEC_CDTBCD): Likewise.
3766 (UNSPEC_CBCDTD): Likewise.
3767 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3768 (UNSPEC_DIVEO): Likewise.
3769 (UNSPEC_DIVEU): Likewise.
3770 (UNSPEC_DIVEUO): Likewise.
3771 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3772 pack/unpack 128-bit types.
3773 (UNSPEC_PACK_128BIT): Likewise.
3774 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3775 (udiv<mode>3): Use idiv_ldiv mode attribute.
3776 (div<mode>3): Likewise.
3777 (addg6s): Add new BCD builtin functions.
3780 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3781 (div_extend): Likewise.
3782 (div<div_extend>_<mode>"): Likewise.
3783 (FP128_64): Add support for new builtin functions to pack/unpack
3785 (unpack<mode>): Likewise.
3786 (unpacktf_0): Likewise.
3787 (unpacktf_1): Likewise.
3788 (unpack<mode>_dm): Likewise.
3789 (unpack<mode>_nodm): Likewise.
3790 (pack<mode>): Likewise.
3791 (unpackv1ti): Likewise.
3792 (packv1ti): Likewise.
3794 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3796 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3799 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3801 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3802 * gimplify.c (omp_is_private): Change last argument's type to int.
3803 Only diagnose lastprivate if the simd argument is 1, only diagnose
3804 linear if the simd argument is 2.
3805 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3806 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3807 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3808 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3809 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3810 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3811 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3812 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3813 * tree-nested.c (convert_nonlocal_omp_clauses,
3814 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3816 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3819 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3822 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3824 * flag-types.h (enum ivar_visibility): Add.
3826 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3828 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3829 function * argument.
3831 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3833 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3835 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3836 Tom de Vries <tom@codesourcery.com>
3838 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3839 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3841 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3842 * emit-rtl.c (try_split): Same.
3844 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3845 Tom de Vries <tom@codesourcery.com>
3847 * common.opt (fuse-caller-save): New option.
3849 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3851 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3852 elements for big-endian.
3854 2014-04-24 Richard Biener <rguenther@suse.de>
3856 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3857 during TER and instead use the sepops interface for expanding
3858 non-GIMPLE_SINGLE_RHS.
3860 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3862 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3863 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3865 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3867 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3868 assembler 64-bit option.
3869 * configure: Regenerate.
3871 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3873 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3874 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3875 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3876 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3877 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3879 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3881 * config/aarch64/aarch64-builtins.c
3882 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3883 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3884 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3885 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3887 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3888 (Vrevsuff): New mode attribute.
3890 2014-04-24 Terry Guo <terry.guo@arm.com>
3892 * config/arm/arm.h (machine_function): Define variable
3893 after_arm_reorg here.
3894 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3895 (arm_split_constant): Update the way to access variable
3898 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3900 2014-04-23 Tom de Vries <tom@codesourcery.com>
3902 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3904 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3906 * is-a.h: Update comments to reflect the following changes to the
3907 "pointerness" of the API, making the template parameter match the
3908 return type, allowing use of is-a.h with typedefs of pointers.
3909 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3910 that the return type matches the parameter to the is_a_helper.
3912 (dyn_cast): Likewise.
3914 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3915 pointer from the is-a.h API.
3917 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3918 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3920 (is_a_helper <varpool_node>::test): Likewise, convert to...
3921 (is_a_helper <varpool_node *>::test): ...this.
3923 (varpool_first_variable): Update for removal of implicit pointer
3924 from the is-a.h API.
3925 (varpool_next_variable): Likewise.
3926 (varpool_first_static_initializer): Likewise.
3927 (varpool_next_static_initializer): Likewise.
3928 (varpool_first_defined_variable): Likewise.
3929 (varpool_next_defined_variable): Likewise.
3930 (cgraph_first_defined_function): Likewise.
3931 (cgraph_next_defined_function): Likewise.
3932 (cgraph_first_function): Likewise.
3933 (cgraph_next_function): Likewise.
3934 (cgraph_first_function_with_gimple_body): Likewise.
3935 (cgraph_next_function_with_gimple_body): Likewise.
3936 (cgraph_alias_target): Likewise.
3937 (varpool_alias_target): Likewise.
3938 (cgraph_function_or_thunk_node): Likewise.
3939 (varpool_variable_node): Likewise.
3940 (symtab_real_symbol_p): Likewise.
3941 * cgraphunit.c (referred_to_p): Likewise.
3942 (analyze_functions): Likewise.
3943 (handle_alias_pairs): Likewise.
3944 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3945 * gimple-ssa.h (gimple_vuse_op): Likewise.
3946 (gimple_vdef_op): Likewise.
3947 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3948 * gimple.c (gimple_build_asm_1): Likewise.
3949 (gimple_build_try): Likewise.
3950 (gimple_build_resx): Likewise.
3951 (gimple_build_eh_dispatch): Likewise.
3952 (gimple_build_omp_for): Likewise.
3953 (gimple_omp_for_set_clauses): Likewise.
3955 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3956 (is_a_helper <gimple_statement_asm *>::test): ...this.
3957 (is_a_helper <gimple_statement_bind>::test): Convert to...
3958 (is_a_helper <gimple_statement_bind *>::test): ...this.
3959 (is_a_helper <gimple_statement_call>::test): Convert to...
3960 (is_a_helper <gimple_statement_call *>::test): ...this.
3961 (is_a_helper <gimple_statement_catch>::test): Convert to...
3962 (is_a_helper <gimple_statement_catch *>::test): ...this.
3963 (is_a_helper <gimple_statement_resx>::test): Convert to...
3964 (is_a_helper <gimple_statement_resx *>::test): ...this.
3965 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3966 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3967 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3968 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3969 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3970 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3971 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3972 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3973 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3974 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3975 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3976 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3977 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3978 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3979 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3980 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3981 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3982 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3983 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3984 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3985 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3986 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3987 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3988 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3989 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3990 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3991 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3992 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3993 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3994 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3995 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3996 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3997 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3998 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3999 (is_a_helper <gimple_statement_phi>::test): Convert to...
4000 (is_a_helper <gimple_statement_phi *>::test): ...this.
4001 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4002 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4003 (is_a_helper <gimple_statement_try>::test): Convert to...
4004 (is_a_helper <gimple_statement_try *>::test): ...this.
4005 (is_a_helper <gimple_statement_wce>::test): Convert to...
4006 (is_a_helper <gimple_statement_wce *>::test): ...this.
4007 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4008 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4009 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4010 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4011 (is_a_helper <const gimple_statement_call>::test): Convert to...
4012 (is_a_helper <const gimple_statement_call *>::test): ...this.
4013 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4014 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4015 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4016 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4017 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4018 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4019 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4020 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4021 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4023 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4025 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4027 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4029 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4030 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4031 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4033 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4034 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4036 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4037 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4038 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4039 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4040 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4041 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4043 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4044 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4045 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4046 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4048 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4049 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4050 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4051 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4052 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4053 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4054 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4055 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4056 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4057 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4058 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4059 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4060 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4061 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4062 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4063 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4065 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4067 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4068 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4070 (gimple_use_ops): Update for removal of implicit pointer from the
4072 (gimple_set_use_ops): Likewise.
4073 (gimple_vuse): Likewise.
4074 (gimple_vdef): Likewise.
4075 (gimple_vuse_ptr): Likewise.
4076 (gimple_vdef_ptr): Likewise.
4077 (gimple_set_vuse): Likewise.
4078 (gimple_set_vdef): Likewise.
4079 (gimple_omp_return_set_lhs): Likewise.
4080 (gimple_omp_return_lhs): Likewise.
4081 (gimple_omp_return_lhs_ptr): Likewise.
4082 (gimple_call_fntype): Likewise.
4083 (gimple_call_set_fntype): Likewise.
4084 (gimple_call_set_internal_fn): Likewise.
4085 (gimple_call_use_set): Likewise.
4086 (gimple_call_clobber_set): Likewise.
4087 (gimple_bind_vars): Likewise.
4088 (gimple_bind_set_vars): Likewise.
4089 (gimple_bind_body_ptr): Likewise.
4090 (gimple_bind_set_body): Likewise.
4091 (gimple_bind_add_stmt): Likewise.
4092 (gimple_bind_block): Likewise.
4093 (gimple_bind_set_block): Likewise.
4094 (gimple_asm_ninputs): Likewise.
4095 (gimple_asm_noutputs): Likewise.
4096 (gimple_asm_nclobbers): Likewise.
4097 (gimple_asm_nlabels): Likewise.
4098 (gimple_asm_input_op): Likewise.
4099 (gimple_asm_input_op_ptr): Likewise.
4100 (gimple_asm_output_op): Likewise.
4101 (gimple_asm_output_op_ptr): Likewise.
4102 (gimple_asm_set_output_op): Likewise.
4103 (gimple_asm_clobber_op): Likewise.
4104 (gimple_asm_set_clobber_op): Likewise.
4105 (gimple_asm_label_op): Likewise.
4106 (gimple_asm_set_label_op): Likewise.
4107 (gimple_asm_string): Likewise.
4108 (gimple_catch_types): Likewise.
4109 (gimple_catch_types_ptr): Likewise.
4110 (gimple_catch_handler_ptr): Likewise.
4111 (gimple_catch_set_types): Likewise.
4112 (gimple_catch_set_handler): Likewise.
4113 (gimple_eh_filter_types): Likewise.
4114 (gimple_eh_filter_types_ptr): Likewise.
4115 (gimple_eh_filter_failure_ptr): Likewise.
4116 (gimple_eh_filter_set_types): Likewise.
4117 (gimple_eh_filter_set_failure): Likewise.
4118 (gimple_eh_must_not_throw_fndecl): Likewise.
4119 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4120 (gimple_eh_else_n_body_ptr): Likewise.
4121 (gimple_eh_else_e_body_ptr): Likewise.
4122 (gimple_eh_else_set_n_body): Likewise.
4123 (gimple_eh_else_set_e_body): Likewise.
4124 (gimple_try_eval_ptr): Likewise.
4125 (gimple_try_cleanup_ptr): Likewise.
4126 (gimple_try_set_eval): Likewise.
4127 (gimple_try_set_cleanup): Likewise.
4128 (gimple_wce_cleanup_ptr): Likewise.
4129 (gimple_wce_set_cleanup): Likewise.
4130 (gimple_phi_capacity): Likewise.
4131 (gimple_phi_num_args): Likewise.
4132 (gimple_phi_result): Likewise.
4133 (gimple_phi_result_ptr): Likewise.
4134 (gimple_phi_set_result): Likewise.
4135 (gimple_phi_arg): Likewise.
4136 (gimple_phi_set_arg): Likewise.
4137 (gimple_resx_region): Likewise.
4138 (gimple_resx_set_region): Likewise.
4139 (gimple_eh_dispatch_region): Likewise.
4140 (gimple_eh_dispatch_set_region): Likewise.
4141 (gimple_omp_critical_name): Likewise.
4142 (gimple_omp_critical_name_ptr): Likewise.
4143 (gimple_omp_critical_set_name): Likewise.
4144 (gimple_omp_for_clauses): Likewise.
4145 (gimple_omp_for_clauses_ptr): Likewise.
4146 (gimple_omp_for_set_clauses): Likewise.
4147 (gimple_omp_for_collapse): Likewise.
4148 (gimple_omp_for_index): Likewise.
4149 (gimple_omp_for_index_ptr): Likewise.
4150 (gimple_omp_for_set_index): Likewise.
4151 (gimple_omp_for_initial): Likewise.
4152 (gimple_omp_for_initial_ptr): Likewise.
4153 (gimple_omp_for_set_initial): Likewise.
4154 (gimple_omp_for_final): Likewise.
4155 (gimple_omp_for_final_ptr): Likewise.
4156 (gimple_omp_for_set_final): Likewise.
4157 (gimple_omp_for_incr): Likewise.
4158 (gimple_omp_for_incr_ptr): Likewise.
4159 (gimple_omp_for_set_incr): Likewise.
4160 (gimple_omp_for_pre_body_ptr): Likewise.
4161 (gimple_omp_for_set_pre_body): Likewise.
4162 (gimple_omp_parallel_clauses): Likewise.
4163 (gimple_omp_parallel_clauses_ptr): Likewise.
4164 (gimple_omp_parallel_set_clauses): Likewise.
4165 (gimple_omp_parallel_child_fn): Likewise.
4166 (gimple_omp_parallel_child_fn_ptr): Likewise.
4167 (gimple_omp_parallel_set_child_fn): Likewise.
4168 (gimple_omp_parallel_data_arg): Likewise.
4169 (gimple_omp_parallel_data_arg_ptr): Likewise.
4170 (gimple_omp_parallel_set_data_arg): Likewise.
4171 (gimple_omp_task_clauses): Likewise.
4172 (gimple_omp_task_clauses_ptr): Likewise.
4173 (gimple_omp_task_set_clauses): Likewise.
4174 (gimple_omp_task_child_fn): Likewise.
4175 (gimple_omp_task_child_fn_ptr): Likewise.
4176 (gimple_omp_task_set_child_fn): Likewise.
4177 (gimple_omp_task_data_arg): Likewise.
4178 (gimple_omp_task_data_arg_ptr): Likewise.
4179 (gimple_omp_task_set_data_arg): Likewise.
4180 (gimple_omp_taskreg_clauses): Likewise.
4181 (gimple_omp_taskreg_clauses_ptr): Likewise.
4182 (gimple_omp_taskreg_set_clauses): Likewise.
4183 (gimple_omp_taskreg_child_fn): Likewise.
4184 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4185 (gimple_omp_taskreg_set_child_fn): Likewise.
4186 (gimple_omp_taskreg_data_arg): Likewise.
4187 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4188 (gimple_omp_taskreg_set_data_arg): Likewise.
4189 (gimple_omp_task_copy_fn): Likewise.
4190 (gimple_omp_task_copy_fn_ptr): Likewise.
4191 (gimple_omp_task_set_copy_fn): Likewise.
4192 (gimple_omp_task_arg_size): Likewise.
4193 (gimple_omp_task_arg_size_ptr): Likewise.
4194 (gimple_omp_task_set_arg_size): Likewise.
4195 (gimple_omp_task_arg_align): Likewise.
4196 (gimple_omp_task_arg_align_ptr): Likewise.
4197 (gimple_omp_task_set_arg_align): Likewise.
4198 (gimple_omp_single_clauses): Likewise.
4199 (gimple_omp_single_clauses_ptr): Likewise.
4200 (gimple_omp_single_set_clauses): Likewise.
4201 (gimple_omp_target_clauses): Likewise.
4202 (gimple_omp_target_clauses_ptr): Likewise.
4203 (gimple_omp_target_set_clauses): Likewise.
4204 (gimple_omp_target_child_fn): Likewise.
4205 (gimple_omp_target_child_fn_ptr): Likewise.
4206 (gimple_omp_target_set_child_fn): Likewise.
4207 (gimple_omp_target_data_arg): Likewise.
4208 (gimple_omp_target_data_arg_ptr): Likewise.
4209 (gimple_omp_target_set_data_arg): Likewise.
4210 (gimple_omp_teams_clauses): Likewise.
4211 (gimple_omp_teams_clauses_ptr): Likewise.
4212 (gimple_omp_teams_set_clauses): Likewise.
4213 (gimple_omp_sections_clauses): Likewise.
4214 (gimple_omp_sections_clauses_ptr): Likewise.
4215 (gimple_omp_sections_set_clauses): Likewise.
4216 (gimple_omp_sections_control): Likewise.
4217 (gimple_omp_sections_control_ptr): Likewise.
4218 (gimple_omp_sections_set_control): Likewise.
4219 (gimple_omp_for_set_cond): Likewise.
4220 (gimple_omp_for_cond): Likewise.
4221 (gimple_omp_atomic_store_set_val): Likewise.
4222 (gimple_omp_atomic_store_val): Likewise.
4223 (gimple_omp_atomic_store_val_ptr): Likewise.
4224 (gimple_omp_atomic_load_set_lhs): Likewise.
4225 (gimple_omp_atomic_load_lhs): Likewise.
4226 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4227 (gimple_omp_atomic_load_set_rhs): Likewise.
4228 (gimple_omp_atomic_load_rhs): Likewise.
4229 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4230 (gimple_omp_continue_control_def): Likewise.
4231 (gimple_omp_continue_control_def_ptr): Likewise.
4232 (gimple_omp_continue_set_control_def): Likewise.
4233 (gimple_omp_continue_control_use): Likewise.
4234 (gimple_omp_continue_control_use_ptr): Likewise.
4235 (gimple_omp_continue_set_control_use): Likewise.
4236 (gimple_transaction_body_ptr): Likewise.
4237 (gimple_transaction_label): Likewise.
4238 (gimple_transaction_label_ptr): Likewise.
4239 (gimple_transaction_set_body): Likewise.
4240 (gimple_transaction_set_label): Likewise.
4242 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4243 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4244 * ipa-ref.c (ipa_record_reference): Likewise.
4245 * ipa-reference.c (analyze_function): Likewise.
4246 (ipa_reference_write_optimization_summary): Likewise.
4247 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4248 (address_taken_from_non_vtable_p): Likewise.
4249 (comdat_can_be_unshared_p_1): Likewise.
4250 * lto-cgraph.c (lto_output_ref): Likewise.
4251 (add_references): Likewise.
4252 (compute_ltrans_boundary): Likewise.
4253 (output_symtab): Likewise.
4254 (input_ref): Likewise.
4255 (input_cgraph_1): Likewise.
4256 (output_cgraph_opt_summary): Likewise.
4257 * lto-streamer-out.c (lto_output): Likewise.
4258 (output_symbol_p): Likewise.
4259 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4260 (lsei_start_function_in_partition): Likewise.
4261 (lsei_next_variable_in_partition): Likewise.
4262 (lsei_start_variable_in_partition): Likewise.
4263 * symtab.c (insert_to_assembler_name_hash): Likewise.
4264 (unlink_from_assembler_name_hash): Likewise.
4265 (symtab_unregister_node): Likewise.
4266 (symtab_remove_node): Likewise.
4267 (dump_symtab_node): Likewise.
4268 (verify_symtab_base): Likewise.
4269 (verify_symtab_node): Likewise.
4270 (symtab_make_decl_local): Likewise.
4271 (symtab_alias_ultimate_target): Likewise.
4272 (symtab_resolve_alias): Likewise.
4273 (symtab_get_symbol_partitioning_class): Likewise.
4274 * tree-phinodes.c (allocate_phi_node): Likewise.
4275 (reserve_phi_args_for_new_edge): Likewise.
4276 (remove_phi_args): Likewise.
4277 * varpool.c (varpool_node_for_asm): Likewise.
4278 (varpool_remove_unreferenced_decls): Likewise.
4280 2014-04-23 Jeff Law <law@redhat.com>
4282 PR tree-optimization/60902
4283 * tree-ssa-threadedge.c
4284 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4285 invalidate outputs from statements that do not produce useful
4286 outputs for threading.
4288 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4290 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4291 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4292 machine descriptions for Stack Smashing Protector.
4294 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4296 * aarch64.md (<optab>_rol<mode>3): New pattern.
4297 (<optab>_rolsi3_uxtw): Likewise.
4298 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4300 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4302 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4303 (arm_cortex_a12_tune): Likewise.
4305 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4307 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4309 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4311 * config/arm/arm.md (arm_rev16si2): New pattern.
4312 (arm_rev16si2_alt): Likewise.
4313 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4315 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4317 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4318 (rev16<mode>2_alt): Likewise.
4319 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4320 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4321 (aarch_rev16_shleft_mask_imm_p): Likewise.
4322 (aarch_rev16_p_1): Likewise.
4323 (aarch_rev16_p): Likewise.
4324 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4325 (aarch_rev16_shright_mask_imm_p): Likewise.
4326 (aarch_rev16_shleft_mask_imm_p): Likewise.
4328 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4330 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4331 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4333 (cortex_a53_extra_costs): Likewise.
4334 (cortex_a57_extra_costs): Likewise.
4335 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4336 (cortexa7_extra_costs): Likewise.
4337 (cortexa8_extra_costs): Likewise.
4338 (cortexa12_extra_costs): Likewise.
4339 (cortexa15_extra_costs): Likewise.
4340 (v7m_extra_costs): Likewise.
4341 (arm_new_rtx_costs): Handle BSWAP.
4343 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4345 * config/arm/arm.c (cortexa8_extra_costs): New table.
4346 (arm_cortex_a8_tune): New tuning struct.
4347 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4349 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4351 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4353 2014-04-23 Richard Biener <rguenther@suse.de>
4355 * Makefile.in (OBJS): Remove loop-unswitch.o.
4356 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4357 * passes.def (pass_rtl_unswitch): Likewise.
4358 * loop-init.c (gate_rtl_unswitch): Likewise.
4359 (rtl_unswitch): Likewise.
4360 (pass_data_rtl_unswitch): Likewise.
4361 (pass_rtl_unswitch): Likewise.
4362 (make_pass_rtl_unswitch): Likewise.
4363 * rtl.h (reversed_condition): Likewise.
4364 (compare_and_jump_seq): Likewise.
4365 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4367 * loop-unroll.c (compare_and_jump_seq): Likewise.
4369 2014-04-23 Richard Biener <rguenther@suse.de>
4371 PR tree-optimization/60903
4372 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4373 commented code block.
4374 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4375 loop flags to newly created BBs and edges.
4377 2014-04-23 Nick Clifton <nickc@redhat.com>
4379 * config/msp430/msp430.c (msp430_handle_option): Move function
4381 (msp430_option_override): Simplify mcu and mcpu option handling.
4382 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4383 support for -mhwmult command line option.
4384 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4385 -mhwmult command line option.
4386 (msp430_hwmult_enabled): Delete.
4387 (msp43o_output_labelref): Add support for -mhwmult command line option.
4388 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4389 (umulsidi3): Likewise.
4390 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4391 (mcpu, mlarge, msmall): Likewise.
4392 (mhwmult): New option.
4393 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4395 (msp430_is_f5_mcu): Remove prototype.
4396 (msp430_use_f5_series_hwmult): Add prototype.
4397 * config/msp430/msp430-opts.h: New file.
4398 * common/config/msp430: New directory.
4399 * common/config/msp430/msp430-common.c: New file.
4400 * config.gcc (msp430): Remove target_has_targetm_common.
4401 * doc/invoke.texi: Document -mhwmult command line option.
4403 2014-04-23 Nick Clifton <nickc@redhat.com>
4405 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4406 default-manifest.o if it can be found in the search path.
4407 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4409 2014-04-23 Terry Guo <terry.guo@arm.com>
4411 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4413 2014-04-23 Richard Biener <rguenther@suse.de>
4416 * tree-inline.c (declare_return_variable): Use mark_addressable.
4418 2014-04-23 Richard Biener <rguenther@suse.de>
4421 * loop-init.c (loop_optimizer_init): Make sure to apply
4422 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4424 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4427 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4429 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4430 if flag_sanitize_undefined_trap_on_error.
4431 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4432 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4433 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4434 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4435 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4436 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4437 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4438 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4439 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4440 * ubsan.c (ubsan_instrument_unreachable): Return
4441 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4442 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4443 if flag_sanitize_undefined_trap_on_error and
4444 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4445 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4446 instrument_bool_enum_load): Emit __builtin_trap () if
4447 flag_sanitize_undefined_trap_on_error and
4448 __builtin_handle_*_abort () if !flag_sanitize_recover.
4449 * doc/invoke.texi (-fsanitize-recover,
4450 -fsanitize-undefined-trap-on-error): Document.
4452 2014-04-22 Christian Bruel <christian.bruel@st.com>
4454 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4455 Force immediates to SImode.
4457 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4459 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4461 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4462 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4463 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4464 (nios2_fpu_insn): Add entry for round.
4465 (N2FPU_NO_ERRNO_P): Define.
4466 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4468 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4470 2014-04-22 Richard Henderson <rth@redhat.com>
4472 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4473 (add<GPI>3_compare0): Remove leading * from name.
4474 (add<GPI>3_carryin): Likewise.
4475 (sub<GPI>3_compare0): Likewise.
4476 (sub<GPI>3_carryin): Likewise.
4477 (<su_optab>mulditi3): New expander.
4478 (multi3): New expander.
4479 (madd<GPI>): Remove leading * from name.
4481 2014-04-22 Martin Jambor <mjambor@suse.cz>
4483 * cgraphclones.c (cgraph_function_versioning): Copy
4484 ipa_transforms_to_apply instead of asserting it is empty.
4486 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4489 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4490 on count_exp to get mode.
4492 2014-04-22 Andrew Pinski <apinski@cavium.com>
4494 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4495 Handle TLS for ILP32.
4496 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4497 (tlsie_small_<mode>): this and handle PTR.
4498 (tlsie_small_sidi): New pattern.
4499 (tlsle_small): Change to an expand to handle ILP32.
4500 (tlsle_small_<mode>): New pattern.
4501 (tlsdesc_small): Rename to ...
4502 (tlsdesc_small_<mode>): this and handle PTR.
4504 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4506 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4508 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4510 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4511 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4512 (aarch64_types_signed_poly_qualifiers): Likewise.
4513 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4514 (aarch64_types_poly_signed_qualifiers): Likewise.
4515 (TYPES_REINTERP_SS): Type macro added.
4516 (TYPES_REINTERP_SU): Likewise.
4517 (TYPES_REINTERP_SP): Likewise.
4518 (TYPES_REINTERP_US): Likewise.
4519 (TYPES_REINTERP_PS): Likewise.
4520 (aarch64_fold_builtin): New expression folding added.
4521 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4522 Declarations removed.
4523 (REINTERP_SS): Declarations added.
4524 (REINTERP_US): Likewise.
4525 (REINTERP_PS): Likewise.
4526 (REINTERP_SU): Likewise.
4527 (REINTERP_SP): Likewise.
4528 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4529 (vreinterpretq_p8_f64): Likewise.
4530 (vreinterpret_p16_f64): Likewise.
4531 (vreinterpretq_p16_f64): Likewise.
4532 (vreinterpret_f32_f64): Likewise.
4533 (vreinterpretq_f32_f64): Likewise.
4534 (vreinterpret_f64_f32): Likewise.
4535 (vreinterpret_f64_p8): Likewise.
4536 (vreinterpret_f64_p16): Likewise.
4537 (vreinterpret_f64_s8): Likewise.
4538 (vreinterpret_f64_s16): Likewise.
4539 (vreinterpret_f64_s32): Likewise.
4540 (vreinterpret_f64_s64): Likewise.
4541 (vreinterpret_f64_u8): Likewise.
4542 (vreinterpret_f64_u16): Likewise.
4543 (vreinterpret_f64_u32): Likewise.
4544 (vreinterpret_f64_u64): Likewise.
4545 (vreinterpretq_f64_f32): Likewise.
4546 (vreinterpretq_f64_p8): Likewise.
4547 (vreinterpretq_f64_p16): Likewise.
4548 (vreinterpretq_f64_s8): Likewise.
4549 (vreinterpretq_f64_s16): Likewise.
4550 (vreinterpretq_f64_s32): Likewise.
4551 (vreinterpretq_f64_s64): Likewise.
4552 (vreinterpretq_f64_u8): Likewise.
4553 (vreinterpretq_f64_u16): Likewise.
4554 (vreinterpretq_f64_u32): Likewise.
4555 (vreinterpretq_f64_u64): Likewise.
4556 (vreinterpret_s64_f64): Likewise.
4557 (vreinterpretq_s64_f64): Likewise.
4558 (vreinterpret_u64_f64): Likewise.
4559 (vreinterpretq_u64_f64): Likewise.
4560 (vreinterpret_s8_f64): Likewise.
4561 (vreinterpretq_s8_f64): Likewise.
4562 (vreinterpret_s16_f64): Likewise.
4563 (vreinterpretq_s16_f64): Likewise.
4564 (vreinterpret_s32_f64): Likewise.
4565 (vreinterpretq_s32_f64): Likewise.
4566 (vreinterpret_u8_f64): Likewise.
4567 (vreinterpretq_u8_f64): Likewise.
4568 (vreinterpret_u16_f64): Likewise.
4569 (vreinterpretq_u16_f64): Likewise.
4570 (vreinterpret_u32_f64): Likewise.
4571 (vreinterpretq_u32_f64): Likewise.
4573 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4575 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4576 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4577 (vreinterpret_p8_s8): Likewise.
4578 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4579 (vreinterpret_p8_s16): Likewise.
4580 (vreinterpret_p8_s32): Likewise.
4581 (vreinterpret_p8_s64): Likewise.
4582 (vreinterpret_p8_f32): Likewise.
4583 (vreinterpret_p8_u8): Likewise.
4584 (vreinterpret_p8_u16): Likewise.
4585 (vreinterpret_p8_u32): Likewise.
4586 (vreinterpret_p8_u64): Likewise.
4587 (vreinterpret_p8_p16): Likewise.
4588 (vreinterpretq_p8_s8): Likewise.
4589 (vreinterpretq_p8_s16): Likewise.
4590 (vreinterpretq_p8_s32): Likewise.
4591 (vreinterpretq_p8_s64): Likewise.
4592 (vreinterpretq_p8_f32): Likewise.
4593 (vreinterpretq_p8_u8): Likewise.
4594 (vreinterpretq_p8_u16): Likewise.
4595 (vreinterpretq_p8_u32): Likewise.
4596 (vreinterpretq_p8_u64): Likewise.
4597 (vreinterpretq_p8_p16): Likewise.
4598 (vreinterpret_p16_s8): Likewise.
4599 (vreinterpret_p16_s16): Likewise.
4600 (vreinterpret_p16_s32): Likewise.
4601 (vreinterpret_p16_s64): Likewise.
4602 (vreinterpret_p16_f32): Likewise.
4603 (vreinterpret_p16_u8): Likewise.
4604 (vreinterpret_p16_u16): Likewise.
4605 (vreinterpret_p16_u32): Likewise.
4606 (vreinterpret_p16_u64): Likewise.
4607 (vreinterpret_p16_p8): Likewise.
4608 (vreinterpretq_p16_s8): Likewise.
4609 (vreinterpretq_p16_s16): Likewise.
4610 (vreinterpretq_p16_s32): Likewise.
4611 (vreinterpretq_p16_s64): Likewise.
4612 (vreinterpretq_p16_f32): Likewise.
4613 (vreinterpretq_p16_u8): Likewise.
4614 (vreinterpretq_p16_u16): Likewise.
4615 (vreinterpretq_p16_u32): Likewise.
4616 (vreinterpretq_p16_u64): Likewise.
4617 (vreinterpretq_p16_p8): Likewise.
4618 (vreinterpret_f32_s8): Likewise.
4619 (vreinterpret_f32_s16): Likewise.
4620 (vreinterpret_f32_s32): Likewise.
4621 (vreinterpret_f32_s64): Likewise.
4622 (vreinterpret_f32_u8): Likewise.
4623 (vreinterpret_f32_u16): Likewise.
4624 (vreinterpret_f32_u32): Likewise.
4625 (vreinterpret_f32_u64): Likewise.
4626 (vreinterpret_f32_p8): Likewise.
4627 (vreinterpret_f32_p16): Likewise.
4628 (vreinterpretq_f32_s8): Likewise.
4629 (vreinterpretq_f32_s16): Likewise.
4630 (vreinterpretq_f32_s32): Likewise.
4631 (vreinterpretq_f32_s64): Likewise.
4632 (vreinterpretq_f32_u8): Likewise.
4633 (vreinterpretq_f32_u16): Likewise.
4634 (vreinterpretq_f32_u32): Likewise.
4635 (vreinterpretq_f32_u64): Likewise.
4636 (vreinterpretq_f32_p8): Likewise.
4637 (vreinterpretq_f32_p16): Likewise.
4638 (vreinterpret_s64_s8): Likewise.
4639 (vreinterpret_s64_s16): Likewise.
4640 (vreinterpret_s64_s32): Likewise.
4641 (vreinterpret_s64_f32): Likewise.
4642 (vreinterpret_s64_u8): Likewise.
4643 (vreinterpret_s64_u16): Likewise.
4644 (vreinterpret_s64_u32): Likewise.
4645 (vreinterpret_s64_u64): Likewise.
4646 (vreinterpret_s64_p8): Likewise.
4647 (vreinterpret_s64_p16): Likewise.
4648 (vreinterpretq_s64_s8): Likewise.
4649 (vreinterpretq_s64_s16): Likewise.
4650 (vreinterpretq_s64_s32): Likewise.
4651 (vreinterpretq_s64_f32): Likewise.
4652 (vreinterpretq_s64_u8): Likewise.
4653 (vreinterpretq_s64_u16): Likewise.
4654 (vreinterpretq_s64_u32): Likewise.
4655 (vreinterpretq_s64_u64): Likewise.
4656 (vreinterpretq_s64_p8): Likewise.
4657 (vreinterpretq_s64_p16): Likewise.
4658 (vreinterpret_u64_s8): Likewise.
4659 (vreinterpret_u64_s16): Likewise.
4660 (vreinterpret_u64_s32): Likewise.
4661 (vreinterpret_u64_s64): Likewise.
4662 (vreinterpret_u64_f32): Likewise.
4663 (vreinterpret_u64_u8): Likewise.
4664 (vreinterpret_u64_u16): Likewise.
4665 (vreinterpret_u64_u32): Likewise.
4666 (vreinterpret_u64_p8): Likewise.
4667 (vreinterpret_u64_p16): Likewise.
4668 (vreinterpretq_u64_s8): Likewise.
4669 (vreinterpretq_u64_s16): Likewise.
4670 (vreinterpretq_u64_s32): Likewise.
4671 (vreinterpretq_u64_s64): Likewise.
4672 (vreinterpretq_u64_f32): Likewise.
4673 (vreinterpretq_u64_u8): Likewise.
4674 (vreinterpretq_u64_u16): Likewise.
4675 (vreinterpretq_u64_u32): Likewise.
4676 (vreinterpretq_u64_p8): Likewise.
4677 (vreinterpretq_u64_p16): Likewise.
4678 (vreinterpret_s8_s16): Likewise.
4679 (vreinterpret_s8_s32): Likewise.
4680 (vreinterpret_s8_s64): Likewise.
4681 (vreinterpret_s8_f32): Likewise.
4682 (vreinterpret_s8_u8): Likewise.
4683 (vreinterpret_s8_u16): Likewise.
4684 (vreinterpret_s8_u32): Likewise.
4685 (vreinterpret_s8_u64): Likewise.
4686 (vreinterpret_s8_p8): Likewise.
4687 (vreinterpret_s8_p16): Likewise.
4688 (vreinterpretq_s8_s16): Likewise.
4689 (vreinterpretq_s8_s32): Likewise.
4690 (vreinterpretq_s8_s64): Likewise.
4691 (vreinterpretq_s8_f32): Likewise.
4692 (vreinterpretq_s8_u8): Likewise.
4693 (vreinterpretq_s8_u16): Likewise.
4694 (vreinterpretq_s8_u32): Likewise.
4695 (vreinterpretq_s8_u64): Likewise.
4696 (vreinterpretq_s8_p8): Likewise.
4697 (vreinterpretq_s8_p16): Likewise.
4698 (vreinterpret_s16_s8): Likewise.
4699 (vreinterpret_s16_s32): Likewise.
4700 (vreinterpret_s16_s64): Likewise.
4701 (vreinterpret_s16_f32): Likewise.
4702 (vreinterpret_s16_u8): Likewise.
4703 (vreinterpret_s16_u16): Likewise.
4704 (vreinterpret_s16_u32): Likewise.
4705 (vreinterpret_s16_u64): Likewise.
4706 (vreinterpret_s16_p8): Likewise.
4707 (vreinterpret_s16_p16): Likewise.
4708 (vreinterpretq_s16_s8): Likewise.
4709 (vreinterpretq_s16_s32): Likewise.
4710 (vreinterpretq_s16_s64): Likewise.
4711 (vreinterpretq_s16_f32): Likewise.
4712 (vreinterpretq_s16_u8): Likewise.
4713 (vreinterpretq_s16_u16): Likewise.
4714 (vreinterpretq_s16_u32): Likewise.
4715 (vreinterpretq_s16_u64): Likewise.
4716 (vreinterpretq_s16_p8): Likewise.
4717 (vreinterpretq_s16_p16): Likewise.
4718 (vreinterpret_s32_s8): Likewise.
4719 (vreinterpret_s32_s16): Likewise.
4720 (vreinterpret_s32_s64): Likewise.
4721 (vreinterpret_s32_f32): Likewise.
4722 (vreinterpret_s32_u8): Likewise.
4723 (vreinterpret_s32_u16): Likewise.
4724 (vreinterpret_s32_u32): Likewise.
4725 (vreinterpret_s32_u64): Likewise.
4726 (vreinterpret_s32_p8): Likewise.
4727 (vreinterpret_s32_p16): Likewise.
4728 (vreinterpretq_s32_s8): Likewise.
4729 (vreinterpretq_s32_s16): Likewise.
4730 (vreinterpretq_s32_s64): Likewise.
4731 (vreinterpretq_s32_f32): Likewise.
4732 (vreinterpretq_s32_u8): Likewise.
4733 (vreinterpretq_s32_u16): Likewise.
4734 (vreinterpretq_s32_u32): Likewise.
4735 (vreinterpretq_s32_u64): Likewise.
4736 (vreinterpretq_s32_p8): Likewise.
4737 (vreinterpretq_s32_p16): Likewise.
4738 (vreinterpret_u8_s8): Likewise.
4739 (vreinterpret_u8_s16): Likewise.
4740 (vreinterpret_u8_s32): Likewise.
4741 (vreinterpret_u8_s64): Likewise.
4742 (vreinterpret_u8_f32): Likewise.
4743 (vreinterpret_u8_u16): Likewise.
4744 (vreinterpret_u8_u32): Likewise.
4745 (vreinterpret_u8_u64): Likewise.
4746 (vreinterpret_u8_p8): Likewise.
4747 (vreinterpret_u8_p16): Likewise.
4748 (vreinterpretq_u8_s8): Likewise.
4749 (vreinterpretq_u8_s16): Likewise.
4750 (vreinterpretq_u8_s32): Likewise.
4751 (vreinterpretq_u8_s64): Likewise.
4752 (vreinterpretq_u8_f32): Likewise.
4753 (vreinterpretq_u8_u16): Likewise.
4754 (vreinterpretq_u8_u32): Likewise.
4755 (vreinterpretq_u8_u64): Likewise.
4756 (vreinterpretq_u8_p8): Likewise.
4757 (vreinterpretq_u8_p16): Likewise.
4758 (vreinterpret_u16_s8): Likewise.
4759 (vreinterpret_u16_s16): Likewise.
4760 (vreinterpret_u16_s32): Likewise.
4761 (vreinterpret_u16_s64): Likewise.
4762 (vreinterpret_u16_f32): Likewise.
4763 (vreinterpret_u16_u8): Likewise.
4764 (vreinterpret_u16_u32): Likewise.
4765 (vreinterpret_u16_u64): Likewise.
4766 (vreinterpret_u16_p8): Likewise.
4767 (vreinterpret_u16_p16): Likewise.
4768 (vreinterpretq_u16_s8): Likewise.
4769 (vreinterpretq_u16_s16): Likewise.
4770 (vreinterpretq_u16_s32): Likewise.
4771 (vreinterpretq_u16_s64): Likewise.
4772 (vreinterpretq_u16_f32): Likewise.
4773 (vreinterpretq_u16_u8): Likewise.
4774 (vreinterpretq_u16_u32): Likewise.
4775 (vreinterpretq_u16_u64): Likewise.
4776 (vreinterpretq_u16_p8): Likewise.
4777 (vreinterpretq_u16_p16): Likewise.
4778 (vreinterpret_u32_s8): Likewise.
4779 (vreinterpret_u32_s16): Likewise.
4780 (vreinterpret_u32_s32): Likewise.
4781 (vreinterpret_u32_s64): Likewise.
4782 (vreinterpret_u32_f32): Likewise.
4783 (vreinterpret_u32_u8): Likewise.
4784 (vreinterpret_u32_u16): Likewise.
4785 (vreinterpret_u32_u64): Likewise.
4786 (vreinterpret_u32_p8): Likewise.
4787 (vreinterpret_u32_p16): Likewise.
4788 (vreinterpretq_u32_s8): Likewise.
4789 (vreinterpretq_u32_s16): Likewise.
4790 (vreinterpretq_u32_s32): Likewise.
4791 (vreinterpretq_u32_s64): Likewise.
4792 (vreinterpretq_u32_f32): Likewise.
4793 (vreinterpretq_u32_u8): Likewise.
4794 (vreinterpretq_u32_u16): Likewise.
4795 (vreinterpretq_u32_u64): Likewise.
4796 (vreinterpretq_u32_p8): Likewise.
4797 (vreinterpretq_u32_p16): Likewise.
4799 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4801 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4803 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4805 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4806 (vqnegd_s64): Likewise.
4807 (vqabs_s64): Likewise.
4808 (vqabsd_s64): Likewise.
4810 2014-04-22 Richard Henderson <rth@redhat.com>
4812 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4813 computation to the top of the loop.
4815 2014-04-22 Renlin <renlin.li@arm.com>
4816 Jiong Wang <jiong.wang@arm.com>
4818 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4819 * config/aarch64/aarch64.c (aarch64_layout_frame)
4820 (aarch64_initial_elimination_offset): Likewise.
4822 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4824 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4827 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4829 * machmode.h (bitwise_mode_for_mode): Declare.
4830 * stor-layout.h (bitwise_type_for_mode): Likewise.
4831 * stor-layout.c (bitwise_mode_for_mode): New function.
4832 (bitwise_type_for_mode): Likewise.
4833 * builtins.c (fold_builtin_memory_op): Use it instead of
4834 int_mode_for_mode and build_nonstandard_integer_type.
4836 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4838 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4839 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4840 (*-*-solaris2*): Simplify.
4841 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4842 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4843 *-*-solaris2.9* handling.
4845 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4847 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4848 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4850 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4851 * configure: Regenerate.
4853 * config/i386/sol2-9.h: Remove.
4855 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4856 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4857 Remove Solaris 9 references.
4859 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4861 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4862 (floatuns<GPI:mode><GPF:mode>2): Remove.
4863 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4864 and floatuns conversions.
4865 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4866 and floatuns conversions.
4867 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4868 (w1,w2): New mode attributes for inequal width conversions.
4870 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4872 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4873 the output asm format.
4875 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4877 * config/aarch64/aarch64-simd.md
4878 (aarch64_cm<optab>di): Always split.
4879 (*aarch64_cm<optab>di): New.
4880 (aarch64_cmtstdi): Always split.
4881 (*aarch64_cmtstdi): New.
4883 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4885 PR tree-optimization/60823
4886 * omp-low.c (ipa_simd_modify_function_body): Go through
4887 all SSA_NAMEs and for those refering to vector arguments
4888 which are going to be replaced adjust SSA_NAME_VAR and,
4889 if it is a default definition, change it into a non-default
4890 definition assigned at the beginning of function from new_decl.
4891 (ipa_simd_modify_stmt_ops): Rewritten.
4892 * tree-dfa.c (set_ssa_default_def): When removing default def,
4893 check for NULL loc instead of NULL *loc.
4895 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4897 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4898 restrictions on core registers for DImode values in Thumb2.
4900 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4902 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4903 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4905 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4907 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4908 (*iordi_notzesidi_di): Likewise.
4909 (*iordi_notsesidi_di): Likewise.
4911 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4913 * config/arm/arm-protos.h (tune_params): New struct members.
4914 * config/arm/arm.c: Initialise tune_params per processor.
4915 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4916 for speed, based on new tune_params.
4918 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4920 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4921 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4922 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4923 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4924 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4925 (vrnda_f64): Likewise.
4926 (vrndi_f64): Likewise.
4927 (vrndm_f64): Likewise.
4928 (vrndn_f64): Likewise.
4929 (vrndp_f64): Likewise.
4930 (vrndx_f64): Likewise.
4932 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4934 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4935 GET_MODE_SIZE argument is enum machine_mode.
4937 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4940 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4941 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4943 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4946 * asan.c (asan_emit_stack_protection): Force the base to align to
4947 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4948 appropriate bits if STRICT_ALIGNMENT.
4949 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4951 (expand_used_vars): Leave a space in the stack frame for alignment
4952 if STRICT_ALIGNMENT.
4954 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4956 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4958 (gimple_store_p): Likewise.
4959 (gimple_assign_load_p): Likewise.
4960 (gimple_assign_cast_p): Likewise.
4961 (gimple_clobber_p): Likewise.
4963 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4964 rather than a gimple.
4965 (gimple_assign_cast_p): Likewise.
4967 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4970 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4971 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4973 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4974 more debug information for E500 if -mdebug=reg.
4976 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4979 * config/i386/i386.c (ix86_expand_builtin)
4980 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4981 register for target RTX.
4982 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4984 2014-04-18 Cong Hou <congh@google.com>
4986 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4987 the widen-mult pattern by handling two operands with different sizes,
4988 and operands whose size is smaller than half of the result type.
4990 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4992 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4993 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4994 (do_estimate_edge_time): Compute it.
4995 * ipa-inline.c (want_inline_small_function_p): Bypass
4996 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4998 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5000 * ipa-inline.c (spec_rem): New static variable.
5001 (dump_overall_stats): New function.
5002 (dump_inline_stats): New function.
5004 2014-04-18 Richard Henderson <rth@redhat.com>
5006 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5007 to GET_MODE_SIZE, not a reg_class_t.
5009 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5011 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5012 (vsx_xxmrglw_<mode>): Likewise.
5014 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5017 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5018 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5019 (rs6000_init_hard_regno_mode_ok): Likewise.
5021 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5023 * ipa-inline.c (inline_small_functions): Account only non-cold
5025 * doc/invoke.texi (inline-unit-growth): Update documentation.
5027 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5029 * config/rs6000/rs6000.md (addti3, subti3): New.
5031 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5034 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5035 comment. Check optimize_insn_for_size_p instead of
5036 optimize_insn_for_speed_p.
5038 2014-04-17 Martin Jambor <mjambor@suse.cz>
5040 * gimple-iterator.c (gsi_start_edge): New function.
5041 * gimple-iterator.h (gsi_start_edge): Declare.
5042 * tree-sra.c (single_non_eh_succ): New function.
5043 (disqualify_ops_if_throwing_stmt): Renamed to
5044 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5045 having one non-EH successor BB.
5046 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5047 generate loads into replacements.
5048 (sra_modify_assign): Likewise and and also use the simple path for
5050 (sra_modify_function_body): Commit statements on edges.
5052 2014-04-17 Richard Biener <rguenther@suse.de>
5055 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5056 comparison results and add clarifying comment.
5058 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5060 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5061 (blank_mode): Initialize it.
5062 (emit_mode_size_inline, emit_mode_nunits_inline,
5063 emit_mode_inner_inline): New functions.
5064 (emit_insn_modes_h): Call them and surround their output with
5065 #if GCC_VERSION >= 4001 ... #endif.
5066 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5067 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5068 mode_* arrays if the argument is __builtin_constant_p.
5069 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5070 is enum machine_mode.
5072 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5074 * passes.c (opt_pass::execute): Adjust.
5075 (pass_manager::execute_pass_mode_switching): Likewise.
5076 (early_local_passes::execute): Likewise.
5077 (execute_one_pass): Pass cfun to the pass's execute method.
5078 * tree-pass.h (opt_pass::execute): Add function * argument.
5079 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5080 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5081 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5082 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5083 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5084 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5085 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5086 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5087 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5088 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5089 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5090 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5091 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5092 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5093 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5094 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5095 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5096 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5097 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5098 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5099 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5100 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5101 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5102 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5103 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5104 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5105 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5106 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5107 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5110 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5112 * passes.c (opt_pass::gate): Take function * argument.
5113 (gate_all_early_local_passes): Merge into
5114 (early_local_passes::gate): this.
5115 (gate_all_early_optimizations): Merge into
5116 (all_early_optimizations::gate): this.
5117 (gate_all_optimizations): Mege into
5118 (all_optimizations::gate): this.
5119 (gate_all_optimizations_g): Merge into
5120 (all_optimizations_g::gate): this.
5121 (gate_rest_of_compilation): Mege into
5122 (rest_of_compilation::gate): this.
5123 (gate_postreload): Merge into
5124 (postreload::gate): this.
5125 (dump_one_pass): Pass cfun to the pass's gate method.
5126 (execute_ipa_summary_passes): Likewise.
5127 (execute_one_pass): Likewise.
5128 (ipa_write_summaries_2): Likewise.
5129 (ipa_write_optimization_summaries_1): Likewise.
5130 (ipa_read_summaries_1): Likewise.
5131 (ipa_read_optimization_summaries_1): Likewise.
5132 (execute_ipa_stmt_fixups): Likewise.
5133 * tree-pass.h (opt_pass::gate): Add function * argument.
5134 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5135 combine-stack-adj.c, combine.c, compare-elim.c,
5136 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5137 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5138 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5139 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5140 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5141 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5142 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5143 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5144 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5145 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5146 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5147 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5148 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5149 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5150 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5151 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5152 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5153 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5154 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5155 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5156 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5157 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5158 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5159 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5160 var-tracking.c, vtable-verify.c, web.c: Adjust.
5162 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5164 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5165 * configure: Regenerate.
5167 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5169 * passes.c (dump_one_pass): don't check pass->has_gate.
5170 (execute_ipa_summary_passes): Likewise.
5171 (execute_one_pass): Likewise.
5172 (ipa_write_summaries_2): Likewise.
5173 (ipa_write_optimization_summaries_1): Likewise.
5174 (ipa_read_optimization_summaries_1): Likewise.
5175 (execute_ipa_stmt_fixups): Likewise.
5176 * tree-pass.h (pass_data::has_gate): Remove.
5177 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5178 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5179 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5180 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5181 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5182 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5183 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5184 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5185 gimple-low.c, gimple-ssa-isolate-paths.c,
5186 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5187 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5188 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5189 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5190 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5191 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5192 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5193 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5194 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5195 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5196 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5197 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5198 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5199 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5200 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5201 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5202 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5203 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5204 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5205 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5206 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5207 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5210 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5212 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5214 * passes.c (pass_manager::register_dump_files_1): Merge into
5215 (pass_manager::register_dump_files): this, and remove its handling of
5216 properties since the pass always has the properties anyway.
5217 (pass_manager::pass_manager): Adjust.
5219 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5221 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5222 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5223 dealing with properties.
5224 (pass_manager::register_dump_files): Adjust.
5226 2014-03-20 Mark Wielaard <mjw@redhat.com>
5228 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5229 then represent the bound as normal constant value.
5231 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5234 Forward port from 4.8 branch
5235 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5237 * config/i386/bmiintrin.h (_blsi_u32): New.
5241 (_blsmsk_u32): Ditto.
5242 (_blsmsk_u64): Ditto.
5243 (_tzcnt_u32): Ditto.
5244 (_tzcnt_u64): Ditto.
5246 2014-04-17 Kito Cheng <kito@0xlab.org>
5248 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5250 2014-04-17 Richard Biener <rguenther@suse.de>
5253 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5254 boolean results for comparisons.
5256 2014-04-17 Richard Biener <rguenther@suse.de>
5258 PR tree-optimization/60836
5259 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5260 initial PHI args to be gimple values.
5262 2014-04-17 Richard Biener <rguenther@suse.de>
5264 PR tree-optimization/60841
5265 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5266 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5267 of stmts to SLP build.
5268 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5269 (vect_analyze_slp): Likewise.
5270 (vect_analyze_slp_instance): Likewise.
5271 (vect_build_slp_tree): Limit overall SLP tree growth.
5272 * tree-vectorizer.h (vect_analyze_data_refs,
5273 vect_analyze_slp): Adjust prototypes.
5275 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5277 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5280 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5282 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5283 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5284 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5285 for TARGET_SLOW_PSHUFB
5287 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5289 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5290 * config/i386/i386.c (intel_cost): Ditto.
5292 2014-04-17 Joey Ye <joey.ye@arm.com>
5294 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5296 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5298 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5301 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5303 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5304 (type_all_derivations_known_p): New predicate.
5305 (type_all_ctors_visible_p): New predicate.
5306 (type_possibly_instantiated_p): New predicate.
5307 (get_odr_type): Compute all_derivations_known.
5308 (dump_odr_type): Dump the flag.
5309 (maybe_record_type): Cleanup.
5310 (record_target_from_binfo): Add bases_to_consider array;
5311 record bases for types w/o instances and skip CXX destructor.
5312 (possible_polymorphic_call_targets_1): Add bases_to_consider
5313 and consider_construction parameters; check if type may have instance.
5314 (get_polymorphic_call_info): Set maybe_in_construction to true
5315 when we know nothing.
5316 (record_targets_from_bases): Skip CXX destructors; they are
5317 never called for types in construction.
5318 (possible_polymorphic_call_targets): Do not record target when
5319 type may not have instance.
5321 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5324 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5325 external aliases alive, too.
5327 2014-04-16 Andrew Pinski <apinski@cavium.com>
5329 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5332 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5334 * final.c (compute_alignments): Do not apply loop alignment to a block
5335 falling through to the exit.
5337 2014-04-16 Catherine Moore <clm@codesourcery.com>
5339 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5340 Adjust constraints for microMIPS store patterns.
5342 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5344 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5346 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5348 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5349 (append_use): Run at -O0.
5350 (append_vdef): Likewise.
5351 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5352 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5354 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5356 PR tree-optimization/60844
5357 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5358 (propagate_op_to_single_use, remove_visited_stmt_chain,
5359 linearize_expr, repropagate_negates, reassociate_bb): Use it
5360 instead of gsi_remove.
5362 2014-04-16 Martin Jambor <mjambor@suse.cz>
5364 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5365 ipa_transforms_to_apply.
5366 (cgraph_function_versioning): Assert that old_node has empty
5367 ipa_transforms_to_apply.
5368 * trans-mem.c (ipa_tm_create_version): Likewise.
5369 * tree-inline.c (tree_function_versioning): Do not duplicate
5370 ipa_transforms_to_apply.
5372 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5375 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5377 Pass necessary as flags on 64-bit Solaris/x86.
5378 Use lowercase relocs for x86_64-*-*.
5379 * configure: Regenerate.
5381 2014-04-15 Jan Hubicka <jh@suse.cz>
5383 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5384 (maybe_record_node, likely_target_p): Use it.
5386 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5389 Revert following patch
5391 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5394 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5395 software floating point or no floating point registers, do not
5396 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5397 in GPRs that occurs after we tested for GPRs that would never be
5400 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5401 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5402 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5403 specifically allow DDmode, since that does not use the SPE SIMD
5406 2014-03-21 Mark Wielaard <mjw@redhat.com>
5408 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5409 as unsigned or int depending on type and value used.
5411 2014-04-15 Richard Biener <rguenther@suse.de>
5413 PR rtl-optimization/56965
5414 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5415 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5417 * alias.c (true_dependence_1): Do not call
5418 nonoverlapping_component_refs_p.
5419 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5420 nonoverlapping_component_refs_p.
5421 (indirect_refs_may_alias_p): Likewise.
5423 2014-04-15 Teresa Johnson <tejohnson@google.com>
5425 * cfg.c (dump_bb_info): Fix flags check.
5426 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5428 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5430 PR rtl-optimization/60663
5431 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5434 2014-04-15 Richard Biener <rguenther@suse.de>
5436 * lto-streamer.h (LTO_major_version): Bump to 4.
5438 2014-04-15 Richard Biener <rguenther@suse.de>
5440 * common.opt (lto_partition_model): New enum.
5441 (flto-partition=): Merge separate options with a single with argument,
5442 add -flto-partition=one support.
5443 * flag-types.h (enum lto_partition_model): Declare.
5444 * opts.c (finish_options): Remove duplicate -flto-partition=
5446 * lto-wrapper.c (run_gcc): Adjust.
5448 2014-04-15 Richard Biener <rguenther@suse.de>
5450 * alias.c (ncr_compar): New function.
5451 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5453 2014-04-15 Richard Biener <rguenther@suse.de>
5455 * alias.c (record_component_aliases): Do not walk BINFOs.
5457 2014-04-15 Richard Biener <rguenther@suse.de>
5459 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5460 Add struct function argument and adjust.
5461 (find_func_aliases_for_call): Likewise.
5462 (find_func_aliases): Likewise.
5463 (find_func_clobbers): Likewise.
5464 (intra_create_variable_infos): Likewise.
5465 (compute_points_to_sets): Likewise.
5466 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5468 2014-04-15 Richard Biener <rguenther@suse.de>
5470 * tree.c (iterative_hash_expr): Use enum tree_code_class
5471 to store TREE_CODE_CLASS.
5472 (tree_block): Likewise.
5473 (tree_set_block): Likewise.
5474 * tree.h (fold_build_pointer_plus_loc): Use
5475 convert_to_ptrofftype_loc.
5477 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5480 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5483 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5485 * cfgloop.h (struct loop): Move force_vectorize down.
5486 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5487 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5488 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5489 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5490 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5491 * tree-core.h (enum annot_expr_kind): Add new kind values.
5492 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5493 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5495 * tree.def (ANNOTATE_EXPR): Tweak comment.
5497 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5499 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5502 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5504 * tree.h (TYPE_IDENTIFIER): Declare.
5505 * tree.c (subrange_type_for_debug_p): Use it.
5506 * godump.c (go_format_type): Likewise.
5507 * dwarf2out.c (is_cxx_auto, modified_type_die,
5508 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5509 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5511 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5514 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5516 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5518 * config/i386/i386.c (examine_argument): Return bool. Return true if
5519 parameter should be passed in memory.
5520 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5521 (construct_container): Update calls to examine_argument.
5522 (function_arg_advance_64): Ditto.
5523 (return_in_memory_32): Merge with ix86_return_in_memory.
5524 (return_in_memory_64): Ditto.
5525 (return_in_memory_ms_64): Ditto.
5527 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5529 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5530 * coverage.c (coverage_compute_profile_id): Handle externally visible
5533 2014-04-14 Martin Jambor <mjambor@suse.cz>
5535 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5536 DECL_DISREGARD_INLINE_LIMITS functions.
5538 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5541 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5543 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5546 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5547 optimize_insn_for_speed_p instead of
5548 optimize_function_for_speed_p.
5550 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5552 * doc/invoke.texi (free): Document AArch64.
5554 2014-04-14 Richard Biener <rguenther@suse.de>
5556 PR tree-optimization/60042
5557 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5558 (insert_into_preds_of_block): Do not prevent PHI insertion
5559 for REFERENCE exprs here ...
5560 (eliminate_dom_walker::before_dom_children): ... but prevent
5561 their use here under similar conditions when applied to the
5562 IL after PRE optimizations.
5564 2014-04-14 Richard Biener <rguenther@suse.de>
5566 * passes.def: Move early points-to after early SRA.
5568 2014-04-14 Richard Biener <rguenther@suse.de>
5570 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5571 check for which sign-changes we allow when forwarding
5572 a converted value into a switch.
5574 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5576 * stor-layout.c (place_field): Finalize non-constant offset for the
5579 2014-04-14 Richard Biener <rguenther@suse.de>
5581 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5583 (expand_switch_using_bit_tests_p): Likewise.
5584 (process_switch): Compute and pass on speed_p based on the
5586 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5587 optimize_bb_for_speed_p.
5589 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5591 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5592 * function.h (struct function): Rename has_force_vect_loops into
5593 has_force_vectorize_loops.
5594 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5595 (input_struct_function_base): Likewise.
5596 * lto-streamer-out.c (output_cfg): Likewise.
5597 (output_struct_function_base): Likewise.
5598 * omp-low.c (expand_omp_simd): Likewise.
5599 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5600 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5601 (version_loop_for_if_conversion): Likewise.
5602 (tree_if_conversion): Likewise.
5603 (main_tree_if_conversion): Likewise.
5604 (gate_tree_if_conversion): Likewise.
5605 * tree-inline.c (copy_loops): Likewise.
5606 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5607 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5608 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5609 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5610 * tree-vectorizer.c (vectorize_loops): Likewise.
5611 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5613 2014-04-14 Richard Biener <rguenther@suse.de>
5616 * lto-streamer-out.c (wrap_refs): New function.
5617 (lto_output): Wrap symbol references in global initializes in
5618 type-preserving MEM_REFs.
5620 2014-04-14 Christian Bruel <christian.bruel@st.com>
5622 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5624 2014-04-14 Christian Bruel <christian.bruel@st.com>
5626 * config/sh/sh.md (setmemqi): New expand pattern.
5627 * config/sh/sh.h (CLEAR_RATIO): Define.
5628 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5629 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5631 2014-04-14 Richard Biener <rguenther@suse.de>
5634 * fold-const.c (negate_expr_p): Don't negate directional rounding
5636 (fold_negate_expr): Likewise.
5638 2014-04-14 Richard Biener <rguenther@suse.de>
5640 PR tree-optimization/59817
5641 PR tree-optimization/60453
5642 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5643 recursion to catch all CHRECs in the scalar evolution and restrict
5644 the predicate for the remains appropriately.
5646 2014-04-12 Catherine Moore <clm@codesourcery.com>
5648 * config/mips/constraints.md: Add new register constraint "kb".
5649 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5650 (*movhi_internal): Likewise.
5651 (*movqi_internal): Likewise.
5652 * config/mips/mips.h (M16_STORE_REGS): New register class.
5653 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5654 (REG_CLASS_CONTENTS): Likewise.
5655 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5657 2014-04-11 Tobias Burnus <burnus@net-b.de>
5660 * doc/invoke.texi (-Wformat-signedness): Document it.
5661 (Wformat=2): Mention that this enables -Wformat-signedness.
5663 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5665 * common/config/epiphany/epiphany-common.c
5666 (epiphany_option_optimization_table): Enable section anchors by
5667 default at -O1 or higher.
5668 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5669 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5670 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5671 carries no extra cost.
5672 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5673 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5674 * config/epiphany/predicates.md (memclob_operand): New predicate.
5675 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5676 Use memclob_operand predicate and X constraint for operand 3.
5678 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5680 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5681 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5684 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5686 PR rtl-optimization/60651
5687 * mode-switching.c (optimize_mode_switching): Make sure to emit
5688 sets of a lower numbered entity before sets of a higher numbered
5689 entity to a mode of the same or lower priority.
5690 When creating a seginfo for a basic block that starts with a code
5691 label, move the insertion point past the code label.
5692 (new_seginfo): Document and enforce requirement that
5693 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5694 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5695 * doc/tm.texi: Regenerate.
5697 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5700 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5702 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5704 * BASE-VER: Set to 4.10.0.
5706 2014-04-11 Tobias Burnus <burnus@net-b.de>
5709 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5710 * doc/gcc.texi (Service): Update description in the @menu
5711 * doc/invoke.texi (Option Summary): Remove misplaced and
5714 2014-04-11 Steve Ellcey <sellcey@mips.com>
5715 Jakub Jelinek <jakub@redhat.com>
5718 * expr.c (convert_move): Use emit_store_flag_force instead of
5719 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5722 2014-04-11 Richard Biener <rguenther@suse.de>
5725 * varasm.c (assemble_alias): Avoid endless error reporting
5726 recursion by setting TREE_ASM_WRITTEN.
5728 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5730 * config/s390/s390.md: Add a splitter for NOT rtx.
5732 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5734 PR rtl-optimization/60663
5735 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5737 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5738 Jakub Jelinek <jakub@redhat.com>
5741 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5742 flag from decl_node to node.
5744 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5747 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5748 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5749 ameliorating the cases where it can be.
5751 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5754 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5756 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5757 (loadsync_<mode>): Change mode.
5758 (load_quadpti, store_quadpti): New.
5759 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5760 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5761 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5763 2014-04-09 Cong Hou <congh@google.com>
5766 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5769 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5771 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5772 instead of vnor to exploit possible fusion opportunity in the
5774 (altivec_expand_vec_perm_const_le): Likewise.
5776 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5778 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5779 (loadsync_<mode>): Change mode.
5780 (load_quadpti, store_quadpti): New.
5781 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5782 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5784 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5787 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5788 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5789 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5791 2014-04-08 Richard Biener <rguenther@suse.de>
5794 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5795 a 64bit widest int print double-int similar to on HWI64 hosts.
5797 2014-04-08 Richard Biener <rguenther@suse.de>
5799 PR tree-optimization/60785
5800 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5801 default defs properly.
5803 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5805 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5806 (Weffc++): Likewise.
5808 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5810 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5811 set completep to false rather than true.
5813 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5816 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5817 ARM_TARGET2_DWARF_FORMAT.
5819 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5822 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5823 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5826 2014-04-07 Richard Biener <rguenther@suse.de>
5828 PR tree-optimization/60766
5829 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5830 (may_eliminate_iv): Convert cand_value_at result to desired type.
5832 2014-04-07 Jason Merrill <jason@redhat.com>
5835 * common.opt (-fno-gnu-unique): Add.
5836 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5838 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5840 * haifa-sched.c: Fix outdated function reference and minor
5841 grammar errors in introductory comment.
5843 2014-04-07 Richard Biener <rguenther@suse.de>
5846 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5848 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5850 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5853 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5854 size accounting for thunks.
5855 (pa_asm_output_mi_thunk): Use final_start_function() and
5856 final_end_function() to output function start and end directives.
5858 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5860 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5861 device specific ISA/ feature information. Remove short_sp and
5862 errata_skip ds. Add avr_device_specific_features enum to have device
5864 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5865 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5866 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5867 updated device specific info.
5868 * config/avr/avr-mcus.def: Merge device specific details to
5869 dev_attribute field.
5870 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5872 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5873 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5874 assembler if RMW isa supported by current device.
5875 * config/avr/genmultilib.awk: Update as device info structure changed.
5876 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5878 2014-04-04 Cong Hou <congh@google.com>
5880 PR tree-optimization/60656
5881 * tree-vect-stmts.c (supportable_widening_operation):
5882 Fix a bug that elements in a vector with vect_used_by_reduction
5883 property are incorrectly reordered when the operation on it is not
5884 consistant with the one in reduction operation.
5886 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5888 PR rtl-optimization/60155
5889 * gcse.c (record_set_data): New function.
5890 (single_set_gcse): New function.
5891 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5892 (hoist_code): Likewise.
5893 (get_pressure_class_and_nregs): Likewise.
5895 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5897 * explow.c (probe_stack_range): Emit a final optimization blockage.
5899 2014-04-04 Anthony Green <green@moxielogic.com>
5901 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5904 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5907 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5908 flags are set only during streaming.
5909 * ipa.c (process_references, walk_polymorphic_call_targets,
5910 symtab_remove_unreachable_nodes): Drop bodies of always inline
5911 after early inlining.
5912 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5914 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5915 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5918 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5921 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5924 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5926 (cortex_a53_fdivd): Likewise.
5928 2014-04-04 Martin Jambor <mjambor@suse.cz>
5931 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5933 * cgraph.c (clone_of_p): Also return true if thunks match.
5934 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5935 cgraph_function_or_thunk_node and an obsolete comment.
5936 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5938 (build_function_decl_skip_args): Likewise.
5939 (set_new_clone_decl_and_node_flags): New function.
5940 (duplicate_thunk_for_node): Likewise.
5941 (redirect_edge_duplicating_thunks): Likewise.
5942 (cgraph_clone_node): New parameter args_to_skip, pass it to
5943 redirect_edge_duplicating_thunks which is called instead of
5944 cgraph_redirect_edge_callee.
5945 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5946 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5948 2014-04-04 Jeff Law <law@redhat.com>
5951 * config/arm/predicates.md (const_int_I_operand): New predicate.
5952 (const_int_M_operand): Similarly.
5953 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5955 (insv_t2, extv_reg, extzv_t2): Likewise.
5956 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5957 (pop_multiple_with_writeback_and_return): Likewise.
5958 (vfp_pop_multiple_with_writeback): Likewise
5960 2014-04-04 Richard Biener <rguenther@suse.de>
5963 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5964 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5966 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5967 * gimplify.c (gimple_add_tmp_var_fn): New function.
5968 * gimple-expr.h (create_tmp_reg_fn): Declare.
5969 * gimple-expr.c (create_tmp_reg_fn): New function.
5970 * gimple-low.c (record_vars_into): Don't change cfun.
5971 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5972 code generation without cfun.
5974 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5977 * Makefile.in (install-driver): Fix shell scripting.
5979 2014-04-03 Cong Hou <congh@google.com>
5981 PR tree-optimization/60505
5982 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5983 threshold of number of iterations below which no vectorization
5985 * tree-vect-loop.c (new_loop_vec_info):
5986 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5987 * tree-vect-loop.c (vect_analyze_loop_operations):
5988 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5989 * tree-vect-loop.c (vect_transform_loop):
5990 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5991 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5992 of iterations of the loop and see if we should build the epilogue.
5994 2014-04-03 Richard Biener <rguenther@suse.de>
5996 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5997 (streamer_tree_cache_create): Adjust.
5998 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5999 to allow optional nodes array.
6000 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6001 (streamer_tree_cache_append): Likewise.
6002 (streamer_tree_cache_create): Create nodes array optionally
6003 as specified by parameter.
6004 * lto-streamer-out.c (create_output_block): Avoid maintaining
6005 the node array in the writer cache.
6006 (DFS_write_tree): Remove assertion.
6007 (produce_asm_for_decls): Free the out decl state hash table early.
6008 * lto-streamer-in.c (lto_data_in_create): Adjust for
6009 streamer_tree_cache_create prototype change.
6011 2014-04-03 Richard Biener <rguenther@suse.de>
6013 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6014 set TREE_CHAIN to NULL_TREE.
6016 2014-04-03 Richard Biener <rguenther@suse.de>
6018 PR tree-optimization/60740
6019 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6020 over all GIMPLE_COND operands.
6022 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6024 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6025 (Weffc++): Remove Scott's numbering, merge lists and reference
6028 2014-04-03 Nick Clifton <nickc@redhat.com>
6030 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6033 2014-04-03 Martin Jambor <mjambor@suse.cz>
6035 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6036 mention gcc_unreachable before failing.
6037 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6040 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6043 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6044 inconsistent code and instead mark the context inconsistent.
6045 (possible_polymorphic_call_targets): For inconsistent contexts
6046 return empty complete list.
6048 2014-04-02 Anthony Green <green@moxielogic.com>
6050 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6051 (extendqisi2, extendhisi2): Define.
6052 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6053 (WCHAR_TYPE): Change to unsigned int.
6055 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6057 PR tree-optimization/60733
6058 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6059 insertion point for PHI candidates to be the end of the feeding
6060 block for the PHI argument.
6062 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6064 PR rtl-optimization/60650
6065 * lra-constraints.c (process_alt_operands): Decrease reject for
6066 earlyclobber matching.
6068 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6070 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6072 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6074 * config/spu/spu.c (pad_bb): Do not crash when the last
6075 insn is CODE_FOR_blockage.
6077 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6079 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6080 lies outside the target mode.
6082 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6085 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6086 software floating point or no floating point registers, do not
6087 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6088 in GPRs that occurs after we tested for GPRs that would never be
6091 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6092 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6093 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6094 specifically allow DDmode, since that does not use the SPE SIMD
6097 2014-04-02 Richard Biener <rguenther@suse.de>
6100 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6101 MODE_INTs. Properly use negv_optab.
6102 (expand_abs): Likewise.
6104 2014-04-02 Richard Biener <rguenther@suse.de>
6107 * Makefile.in (install-driver): Guard extra installs with special
6110 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6112 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6115 2014-04-01 Richard Henderson <rth@redhat.com>
6118 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6119 alternative enabled before register allocation.
6121 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6123 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6124 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6126 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6127 (nios2_got_address): Update nios2_large_got_address call site.
6128 (nios2_delegitimize_address): New function.
6129 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6130 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6131 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6133 2014-04-01 Martin Husemann <martin@duskware.de>
6135 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6138 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6140 PR rtl-optimization/60604
6141 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6142 check from register_operand.
6143 (register_operand): Redefine in terms of general_operand.
6144 (nonmemory_operand): Use register_operand for the non-constant cases.
6146 2014-04-01 Richard Biener <rguenther@suse.de>
6148 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6150 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6152 * doc/invoke.texi (mapp-regs): Clarify.
6154 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6156 * config/i386/avx512fintrin.h (__v32hi): Define type.
6157 (__v64qi): Likewise.
6158 (_mm512_set1_epi8): Define.
6159 (_mm512_set1_epi16): Define.
6160 (_mm512_set4_epi32): Define.
6161 (_mm512_set4_epi64): Define.
6162 (_mm512_set4_pd): Define.
6163 (_mm512_set4_ps): Define.
6164 (_mm512_setr4_epi64): Define.
6165 (_mm512_setr4_epi32): Define.
6166 (_mm512_setr4_pd): Define.
6167 (_mm512_setr4_ps): Define.
6168 (_mm512_setzero_epi32): Define.
6170 2014-03-31 Martin Jambor <mjambor@suse.cz>
6173 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6174 callsite_arguments_match_p. Updated all callers. Also check types of
6175 corresponding formal parameters and actual arguments.
6176 (not_all_callers_have_enough_arguments_p) Renamed to
6177 some_callers_have_mismatched_arguments_p.
6179 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6181 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6183 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6186 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6189 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6191 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6192 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6194 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6195 Use FMAMODE_NOVF512 mode iterator.
6196 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6197 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6198 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6199 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6201 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6202 Use VF_128_256 mode iterator.
6203 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6206 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6208 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6209 static chain if needed.
6211 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6214 * lra-constraints.c (index_part_to_reg): New.
6215 (process_address): Use it.
6217 2014-03-27 Jeff Law <law@redhat.com>
6218 Jakub Jelinek <jakub@redhat.com>
6221 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6222 gen_rtx_{PLUS,MULT} to build up the address expression.
6224 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6225 creating non-canonical RTL.
6227 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6230 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6231 functions; reorganize to make cheap checks first.
6232 (inline_small_functions): Do not estimate growth when dumping;
6234 * ipa-inline.h (inline_summary): Add min_size.
6235 (growth_likely_positive): New function.
6236 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6237 (set_cond_stmt_execution_predicate): Cleanup.
6238 (estimate_edge_size_and_time): Compute min_size.
6239 (estimate_calls_size_and_time): Likewise.
6240 (estimate_node_size_and_time): Likewise.
6241 (inline_update_overall_summary): Update min_size.
6242 (do_estimate_edge_time): Likewise.
6243 (do_estimate_edge_size): Update.
6244 (do_estimate_edge_hints): Update.
6245 (growth_likely_positive): New function.
6247 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6250 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6251 also if addr has VOIDmode.
6253 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6255 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6256 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6258 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6259 instructions as well as AdvancedSIMD loads.
6261 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6263 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6264 Use crypto_aese type.
6265 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6266 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6267 crypto_aese, crypto_aesmc. Move to types.md.
6268 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6270 * config/arm/iterators.md (crypto_type): Likewise.
6272 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6274 * cgraph.c: Include expr.h and tree-dfa.h.
6275 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6278 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6281 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6282 regs from checking multi-reg pseudos.
6284 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6286 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6288 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6290 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6291 if it would clobber the stack pointer, even temporarily.
6293 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6295 * mode-switching.c: Make small adjustments to the top comment.
6297 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6299 * config/rs6000/constraints.md (wD constraint): New constraint to
6300 match the constant integer to get the top DImode/DFmode out of a
6301 vector in a VSX register.
6303 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6304 match the constant integer to get the top DImode/DFmode out of a
6305 vector in a VSX register.
6307 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6310 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6313 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6314 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6316 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6317 Optimize vec_extract of 64-bit values, where the value being
6318 extracted is in the top word, where we can use scalar
6319 instructions. Add direct move and store support. Combine the big
6320 endian/little endian vector select load support into a single insn.
6321 (vsx_extract_<mode>_internal1): Likewise.
6322 (vsx_extract_<mode>_internal2): Likewise.
6323 (vsx_extract_<mode>_load): Likewise.
6324 (vsx_extract_<mode>_store): Likewise.
6325 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6326 combined into vsx_extract_<mode>_load.
6327 (vsx_extract_<mode>_one_le): Likewise.
6329 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6330 define the top 64-bit vector element.
6332 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6335 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6336 Document vec_vbpermq builtin.
6339 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6340 enable use of xxsldwi and xxpermdi builtin functions.
6341 (vec_xxpermdi): Likewise.
6343 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6344 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6346 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6348 PR rtl-optimization/60650
6349 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6351 (find_spills_for): New.
6352 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6353 Spill all pseudos on the second iteration.
6355 2014-03-27 Marek Polacek <polacek@redhat.com>
6358 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6361 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6363 * config/s390/s390.c (s390_can_use_return_insn): Check for
6364 call-saved FPRs on 31 bit.
6366 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6369 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6370 if they need regimplification, just drop them instead of
6371 calling gimple_regimplify_operands on them.
6373 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6376 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6377 (aarch64_frame_pointer_required): Adjust logic.
6378 (aarch64_can_eliminate): Adjust logic.
6379 (aarch64_override_options_after_change): Adjust logic.
6381 2014-03-27 Dehao Chen <dehao@google.com>
6383 * ipa-inline.c (early_inliner): Update node's inline info.
6385 2014-03-26 Dehao Chen <dehao@google.com>
6387 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6388 compiler inserted conditional jumps for NAN float check.
6390 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6392 * ubsan.h (ubsan_create_data): Change second argument's type
6393 to const location_t *.
6394 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6396 (ubsan_create_data): Change second argument to const location_t *PLOC.
6397 Create Loc field whenever PLOC is non-NULL.
6398 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6399 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6403 * real.c (real_to_integer2): Change type of low to UHWI.
6405 2014-03-26 Tobias Burnus <burnus@net-b.de>
6407 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6408 (CILK_SELF_SPECS): New define.
6409 (driver_self_specs): Use it.
6411 2014-03-26 Richard Biener <rguenther@suse.de>
6413 * tree-pretty-print.c (percent_K_format): Implement special
6414 case for LTO and its stripped down BLOCK tree.
6416 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6419 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6421 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6422 one range is range_int_cst_p, but not both, at least optimize
6423 addition/subtraction of 0 and multiplication by 0 or 1.
6424 * gimple-fold.c (gimple_fold_call): Fold
6425 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6426 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6427 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6429 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6431 PR rtl-optimization/60452
6432 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6433 <case REG>: Return 1 for invalid offsets from the frame pointer.
6435 2014-03-26 Marek Polacek <polacek@redhat.com>
6438 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6441 2014-03-26 Marek Polacek <polacek@redhat.com>
6444 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6447 2014-03-26 Marek Polacek <polacek@redhat.com>
6450 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6451 multiplication in unsigned type.
6453 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6455 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6457 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6459 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6461 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6464 * cif-code.def (UNREACHABLE) New code.
6465 * ipa-inline.c (inline_small_functions): Skip edges to
6466 __builtlin_unreachable.
6467 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6468 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6469 predicate to __bulitin_unreachable.
6470 (set_cond_stmt_execution_predicate): Fix issue when
6471 invert_tree_comparison returns ERROR_MARK.
6472 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6473 propagate to inline clones.
6474 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6476 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6477 * cgraphclones.c (cgraph_clone_node): If call destination is already
6478 ureachable, do not redirect it back.
6479 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6482 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6484 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6485 Do not modify inline clones.
6487 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6489 * config/i386/i386.md (general_sext_operand): New mode attr.
6490 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6491 don't generate (sign_extend (const_int)).
6492 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6493 operands[2]. Use We constraint instead of <i> and
6494 <general_sext_operand> predicate instead of <general_operand>.
6495 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6496 * config/i386/constraints.md (We): New constraint.
6497 * config/i386/predicates.md (x86_64_sext_operand,
6498 sext_operand): New predicates.
6500 2014-03-25 Martin Jambor <mjambor@suse.cz>
6503 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6504 inconsistent devirtualizations to __builtin_unreachable.
6506 2014-03-25 Marek Polacek <polacek@redhat.com>
6509 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6511 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6513 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6514 order of elements for big-endian.
6516 2014-03-25 Richard Biener <rguenther@suse.de>
6519 * gimplify-me.c (gimple_regimplify_operands): Update the
6522 2014-03-25 Martin Jambor <mjambor@suse.cz>
6525 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6526 (lto_output_varpool_node): Likewise.
6527 (input_overwrite_node): Likewise.
6528 (input_varpool_node): Likewise.
6530 2014-03-25 Richard Biener <rguenther@suse.de>
6532 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6533 (run_gcc): Likewise.
6535 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6537 * combine.c (simplify_compare_const): Add MODE argument.
6538 Handle mode_width 0 as very large mode_width.
6539 (try_combine, simplify_comparison): Adjust callers.
6541 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6542 type to avoid signed integer overflow.
6543 * explow.c (plus_constant): Likewise.
6545 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6547 * doc/generic.texi: Correct typos.
6549 2014-03-24 Tobias Burnus <burnus@net-b.de>
6551 * doc/invoke.texi (-flto): Expand section about
6552 using static libraries with LTO.
6554 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6556 PR rtl-optimization/60501
6557 * optabs.def (addptr3_optab): New optab.
6558 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6559 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6560 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6562 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6564 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6566 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6568 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6571 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6572 (_mm256_undefined_ps): Define.
6573 (_mm256_undefined_pd): Define.
6574 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6575 (_mm_undefined_pd): Define.
6576 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6577 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6578 (_mm512_undefined_ps): Define.
6579 (_mm512_undefined_pd): Define.
6580 Use _mm*_undefined_*.
6581 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6583 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6585 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6586 (lshr_simd): DI mode added.
6587 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6588 (aarch64_ushr_simddi): Likewise.
6589 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6590 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6591 (vshrd_n_u64): Likewise.
6593 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6595 * Makefile.in (s-macro_list): Depend on cc1.
6597 2014-03-23 Teresa Johnson <tejohnson@google.com>
6599 * ipa-utils.c (ipa_print_order): Use specified dump file.
6601 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6603 PR rtl-optimization/60601
6604 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6606 * gcc.c (eval_spec_function): Initialize save_growing_value.
6608 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6611 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6612 code == MINUS_EXPR, never swap op0 with op1.
6614 * toplev.c (init_local_tick): Avoid signed integer multiplication
6616 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6617 shift by first operand's bitsize.
6619 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6622 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6624 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6625 TARGET_ISA_64BIT_P(x).
6627 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6629 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6630 pattern for vector nor instead of subtract from splat(-1).
6631 (altivec_expand_vec_perm_const_le): Likewise.
6633 2014-03-21 Richard Henderson <rth@twiddle.net>
6636 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6637 related insns after epilogue_completed.
6639 2014-03-21 Martin Jambor <mjambor@suse.cz>
6642 * cgraph.h (symtab_node): New flag body_removed.
6643 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6644 when removing bodies.
6645 * symtab.c (dump_symtab_base): Dump body_removed flag.
6646 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6647 had their bodies removed.
6649 2014-03-21 Martin Jambor <mjambor@suse.cz>
6652 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6655 2014-03-21 Richard Biener <rguenther@suse.de>
6657 PR tree-optimization/60577
6658 * tree-core.h (struct tree_base): Document nothrow_flag use
6659 in VAR_DECL_NONALIASED.
6660 * tree.h (VAR_DECL_NONALIASED): New.
6661 (may_be_aliased): Adjust.
6662 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6664 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6666 * expr.c (expand_expr_real_1): Remove outdated comment.
6668 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6671 * ira.c (adjust_cleared_regs): Call copy_rtx on
6672 *reg_equiv[REGNO (loc)].src_p before passing it to
6673 simplify_replace_fn_rtx.
6676 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6677 into CONST, put pic register as first operand of PLUS. Use
6678 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6680 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6682 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6684 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6686 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6687 around for store forwarding issue in the FPU on the UT699.
6688 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6689 loads and operations if -mfix-ut699 is specified.
6690 (divtf3_hq): Tweak attribute.
6691 (sqrttf2_hq): Likewise.
6693 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6695 * calls.c (store_one_arg): Remove incorrect const qualification on the
6696 type of the temporary.
6697 * cfgexpand.c (expand_return): Likewise.
6698 * expr.c (expand_constructor): Likewise.
6699 (expand_expr_real_1): Likewise.
6701 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6703 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6706 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6709 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6711 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6713 * config/arm/aarch-common-protos.h
6714 (alu_cost_table): Fix spelling of "extend".
6715 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6717 2014-03-19 Richard Biener <rguenther@suse.de>
6720 * tree-core.h (tree_type_common): Re-order pointer members
6721 to reduce recursion depth during GC walks.
6723 2014-03-19 Marek Polacek <polacek@redhat.com>
6726 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6727 before accessing it.
6729 2014-03-19 Richard Biener <rguenther@suse.de>
6732 * lto-streamer-in.c (input_function): In WPA stage do not drop
6735 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6737 PR tree-optimization/60559
6738 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6739 with build_zero_cst assignment.
6741 2014-03-18 Kai Tietz <ktietz@redhat.com>
6743 PR rtl-optimization/56356
6744 * sdbout.c (sdbout_parms): Verify that parms'
6745 incoming argument is valid.
6746 (sdbout_reg_parms): Likewise.
6748 2014-03-18 Richard Henderson <rth@redhat.com>
6751 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6752 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6753 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6755 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6757 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6758 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6759 Italicize plugin event names in description. Explain that
6760 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6761 Remind that no GCC functions should be called after PLUGIN_FINISH.
6762 Explain what pragmas with expansion are.
6764 2014-03-18 Martin Liska <mliska@suse.cz>
6766 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6767 gimple call statement is update.
6768 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6769 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6771 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6774 * ubsan.c (ubsan_instrument_unreachable): Call
6775 initialize_sanitizer_builtins.
6776 (ubsan_pass): Likewise.
6779 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6780 varpool_finalize_decl instead of rest_of_decl_compilation.
6782 2014-03-18 Richard Biener <rguenther@suse.de>
6784 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6785 by using bitmap_and_compl instead of bitmap_and_compl_into.
6786 (df_rd_transfer_function): Likewise.
6788 2014-03-18 Richard Biener <rguenther@suse.de>
6790 * doc/lto.texi (fresolution): Fix typo.
6792 2014-03-18 Richard Biener <rguenther@suse.de>
6794 * doc/invoke.texi (flto): Update for changes in 4.9.
6796 2014-03-18 Richard Biener <rguenther@suse.de>
6798 * doc/loop.texi: Remove section on the removed lambda framework.
6799 Update loop docs with recent changes in preserving loop structure.
6801 2014-03-18 Richard Biener <rguenther@suse.de>
6803 * doc/lto.texi (-fresolution): Document.
6805 2014-03-18 Richard Biener <rguenther@suse.de>
6807 * doc/contrib.texi: Adjust my name.
6809 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6812 * internal-fn.c: Include diagnostic-core.h.
6813 (expand_BUILTIN_EXPECT): New function.
6814 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6815 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6816 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6817 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6819 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6820 Revert 3 argument __builtin_expect code.
6821 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6822 * gimple-fold.c (gimple_fold_call): Likewise.
6823 * tree.h (fold_builtin_expect): New prototype.
6824 * builtins.c (build_builtin_expect_predicate): Add predictor
6825 argument, if non-NULL, create 3 argument __builtin_expect.
6826 (fold_builtin_expect): No longer static. Add ARG2 argument,
6827 pass it through to build_builtin_expect_predicate.
6828 (fold_builtin_2): Adjust caller.
6829 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6830 * internal-fn.def (BUILTIN_EXPECT): New.
6832 2014-03-18 Tobias Burnus <burnus@net-b.de>
6835 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6836 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6837 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6839 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6842 * predict.c (combine_predictions_for_bb): Fix up formatting.
6843 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6844 fill what it points to if non-NULL.
6845 (tree_predict_by_opcode): Adjust caller, use the predictor.
6846 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6848 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6850 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6851 proper constant for the store mode.
6853 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6855 * symtab.c (change_decl_assembler_name): Fix transparent alias
6858 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6860 * config/aarch64/aarch64.c: Correct the comments about the
6861 aarch64 stack layout.
6863 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6865 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6866 check for GF_OMP_FOR_KIND_FOR.
6868 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6870 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6871 ymm and zmm register names.
6873 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6876 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6877 note creation for the 2010-08-31 changes.
6879 2014-03-17 Marek Polacek <polacek@redhat.com>
6882 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6883 as -fno-tree-loop-vectorize.
6884 (expand_omp_simd): Likewise.
6886 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6888 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6889 (eligible_for_call_delay): New prototype.
6890 * config/sparc/sparc.c (tls_call_delay): Rename into...
6891 (eligible_for_call_delay): ...this. Return false if the instruction
6892 cannot be put in the delay slot of a branch.
6893 (eligible_for_restore_insn): Simplify.
6894 (eligible_for_return_delay): Return false if the instruction cannot be
6895 put in the delay slot of a branch and simplify.
6896 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6897 put in the delay slot of a branch.
6898 * config/sparc/sparc.md (fix_ut699): New attribute.
6899 (tls_call_delay): Delete.
6900 (in_call_delay): Reimplement.
6901 (eligible_for_sibcall_delay): Rename into...
6902 (in_sibcall_delay): ...this.
6903 (eligible_for_return_delay): Rename into...
6904 (in_return_delay): ...this.
6905 (in_branch_delay): Reimplement.
6906 (in_uncond_branch_delay): Delete.
6907 (in_annul_branch_delay): Delete.
6909 2014-03-14 Richard Henderson <rth@redhat.com>
6912 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6913 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6914 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6915 (floathi splitters): Remove.
6916 (float<SWI48x>xf2): New pattern.
6917 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6918 code that tried to handle DImode for 32-bit, but which was excluded
6919 by the pattern's condition. Drop allocation of stack temporary.
6920 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6921 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6922 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6923 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6924 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6925 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6926 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6927 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6928 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6929 (*float<SWI48x><X87MODEF>2_i387): Remove.
6930 (all float _with_temp splitters): Remove.
6931 (*float<SWI48x><MODEF>2_i387): New pattern.
6932 (*float<SWI48><MODEF>2_sse): New pattern.
6933 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6934 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6936 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6937 Marek Polacek <polacek@redhat.com>
6940 * common.opt (dump_base_name_prefixed): New Variable.
6941 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6942 if x_dump_base_name_prefixed is already set, set it at the end.
6944 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6946 PR rtl-optimization/60508
6947 * lra-constraints.c (get_reload_reg): Add new parameter
6949 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6950 Pass the new parameter values.
6952 2014-03-14 Richard Biener <rguenther@suse.de>
6954 * common.opt: Revert unintented changes from r205065.
6957 2014-03-14 Richard Biener <rguenther@suse.de>
6960 * cfghooks.c (split_block): Properly adjust all loops the
6961 block was a latch of.
6963 2014-03-14 Martin Jambor <mjambor@suse.cz>
6966 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6969 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6972 * config/avr/avr.c (avr_set_current_function): Pass function name
6973 through default_strip_name_encoding before sanity checking instead
6974 of skipping the first char of the assembler name.
6976 2014-03-13 Richard Henderson <rth@redhat.com>
6979 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6980 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6981 * config/i386/i386-protos.h: Likewise.
6982 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6983 in the expander instead of a splitter.
6984 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6985 any possibility of requiring a memory.
6986 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6987 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6988 (fp branch splitters): Update for ix86_split_fp_branch.
6989 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6990 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6991 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6992 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6993 (*fop_<MODEF>_3_i387): Likewise.
6994 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6995 (splitters for the fop_* register patterns): Remove.
6996 (fscalexf4_i387): Rename from *fscalexf4_i387.
6997 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6999 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7001 PR tree-optimization/59779
7002 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7003 type for bitsize and maxsize instead of HOST_WIDE_INT.
7005 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7007 PR rtl-optimization/57320
7008 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7009 the CFG after thread_prologue_and_epilogue_insns.
7011 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7013 PR rtl-optimization/57189
7014 * lra-constraints.c (process_alt_operands): Disfavor spilling
7017 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7019 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7021 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7023 PR tree-optimization/59025
7025 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7026 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7028 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7031 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7032 calls of avr_out_plus_1.
7034 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7036 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7037 BB's single pred and update the father loop's latch info later.
7039 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7041 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7045 (VEC_base): Likewise.
7046 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7047 registers, we need to swap double words in little endian mode.
7049 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7050 to be a container mode for 128-bit integer operations added in ISA
7051 2.07. Unlike TImode and PTImode, the preferred register set is
7052 the Altivec/VMX registers for the 128-bit operations.
7054 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7056 (rs6000_split_128bit_ok_p): Likewise.
7058 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7059 macros for creating ISA 2.07 normal and overloaded builtin
7060 functions with 3 arguments.
7061 (BU_P8V_OVERLOAD_3): Likewise.
7062 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7063 for use as overloaded functions.
7064 (VPERM_1TI_UNS): Likewise.
7065 (VSEL_1TI): Likewise.
7066 (VSEL_1TI_UNS): Likewise.
7067 (ST_INTERNAL_1ti): Likewise.
7068 (LD_INTERNAL_1ti): Likewise.
7069 (XXSEL_1TI): Likewise.
7070 (XXSEL_1TI_UNS): Likewise.
7071 (VPERM_1TI): Likewise.
7072 (VPERM_1TI_UNS): Likewise.
7073 (XXPERMDI_1TI): Likewise.
7074 (SET_1TI): Likewise.
7075 (LXVD2X_V1TI): Likewise.
7076 (STXVD2X_V1TI): Likewise.
7077 (VEC_INIT_V1TI): Likewise.
7078 (VEC_SET_V1TI): Likewise.
7079 (VEC_EXT_V1TI): Likewise.
7080 (EQV_V1TI): Likewise.
7081 (NAND_V1TI): Likewise.
7082 (ORC_V1TI): Likewise.
7083 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7084 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7086 (VADDUQM): Likewise.
7087 (VSUBCUQ): Likewise.
7088 (VADDEUQM): Likewise.
7089 (VADDECUQ): Likewise.
7090 (VSUBEUQM): Likewise.
7091 (VSUBECUQ): Likewise.
7093 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7094 __int128_t and __uint128_t types.
7095 (__uint128_type): Likewise.
7096 (altivec_categorize_keyword): Add support for vector __int128_t,
7097 vector __uint128_t, vector __int128, and vector unsigned __int128
7098 as a container type for TImode operations that need to be done in
7099 VSX/Altivec registers.
7100 (rs6000_macro_to_expand): Likewise.
7101 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7102 to support 128-bit integer instructions vaddcuq, vadduqm,
7103 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7104 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7106 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7107 for V1TImode, and set up preferences to use VSX/Altivec registers.
7108 Setup VSX reload handlers.
7109 (rs6000_debug_reg_global): Likewise.
7110 (rs6000_init_hard_regno_mode_ok): Likewise.
7111 (rs6000_preferred_simd_mode): Likewise.
7112 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7113 (easy_altivec_constant): Likewise.
7114 (output_vec_const_move): Likewise.
7115 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7117 (rs6000_expand_vector_extract): Likewise.
7118 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7120 (rs6000_const_vec): Add support for V1TImode.
7121 (rs6000_emit_le_vsx_load): Swap double words when loading or
7122 storing TImode/V1TImode.
7123 (rs6000_emit_le_vsx_store): Likewise.
7124 (rs6000_emit_le_vsx_move): Likewise.
7125 (rs6000_emit_move): Add support for V1TImode.
7126 (altivec_expand_ld_builtin): Likewise.
7127 (altivec_expand_st_builtin): Likewise.
7128 (altivec_expand_vec_init_builtin): Likewise.
7129 (altivec_expand_builtin): Likewise.
7130 (rs6000_init_builtins): Add support for V1TImode type. Add
7131 support for ISA 2.07 128-bit integer builtins. Define type names
7132 for the VSX/Altivec vector types.
7133 (altivec_init_builtins): Add support for overloaded vector
7134 functions with V1TImode type.
7135 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7136 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7138 (rs6000_split_128bit_ok_p): Likewise.
7139 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7140 __int128_t and vector __uint128_t.
7142 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7143 and mode attributes.
7150 (VS_scalar): Likewise.
7151 (VS_double): Likewise.
7152 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7154 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7155 we support the ISA 2.07 128-bit integer arithmetic instructions.
7156 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7157 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7158 and TImode types for use with the builtin functions.
7159 (V1TI_type_node): Likewise.
7160 (unsigned_V1TI_type_node): Likewise.
7161 (intTI_type_internal_node): Likewise.
7162 (uintTI_type_internal_node): Likewise.
7164 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7165 128-bit builtin functions.
7166 (UNSPEC_VADDEUQM): Likewise.
7167 (UNSPEC_VADDECUQ): Likewise.
7168 (UNSPEC_VSUBCUQ): Likewise.
7169 (UNSPEC_VSUBEUQM): Likewise.
7170 (UNSPEC_VSUBECUQ): Likewise.
7171 (VM): Add V1TImode to vector mode iterators.
7173 (VI_unit): Likewise.
7174 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7175 (altivec_vaddcuq): Likewise.
7176 (altivec_vsubuqm): Likewise.
7177 (altivec_vsubcuq): Likewise.
7178 (altivec_vaddeuqm): Likewise.
7179 (altivec_vaddecuq): Likewise.
7180 (altivec_vsubeuqm): Likewise.
7181 (altivec_vsubecuq): Likewise.
7183 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7185 (BOOL_128): Likewise.
7186 (BOOL_REGS_OUTPUT): Likewise.
7187 (BOOL_REGS_OP1): Likewise.
7188 (BOOL_REGS_OP2): Likewise.
7189 (BOOL_REGS_UNARY): Likewise.
7190 (BOOL_REGS_AND_CR0): Likewise.
7192 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7193 128-bit integer builtin support.
7194 (vec_vadduqm): Likewise.
7195 (vec_vaddecuq): Likewise.
7196 (vec_vaddeuqm): Likewise.
7197 (vec_vsubecuq): Likewise.
7198 (vec_vsubeuqm): Likewise.
7199 (vec_vsubcuq): Likewise.
7200 (vec_vsubuqm): Likewise.
7202 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7203 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7204 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7205 128-bit integer add/subtract to ISA 2.07.
7207 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7209 * config/arc/arc.c (arc_predicate_delay_insns):
7210 Fix third argument passed to conditionalize_nonjump.
7212 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7214 * config/aarch64/aarch64-builtins.c
7215 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7216 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7217 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7218 instead of __builtin_lfloor.
7219 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7221 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7223 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7224 (tree_ssa_ifcombine_bb_1): New function.
7225 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7226 is an empty forwarder block to then_bb or vice versa and then_bb
7227 and else_bb are effectively swapped.
7229 2014-03-12 Christian Bruel <christian.bruel@st.com>
7232 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7233 REG_CFA_DEF_CFA note.
7234 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7235 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7237 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7239 PR tree-optimization/60454
7240 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7242 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7244 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7245 Do not define target_cpu_default2 to generic.
7246 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7247 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7248 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7250 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7251 Marc Glisse <marc.glisse@inria.fr>
7253 PR tree-optimization/60502
7254 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7255 instead of build_low_bits_mask.
7257 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7260 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7261 if there are multiple uses, but op doesn't live on E edge.
7262 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7263 clobber stmts before __builtin_unreachable.
7265 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7267 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7268 hard_frame_pointer_rtx.
7269 * cse.c (cse_insn): Remove volatile check.
7270 * cselib.c (cselib_process_insn): Likewise.
7271 * dse.c (scan_insn): Likewise.
7273 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7275 * config/arc/arc.c (conditionalize_nonjump): New function,
7277 (arc_ifcvt): ... this.
7278 (arc_predicate_delay_insns): Use it.
7280 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7282 * config/arc/predicates.md (extend_operand): During/after reload,
7283 allow const_int_operand.
7284 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7285 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7286 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7288 (umulsi3_highpart_i): Likewise.
7290 2014-03-11 Richard Biener <rguenther@suse.de>
7292 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7293 Add asserts to guard possible wrong-code bugs.
7295 2014-03-11 Richard Biener <rguenther@suse.de>
7297 PR tree-optimization/60429
7298 PR tree-optimization/60485
7299 * tree-ssa-structalias.c (set_union_with_increment): Properly
7300 take into account all fields that overlap the shifted vars.
7301 (do_sd_constraint): Likewise.
7302 (do_ds_constraint): Likewise.
7303 (get_constraint_for_ptr_offset): Likewise.
7305 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7307 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7308 (nios2_compute_frame_layout):
7309 Add calculation of cfun->machine->fp_save_offset.
7310 (nios2_expand_prologue): Correct setting of frame pointer register
7312 (nios2_expand_epilogue): Update recovery of stack pointer from
7313 frame pointer accordingly.
7314 (nios2_initial_elimination_offset): Update calculation of offset
7315 for eliminating to HARD_FRAME_POINTER_REGNUM.
7317 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7320 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7321 cgraph_get_create_node on VAR_DECLs.
7323 2014-03-10 Richard Biener <rguenther@suse.de>
7326 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7328 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7330 * config/vms/vms.opt (vms_float_format): New variable.
7332 2014-03-08 Tobias Burnus <burnus@net-b.de>
7334 * doc/invoke.texi (-fcilkplus): Update implementation status.
7336 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7337 Richard Biener <rguenther@suse.de>
7339 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7340 consistently accross all TUs.
7341 (run_gcc): Enable -fshort-double automatically at link at link-time
7342 and disallow override.
7344 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7347 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7348 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7349 if they can't be used.
7351 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7353 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7354 for Solaris 11/x86 ld.
7355 * configure: Regenerate.
7357 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7359 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7360 (LIB_TLS_SPEC): Save as ld_tls_libs.
7361 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7362 (HAVE_AS_IX86_TLSLDM): New test.
7363 * configure, config.in: Regenerate.
7364 * config/i386/i386.c (legitimize_tls_address): Fall back to
7365 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7366 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7367 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7368 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7370 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7372 * common.opt (fira-loop-pressure): Mark as optimization.
7374 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7376 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7377 an OpenMP mappable type.
7379 2014-03-06 Matthias Klose <doko@ubuntu.com>
7381 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7382 MULTILIB_OSDIRNAMES is not defined.
7384 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7385 Meador Inge <meadori@codesourcery.com>
7388 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7389 (arm_legitimize_address): Call legitimize_tls_address for any
7390 arm_tls_referenced_p expression, handle constant addend. Call it
7391 before testing for !TARGET_ARM.
7392 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7394 2014-03-06 Richard Biener <rguenther@suse.de>
7400 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7402 * tree-streamer.c (record_common_node): Assert we don't record
7403 nodes with type double.
7404 (preload_common_node): Skip type double, complex double and double
7405 pointer since it is now frontend dependent due to fshort-double option.
7407 2014-03-06 Richard Biener <rguenther@suse.de>
7409 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7410 or -fno-lto is specified and the linker has full plugin support.
7411 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7412 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7413 * lto-wrapper.c (merge_and_complain): Merge compile-time
7414 optimization levels.
7415 (run_gcc): And pass it through to the link options.
7417 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7421 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7423 * cselib.c (remove_useless_values): Skip to avoid quadratic
7424 behavior if the condition moved from...
7425 (cselib_process_insn): ... here holds.
7427 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7430 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7431 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7434 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7435 (TM_H): Add x86-tune.def.
7437 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7439 * config/aarch64/aarch64.c (generic_tunings):
7440 Use cortexa57_extra_costs.
7442 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7445 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7446 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7447 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7450 2014-03-04 Heiher <r@hev.cc>
7452 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7453 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7455 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7457 * config/i386/predicates.md (const2356_operand): Change to ...
7458 (const2367_operand): ... this.
7459 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7461 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7462 (*avx512pf_scatterpf<mode>sf): Ditto.
7463 (avx512pf_scatterpf<mode>df): Ditto.
7464 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7465 (*avx512pf_scatterpf<mode>df): Ditto.
7466 * config/i386/i386.c (ix86_expand_builtin): Update
7467 incorrect hint operand error message.
7469 2014-03-04 Richard Biener <rguenther@suse.de>
7471 * lto-section-in.c (lto_get_section_data): Fix const cast.
7473 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7475 * tree-streamer.c (record_common_node): Assert we don't record
7476 nodes with type double.
7477 (preload_common_node): Skip type double, complex double and double
7478 pointer since it is now frontend dependent due to fshort-double option.
7480 2014-03-04 Richard Biener <rguenther@suse.de>
7483 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7484 (lto_input_toplevel_asms): Likewise.
7485 * lto-section-in.c (lto_get_section_data): Instead do it here
7488 2014-03-04 Richard Biener <rguenther@suse.de>
7490 PR tree-optimization/60382
7491 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7492 dead PHIs a reduction.
7494 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7496 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7498 (_mm_prefetch): Move out of GCC target("sse") pragma.
7499 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7500 GCC target("prfchw") pragma.
7501 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7503 * config/i386/i386.c (ix86_option_override_internal): Enable
7504 -mprfchw with -mprefetchwt1.
7506 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7508 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7511 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7513 * opts.h (CL_PCH_IGNORE): Define.
7514 * targhooks.c (option_affects_pch_p):
7515 Return false for options that have CL_PCH_IGNORE set.
7516 * opt-functions.awk: Process PchIgnore.
7517 * doc/options.texi: Document PchIgnore.
7519 * config/arc/arc.opt (misize): Add PchIgnore property.
7521 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7523 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7524 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7525 constraint on constants to permit them being loaded into
7526 GENERAL_REGS or BASE_REGS.
7528 2014-03-03 Nick Clifton <nickc@redhat.com>
7530 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7531 anti-cacnonical alternatives.
7532 (negandhi3_real): New pattern.
7533 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7535 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7537 * config/avr/avr-mcus.def: Remove atxmega16x1.
7538 * config/avr/avr-tables.opt: Regenerate.
7539 * config/avr/t-multilib: Regenerate.
7540 * doc/avr-mmcu.texi: Regenerate.
7542 2014-03-03 Tobias Grosser <tobias@grosser.es>
7543 Mircea Namolaru <mircea.namolaru@inria.fr>
7545 PR tree-optimization/58028
7546 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7549 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7551 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7552 not handled by recognizers.
7554 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7557 * function.c (expand_function_end): Don't emit
7558 clobber_return_register sequence if clobber_after is a BARRIER.
7559 * cfgexpand.c (construct_exit_block): Append instructions before
7560 return_label to prev_bb.
7562 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7564 * config/rs6000/constraints.md: Document reserved use of "wc".
7566 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7569 * ipa.c (function_and_variable_visibility): When dissolving comdat
7570 group, also set all symbols to local.
7572 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7577 2013-12-14 Jan Hubicka <jh@suse.cz>
7579 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7581 2014-03-02 Jon Beniston <jon@beniston.com>
7587 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7588 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7589 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7590 (simple_return, *simple_return): New patterns
7591 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7592 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7594 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7596 * dwarf2out.c (gen_subprogram_die): Tidy.
7598 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7601 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7602 (*mov_t_msb_neg_negc): ... this new insn.
7604 2014-02-28 Jason Merrill <jason@redhat.com>
7607 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7610 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7613 * dwarf2out.c (decltype_auto_die): New static.
7614 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7615 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7616 (is_cxx_auto): Likewise.
7618 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7620 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7621 we are not using general regs only.
7623 2014-02-28 Richard Biener <rguenther@suse.de>
7626 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7627 previous fix and only allow to remove trivial pre-headers
7628 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7629 (remove_forwarder_block): Properly update the latch of a loop.
7631 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7634 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7635 (cselib_preserved_hash_table): New.
7636 (preserve_constants_and_equivs): Move preserved vals to it.
7637 (cselib_find_slot): Look it up first.
7638 (cselib_init): Initialize it.
7639 (cselib_finish): Release it.
7640 (dump_cselib_table): Dump it.
7642 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7645 * cselib.c (remove_useless_values): Skip to avoid quadratic
7646 behavior if the condition moved from...
7647 (cselib_process_insn): ... here holds.
7649 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7652 * var-tracking.c (vt_initialize): Apply the same condition to
7653 preserve the CFA base value.
7655 2014-02-28 Joey Ye <joey.ye@arm.com>
7658 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7659 if reload in progress or completed.
7661 2014-02-28 Tobias Burnus <burnus@net-b.de>
7664 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7667 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7669 * doc/tm.texi.in (Condition Code Status): Update documention for
7670 relative locations of cc0-setter and cc0-user.
7672 2014-02-27 Jeff Law <law@redhat.com>
7674 PR rtl-optimization/52714
7675 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7676 into two independent simple sets, if I3 is a jump, ensure the
7677 pattern we place into I3 is a (set (pc) ...).
7679 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7680 Jeff Law <law@redhat.com>
7682 PR rtl-optimization/49847
7683 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7684 are in different blocks.
7685 * doc/tm.texi (Condition Code Status): Update documention for
7686 relative locations of cc0-setter and cc0-user.
7688 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7691 * lra.c (lra_emit_add): Check SUBREG too.
7693 2014-02-27 Andreas Schwab <schwab@suse.de>
7695 * config/m68k/m68k.c (m68k_option_override): Disable
7696 -flive-range-shrinkage for classic m68k.
7697 (m68k_override_options_after_change): Likewise.
7699 2014-02-27 Marek Polacek <polacek@redhat.com>
7702 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7703 -Wmaybe-uninitialized.
7705 2014-02-27 Alan Modra <amodra@gmail.com>
7708 * reload1.c (emit_input_reload_insns): When reload_override_in,
7709 set old to rl->in_reg when rl->in_reg is a subreg.
7711 2014-02-26 Richard Biener <rguenther@suse.de>
7714 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7716 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7718 * common/config/i386/predicates.md (const1256_operand): Remove.
7719 (const2356_operand): New.
7720 (const_1_to_2_operand): Remove.
7721 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7722 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7723 (*avx512pf_gatherpf<mode>sf): Ditto.
7724 (avx512pf_gatherpf<mode>df): Ditto.
7725 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7726 (*avx512pf_gatherpf<mode>df): Ditto.
7727 (avx512pf_scatterpf<mode>sf): Ditto.
7728 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7729 (*avx512pf_scatterpf<mode>sf): Ditto.
7730 (avx512pf_scatterpf<mode>df): Ditto.
7731 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7732 (*avx512pf_scatterpf<mode>df): Ditto.
7733 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7735 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7737 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7738 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7739 (_mm512_mask_testn_epi64_mask): Move to ...
7740 * config/i386/avx512cdintrin.h: Here.
7741 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7742 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7743 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7744 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7745 TARGET_AVX512F from TARGET_AVX512CD.
7747 2014-02-26 Richard Biener <rguenther@suse.de>
7750 * ipa.c (walk_polymorphic_call_targets): Properly guard
7751 call to inline_update_overall_summary.
7753 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7756 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7757 and latches only if requested. Fix latch if it is removed.
7758 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7759 LOOPS_HAVE_PREHEADERS.
7761 2014-02-25 Andrew Pinski <apinski@cavium.com>
7763 * builtins.c (expand_builtin_thread_pointer): Create a new target
7764 when the target is NULL.
7766 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7768 PR rtl-optimization/60317
7769 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7770 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7771 * lra-assigns.c: Include params.h.
7772 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7773 other reload pseudos considerations.
7775 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7777 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7778 to use canonical form for nor<mode>3.
7780 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7783 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7786 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7788 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7789 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7790 (ix86_handle_option): Handle OPT_mprefetchwt1.
7791 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7792 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7794 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7795 OPTION_MASK_ISA_PREFETCHWT1.
7796 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7797 (PTA_PREFETCHWT1): New.
7798 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7799 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7800 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7801 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7802 (*prefetch_avx512pf_<mode>_: Change into ...
7803 (*prefetch_prefetchwt1_<mode>: This.
7804 * config/i386/i386.opt (mprefetchwt1): New.
7805 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7806 (_mm_prefetch): Handle intent to write.
7807 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7809 2014-02-25 Richard Biener <rguenther@suse.de>
7812 * emit-rtl.c (mem_attrs_htab): Remove.
7813 (mem_attrs_htab_hash): Likewise.
7814 (mem_attrs_htab_eq): Likewise.
7815 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7816 (init_emit_once): Do not allocate mem_attrs_htab.
7818 2014-02-25 Richard Biener <rguenther@suse.de>
7821 * lto-opts.c (lto_write_options): Output non-explicit conservative
7822 -fwrapv, -fno-trapv and -fno-strict-overflow.
7823 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7824 (run_gcc): And pass them through.
7826 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7828 * sel-sched.c (calculate_new_fences): New parameter ptime.
7829 Calculate it as a maximum over all fence cycles.
7830 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7831 Print the final schedule timing when sched_verbose.
7833 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7835 PR rtl-optimization/60292
7836 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7837 bit fot the fence instruction.
7839 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7841 * calls.h: Fix typo in comment.
7843 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7845 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7846 adjusting offsetable addresses.
7848 2014-02-24 Guozhi Wei <carrot@google.com>
7850 * sparseset.h (sparseset_pop): Fix the wrong index.
7852 2014-02-24 Walter Lee <walt@tilera.com>
7854 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7855 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7857 * common/config/tilegx/tilegx-common.c
7858 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7859 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7861 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7862 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7863 (tilegx_gimplify_va_arg_expr): Handle big endian.
7864 (tilegx_expand_unaligned_load): Ditto.
7865 (tilegx_expand_unaligned_store): Ditto.
7866 (TARGET_RETURN_IN_MSB): New.
7867 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7868 (TARGET_ENDIAN_DEFAULT): New.
7869 (TARGET_BIG_ENDIAN): Handle big endian.
7870 (BYTES_BIG_ENDIAN): Ditto.
7871 (WORDS_BIG_ENDIAN): Ditto.
7872 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7875 * config/tilegx/tilegx.md (extv): Handle big endian.
7877 (insn_st<n>): Ditto.
7878 (insn_st<n>_add<bitsuffix>): Ditto.
7879 (insn_stnt<n>): Ditto.
7880 (insn_stnt<n>_add<bitsuffix>):Ditto.
7881 (vec_interleave_highv8qi): Handle big endian.
7882 (vec_interleave_highv8qi_be): New.
7883 (vec_interleave_highv8qi_le): New.
7884 (insn_v1int_h): Handle big endian.
7885 (vec_interleave_lowv8qi): Handle big endian.
7886 (vec_interleave_lowv8qi_be): New.
7887 (vec_interleave_lowv8qi_le): New.
7888 (insn_v1int_l): Handle big endian.
7889 (vec_interleave_highv4hi): Handle big endian.
7890 (vec_interleave_highv4hi_be): New.
7891 (vec_interleave_highv4hi_le): New.
7892 (insn_v2int_h): Handle big endian.
7893 (vec_interleave_lowv4hi): Handle big endian.
7894 (vec_interleave_lowv4hi_be): New.
7895 (vec_interleave_lowv4hi_le): New.
7896 (insn_v2int_l): Handle big endian.
7897 (vec_interleave_highv2si): Handle big endian.
7898 (vec_interleave_highv2si_be): New.
7899 (vec_interleave_highv2si_le): New.
7900 (insn_v4int_h): Handle big endian.
7901 (vec_interleave_lowv2si): Handle big endian.
7902 (vec_interleave_lowv2si_be): New.
7903 (vec_interleave_lowv2si_le): New.
7904 (insn_v4int_l): Handle big endian.
7905 * config/tilegx/tilegx.opt (mbig-endian): New option.
7906 (mlittle-endian): New option.
7907 * doc/install.texi: Document tilegxbe-linux.
7908 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7910 2014-02-24 Martin Jambor <mjambor@suse.cz>
7913 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7914 there are no parameter descriptors.
7916 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7918 PR rtl-optimization/60268
7919 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7920 initialization to ...
7921 (sched_rgn_init): ... here.
7922 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7924 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7926 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7929 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7931 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7934 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7936 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7937 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7939 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7941 * config/microblaze/predicates.md: Add cmp_op predicate.
7942 * config/microblaze/microblaze.md: Add branch_compare instruction
7943 which uses cmp_op predicate and emits cmp insn before branch.
7944 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7945 to microblaze_expand_conditional_branch and consolidate logic.
7946 (microblaze_expand_conditional_branch): emit branch_compare
7947 insn instead of handling cmp op separate from branch insn.
7949 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7951 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7954 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7956 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7957 define_insn with define_expand and new define_insn
7958 *altivec_lve<VI_char>x_internal.
7959 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7960 and new define_insn *altivec_stve<VI_char>x_internal.
7961 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7963 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7965 (altivec_expand_stvex_be): New function.
7967 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7969 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7970 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7971 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7972 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7974 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7977 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7978 instead of emit_move_insn.
7980 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7982 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7984 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7985 gen_altivec_vsumsws.
7987 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7989 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7990 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7991 (altivec_lvxl_<mode>): New define_expand incorporating
7992 -maltivec=be semantics where needed.
7993 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7994 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7995 semantics where needed.
7996 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7997 (altivec_stvx_<mode>): New define_expand incorporating
7998 -maltivec=be semantics where needed.
7999 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8000 VM2 iterator instead of V4SI.
8001 (altivec_stvxl_<mode>): New define_expand incorporating
8002 -maltivec=be semantics where needed.
8003 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8004 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8005 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8006 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8007 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8008 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8009 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8010 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8011 ALTIVEC_BUILTIN_STVXL.
8012 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8013 (altivec_expand_stvx_be): Likewise.
8014 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8015 (altivec_expand_lvx_be): Likewise.
8016 (altivec_expand_stvx_be): Likewise.
8017 (altivec_expand_builtin): Add cases for
8018 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8019 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8020 (altivec_init_builtins): Add definitions for
8021 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8022 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8024 2014-02-21 Catherine Moore <clm@codesourcery.com>
8026 * doc/invoke.texi (mvirt, mno-virt): Document.
8027 * config/mips/mips.opt (mvirt): New option.
8028 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8030 2014-02-21 Richard Biener <rguenther@suse.de>
8032 PR tree-optimization/60276
8033 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8034 (STMT_VINFO_MIN_NEG_DIST): New macro.
8035 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8036 STMT_VINFO_MIN_NEG_DIST.
8037 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8038 made for negative dependence distances still hold.
8040 2014-02-21 Richard Biener <rguenther@suse.de>
8043 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8044 DECL_INITIAL for globals not in the current function context.
8046 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8048 PR tree-optimization/56490
8049 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8050 * tree-ssa-uninit.c: Include params.h.
8051 (compute_control_dep_chain): Add num_calls argument, return false
8052 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8053 num_calls to recursive call.
8054 (find_predicates): Change dep_chain into normal array,
8055 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8056 variable and adjust compute_control_dep_chain caller.
8057 (find_def_preds): Likewise.
8059 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8061 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8062 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8064 2014-02-21 Nick Clifton <nickc@redhat.com>
8066 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8067 (pushhi1): Likewise.
8068 (popqi1): Add mode to pre_dec.
8071 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8073 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8074 mode for mask of V8SFmode permutation.
8076 2014-02-20 Richard Henderson <rth@redhat.com>
8079 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8080 a new pseudo for OLDVAL.
8082 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8085 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8086 gen_reg_rtx if d->testing_p.
8087 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8088 if d->testing_p and we will certainly return true.
8089 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8092 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8094 * emit-rtl.c (gen_reg_rtx): Assert that
8095 crtl->emit.regno_pointer_align_length is non-zero.
8097 2014-02-20 Richard Henderson <rth@redhat.com>
8100 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8101 on failure the store back into EXPECT.
8103 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8104 Sandra Loosemore <sandra@codesourcery.com>
8106 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8107 * config/nios2/nios2.c (nios2_function_profiler): Add
8108 -fPIC (flag_pic == 2) support.
8109 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8110 (nios2_large_offset_p): New function.
8111 (nios2_unspec_reloc_p): Move up position, update to use
8112 nios2_large_offset_p.
8113 (nios2_unspec_address): Remove function.
8114 (nios2_unspec_offset): New function.
8115 (nios2_large_got_address): New function.
8116 (nios2_got_address): Add large offset support.
8117 (nios2_legitimize_tls_address): Update usage of removed and new
8119 (nios2_symbol_binds_local_p): New function.
8120 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8121 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8122 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8123 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8124 processing for (const (unspec ...)).
8125 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8127 2014-02-20 Richard Biener <rguenther@suse.de>
8129 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8130 doing the substitution.
8131 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8133 2014-02-20 Martin Jambor <mjambor@suse.cz>
8136 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8137 info when checking whether lattices are bottom.
8139 2014-02-20 Richard Biener <rguenther@suse.de>
8142 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8145 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8148 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8149 parameter specifying the scaling.
8150 (inline_call): Update.
8151 (want_inline_recursively): Guard division by zero.
8152 (recursive_inlining): Update.
8153 * ipa-inline.h (clone_inlined_nodes): Update.
8155 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8158 * config/i386/i386.c (classify_argument): Pass structures of size
8159 64 bytes or less in register.
8161 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8162 Kirill Yukhin <kirill.yukhin@intel.com>
8164 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8165 (_mm_rcp28_round_ss): Ditto.
8166 (_mm_rsqrt28_round_sd): Ditto.
8167 (_mm_rsqrt28_round_ss): Ditto.
8168 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8169 (_mm_rcp14_round_ss): Ditto.
8170 (_mm_rsqrt14_round_sd): Ditto.
8171 (_mm_rsqrt14_round_ss): Ditto.
8172 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8173 the first input operand, get rid of match_dup.
8174 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8176 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8178 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8179 operand as the first input operand, set type attribute.
8180 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8182 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8183 operand as the first input operand, set type attribute.
8185 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8187 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8190 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8193 * config/i386/i386.c (construct_container): Remove TFmode check
8194 for X86_64_INTEGER_CLASS.
8196 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8199 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8200 only when -Wpsabi is enabled.
8202 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8205 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8206 passing arrays in registers are the same as for structs, so remove the
8207 special case for them.
8209 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8211 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8212 destination type, extract only the valid bits if the source type is not
8213 integral and has a different mode.
8215 2014-02-19 Richard Biener <rguenther@suse.de>
8218 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8221 2014-02-19 Richard Biener <rguenther@suse.de>
8224 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8225 (ipa_modify_call_arguments): Emit an argument load explicitely and
8226 preserve virtual SSA form there and for the replacement call.
8227 Do not update SSA form nor free dominance info.
8229 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8231 * ipa.c (function_and_variable_visibility): Also clear WEAK
8232 flag when disolving COMDAT_GROUP.
8234 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8236 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8237 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8239 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8240 do more sanity checks.
8241 (detect_type_change): Return true when giving up early.
8242 (compute_complex_assign_jump_func): Fix type parameter of
8243 ipa_set_ancestor_jf.
8244 (compute_complex_ancestor_jump_func): Likewise.
8245 (update_jump_functions_after_inlining): Fix updating of
8247 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8249 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8251 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8252 inline clones when edge disappears.
8254 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8257 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8258 Split 64-bit moves into 2 patterns. Do not allow the use of
8259 direct move for TDmode in little endian, since the decimal value
8260 has little endian bytes within a word, but the 64-bit pieces are
8261 ordered in a big endian fashion, and normal subreg's of TDmode are
8263 (mov<mode>_64bit_dm): Likewise.
8264 (movtd_64bit_nodm): Likewise.
8266 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8268 PR tree-optimization/60174
8269 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8270 statement of an SSA_NAME that occurs in an abnormal PHI node.
8272 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8275 * final.c (SEEN_BB): Remove.
8276 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8277 (final_scan_insn): Don't force_source_line on second
8278 NOTE_INSN_BASIC_BLOCK.
8280 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8283 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8284 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8285 (type_natural_mode): Warn ABI change when %zmm register is not
8286 available for AVX512F vector value passing.
8288 2014-02-18 Kai Tietz <ktietz@redhat.com>
8291 * config/i386/i386.c (ix86_expand_prologue): Use value in
8292 rax register as displacement when restoring %r10 or %rax.
8293 Fix wrong offset when restoring both registers.
8295 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8297 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8298 assertion with conditional return.
8300 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8301 Uros Bizjak <ubizjak@gmail.com>
8304 * config/i386/driver-i386.c (host_detect_local_cpu): If
8305 YMM state is not saved by the OS, also clear has_f16c. Move
8306 CPUID 0x80000001 handling before YMM state saving checking.
8308 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8310 PR rtl-optimization/58960
8311 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8312 factored out from ...
8313 (sched_init): ... here.
8314 (free_global_sched_pressure_data): New, factored out from ...
8315 (sched_finish): ... here.
8316 * sched-int.h (free_global_sched_pressure_data): Declare.
8317 * sched-rgn.c (nr_regions_initial): New static global.
8318 (haifa_find_rgns): Initialize it.
8319 (schedule_region): Disable sched-pressure for the newly
8322 2014-02-17 Richard Biener <rguenther@suse.de>
8324 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8325 release SSA defs of pattern stmts.
8327 2014-02-17 Richard Biener <rguenther@suse.de>
8329 * tree-inline.c (expand_call_inline): Release the virtual
8330 operand defined by the call we are about to inline.
8332 2014-02-17 Richard Biener <rguenther@suse.de>
8334 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8336 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8337 Ilya Tocar <ilya.tocar@intel.com>
8339 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8340 arguments order in builtin.
8341 (_mm512_permutexvar_epi64): Ditto.
8342 (_mm512_mask_permutexvar_epi64): Ditto
8343 (_mm512_maskz_permutexvar_epi32): Ditto
8344 (_mm512_permutexvar_epi32): Ditto
8345 (_mm512_mask_permutexvar_epi32): Ditto
8347 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8349 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8350 (p8_vmrgow): Likewise.
8352 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8354 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8357 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8360 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8361 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8362 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8363 using direct move instructions on ISA 2.07. Also adjust
8364 instruction length for 64-bit.
8365 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8366 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8368 2014-02-15 Alan Modra <amodra@gmail.com>
8372 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8373 find_replacement on parts of insn rtl that might be reloaded.
8375 2014-02-15 Richard Biener <rguenther@suse.de>
8377 PR tree-optimization/60183
8378 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8379 (tree_ssa_phiprop): Calculate and free post-dominators.
8381 2014-02-14 Jeff Law <law@redhat.com>
8383 PR rtl-optimization/60131
8384 * ree.c (get_extended_src_reg): New function.
8385 (combine_reaching_defs): Use it rather than assuming location of REG.
8386 (find_and_remove_re): Verify first operand of extension is
8387 a REG before adding the insns to the copy list.
8389 2014-02-14 Roland McGrath <mcgrathr@google.com>
8391 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8392 * configure: Regenerated.
8393 * config.in: Regenerated.
8394 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8395 instead of ASM_SHORT.
8397 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8398 Richard Earnshaw <rearnsha@arm.com>
8400 PR rtl-optimization/59535
8401 * lra-constraints.c (process_alt_operands): Encourage alternative
8402 when unassigned pseudo class is superset of the alternative class.
8403 (inherit_reload_reg): Don't inherit when optimizing for code size.
8404 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8405 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8406 modes not less than 4 for Thumb1.
8408 2014-02-14 Kyle McMartin <kyle@redhat.com>
8411 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8413 2014-02-14 Richard Biener <rguenther@suse.de>
8415 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8416 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8417 Do not strip INDIRECT_REFs.
8419 2014-02-14 Richard Biener <rguenther@suse.de>
8422 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8423 DECL_FUNCTION_SPECIFIC_TARGET.
8424 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8425 * tree-streamer-out.c (pack_ts_target_option): Remove.
8426 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8427 (write_ts_function_decl_tree_pointers): Do not stream
8428 DECL_FUNCTION_SPECIFIC_TARGET.
8429 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8430 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8431 (lto_input_ts_function_decl_tree_pointers): Do not stream
8432 DECL_FUNCTION_SPECIFIC_TARGET.
8434 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8436 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8437 (get_initial_def_for_induction, vectorizable_induction): Ignore
8438 debug stmts when looking for exit_phi.
8439 (vectorizable_live_operation): Fix up condition.
8441 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8443 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8444 nreverse() because it changes the content of original tree list.
8446 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8448 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8449 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8451 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8453 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8454 GNU coding standards.
8456 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8459 * dwarf2out.c (gen_subprogram_die): Don't call
8460 add_calling_convention_attribute if subr_die is old_die.
8462 2014-02-13 Sharad Singhai <singhai@google.com>
8464 * doc/optinfo.texi: Fix order of nodes.
8466 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8468 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8469 operands[2], not operands[3].
8471 2014-02-13 Richard Biener <rguenther@suse.de>
8474 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8475 mention the possibility of an in-tree build.
8476 (CLooG): Update recommended version to 0.18.1, mention the
8477 possibility of an in-tree build and clarify that the ISL
8478 bundled with CLooG does not work.
8480 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8483 * expr.c (compress_float_constant): If x is a hard register,
8484 extend into a pseudo and then move to x.
8486 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8488 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8489 caused by bad second argument to warning_at() with -mhotpatch and
8490 nested functions (e.g. with gfortran).
8492 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8494 * opts.c (option_name): Remove "enabled by default" rider.
8496 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8498 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8500 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8501 Uros Bizjak <ubizjak@gmail.com>
8504 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8505 * configure: Regenerated.
8507 2014-02-12 Richard Biener <rguenther@suse.de>
8509 * vec.c (vec_prefix::calculate_allocation): Move as
8510 inline variant to vec.h.
8511 (vec_prefix::calculate_allocation_1): New out-of-line version.
8512 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8513 (vec_prefix::m_has_auto_buf): Rename to ...
8514 (vec_prefix::m_using_auto_storage): ... this.
8515 (vec_prefix::calculate_allocation): Inline the easy cases
8516 and dispatch to calculate_allocation_1 which doesn't need the
8518 (va_heap::reserve): Use gcc_checking_assert.
8519 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8520 m_using_auto_storage.
8521 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8523 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8524 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8525 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8527 2014-02-12 Richard Biener <rguenther@suse.de>
8529 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8530 when we found a dependence.
8532 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8534 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8536 (maybe_fold_stmt): ... into this new function.
8537 * omp-low.c (lower_omp): Update comment.
8539 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8542 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8545 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8547 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8548 identifiers in comments.
8549 (cortexa53_extra_costs): Likewise.
8550 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8551 (cortexa7_extra_costs): Likewise.
8552 (cortexa12_extra_costs): Likewise.
8553 (cortexa15_extra_costs): Likewise.
8554 (v7m_extra_costs): Likewise.
8556 2014-02-12 Richard Biener <rguenther@suse.de>
8559 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8560 of posix_memalign being successful.
8561 (lower_stmt): Restrict lowering of posix_memalign to when
8562 -ftree-bit-ccp is enabled.
8564 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8566 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8568 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8570 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8572 PR rtl-optimization/60116
8573 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8574 other_insn once the combination has been validated.
8576 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8579 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8581 * ipa-devirt.c: Include demangle.h
8582 (odr_violation_reported): New static variable.
8583 (add_type_duplicate): Update odr_violations.
8584 (maybe_record_node): Add completep parameter; update it.
8585 (record_target_from_binfo): Add COMPLETEP parameter;
8586 update it as needed.
8587 (possible_polymorphic_call_targets_1): Likewise.
8588 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8589 rename FINAL to COMPLETE.
8590 (record_targets_from_bases): Sanity check we found the binfo;
8591 fix COMPLETEP updating.
8592 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8593 parameter, fix computing of COMPLETEP.
8594 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8595 at LTO time do demangling.
8596 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8597 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8599 (gimple_get_virt_method_for_binfo): Likewise.
8600 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8601 gimple_get_virt_method_for_vtable): Update prototypes.
8603 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8606 * genautomata.c (add_presence_absence): Fix typo with
8607 {final_}presence_list.
8609 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8612 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8613 for VSX/Altivec vectors that land in GPR registers.
8615 2014-02-11 Richard Henderson <rth@redhat.com>
8616 Jakub Jelinek <jakub@redhat.com>
8619 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8620 around drhs if type conversion to lacc->type is not useless.
8622 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8624 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8626 (cortex-a57.cortex-a53): Likewise.
8627 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8629 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8631 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8634 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8636 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8637 add_options_for_arm_vfp3.
8639 2014-02-11 Jeff Law <law@redhat.com>
8642 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8643 object with an undesirable mode.
8645 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8648 * config/i386/sol2-9.h: New file.
8649 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8650 *-*-solaris2.9*): Use it.
8652 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8654 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8655 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8657 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8659 * config/microblaze/microblaze.c: Extend mcpu version format
8661 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8663 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8665 2014-02-10 Richard Henderson <rth@redhat.com>
8668 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8669 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8670 ms-abi vs -mno-accumulate-outgoing-args.
8671 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8672 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8675 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8678 * cfgexpand.c (expand_asm_operands): Attach source location to
8679 ASM_INPUT rtx objects.
8680 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8682 2014-02-10 Nick Clifton <nickc@redhat.com>
8684 * config/mn10300/mn10300.c (popcount): New function.
8685 (mn10300_expand_prologue): Include saved registers in stack usage
8688 2014-02-10 Jeff Law <law@redhat.com>
8691 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8692 when changing the SET_DEST of a prior insn to avoid an input reload.
8694 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8696 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8697 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8698 -mcall-openbsd, or -mcall-linux.
8699 (CC1_ENDIAN_BIG_SPEC): Remove.
8700 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8701 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8702 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8703 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8704 and %cc1_endian_default.
8705 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8707 2014-02-10 Richard Biener <rguenther@suse.de>
8709 PR tree-optimization/60115
8710 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8711 MEM_REF handling. Properly verify that the accesses are not
8712 out of the objects bound.
8714 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8716 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8719 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8721 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8722 proper constants and fix formatting.
8723 (possible_polymorphic_call_targets): Fix formatting.
8725 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8726 Ilya Tocar <ilya.tocar@intel.com>
8728 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8729 (_mm512_loadu_epi32): Renamed into...
8730 (_mm512_loadu_si512): This.
8731 (_mm512_storeu_epi32): Renamed into...
8732 (_mm512_storeu_si512): This.
8733 (_mm512_maskz_ceil_ps): Removed.
8734 (_mm512_maskz_ceil_pd): Ditto.
8735 (_mm512_maskz_floor_ps): Ditto.
8736 (_mm512_maskz_floor_pd): Ditto.
8737 (_mm512_floor_round_ps): Ditto.
8738 (_mm512_floor_round_pd): Ditto.
8739 (_mm512_ceil_round_ps): Ditto.
8740 (_mm512_ceil_round_pd): Ditto.
8741 (_mm512_mask_floor_round_ps): Ditto.
8742 (_mm512_mask_floor_round_pd): Ditto.
8743 (_mm512_mask_ceil_round_ps): Ditto.
8744 (_mm512_mask_ceil_round_pd): Ditto.
8745 (_mm512_maskz_floor_round_ps): Ditto.
8746 (_mm512_maskz_floor_round_pd): Ditto.
8747 (_mm512_maskz_ceil_round_ps): Ditto.
8748 (_mm512_maskz_ceil_round_pd): Ditto.
8749 (_mm512_expand_pd): Ditto.
8750 (_mm512_expand_ps): Ditto.
8751 * config/i386/i386.c (ix86_builtins): Remove
8752 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8753 (bdesc_args): Ditto.
8754 * config/i386/predicates.md (const1256_operand): New.
8755 (const_1_to_2_operand): Ditto.
8756 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8757 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8758 (*avx512pf_gatherpf<mode>sf): Ditto.
8759 (avx512pf_gatherpf<mode>df): Ditto.
8760 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8761 (*avx512pf_gatherpf<mode>df): Ditto.
8762 (avx512pf_scatterpf<mode>sf): Ditto.
8763 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8764 (*avx512pf_scatterpf<mode>sf): Ditto.
8765 (avx512pf_scatterpf<mode>df): Ditto.
8766 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8767 (*avx512pf_scatterpf<mode>df): Ditto.
8768 (avx512f_expand<mode>): Removed.
8769 (<shift_insn><mode>3<mask_name>): Change predicate type.
8771 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8773 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8774 not at the end of datarefs vector use ordered_remove to avoid
8775 reordering datarefs vector.
8778 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8779 mark local addressable non-static vars as GOVD_PRIVATE
8780 instead of GOVD_LOCAL.
8781 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8782 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8786 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8787 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8788 assume_aligned or alloc_align attributes.
8789 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8790 arguments. Handle also assume_aligned and alloc_align attributes.
8791 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8792 calls to functions with assume_aligned or alloc_align attributes.
8793 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8795 2014-02-08 Terry Guo <terry.guo@arm.com>
8797 * doc/invoke.texi: Document ARM -march=armv7e-m.
8799 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8801 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8802 flag on __cilkrts_rethrow builtin.
8805 * ipa-cp.c (determine_versionability): Fail at -O0
8806 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8807 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8810 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8813 * tree-inline.c (copy_forbidden): Fail for
8814 __attribute__((optimize (0))) functions.
8816 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8818 * varpool.c: Include pointer-set.h.
8819 (varpool_remove_unreferenced_decls): Variables in other partitions
8820 will not be output; be however careful to not lose information
8823 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8825 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8826 lookup in the vtable constructor.
8828 2014-02-07 Jeff Law <law@redhat.com>
8831 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8832 define_insn_and_split.
8834 * ipa-inline.c (inline_small_functions): Fix typos.
8836 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8838 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8839 (s390_can_use_return_insn): Declare.
8840 * config/s390/s390.h (EPILOGUE_USES): Define.
8841 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8843 (s390_chunkify_start): Handle return JUMP_LABELs.
8844 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8845 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8846 (s390_can_use_return_insn): New functions.
8847 (s390_fix_long_loop_prediction): Handle conditional returns.
8848 (TARGET_SET_UP_BY_PROLOGUE): Define.
8849 * config/s390/s390.md (ANY_RETURN): New code iterator.
8850 (*creturn, *csimple_return, return, simple_return): New patterns.
8852 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8854 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8855 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8856 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8857 REG_CFA_RESTORE list when deciding not to restore a register.
8859 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8861 * config/s390/s390.c: Include tree-pass.h and context.h.
8862 (s390_early_mach): New function, split out from...
8863 (s390_emit_prologue): ...here.
8864 (pass_data_s390_early_mach): New pass structure.
8865 (pass_s390_early_mach): New class.
8866 (s390_option_override): Create and register early_mach pass.
8867 Move to end of file.
8869 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8871 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8872 to match for the exit block.
8874 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8876 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8877 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8878 Reject misaligned operands.
8880 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8882 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8884 2014-02-07 Richard Biener <rguenther@suse.de>
8887 * gimple-low.c (lower_builtin_posix_memalign): New function.
8888 (lower_stmt): Call it to lower posix_memalign in a way
8889 to make alignment info accessible.
8891 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8894 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8895 __builtin_setjmp_receiver.
8897 2014-02-07 Richard Biener <rguenther@suse.de>
8900 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8901 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8902 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8903 Handle BUILT_IN_POSIX_MEMALIGN.
8904 (find_func_clobbers): Likewise.
8905 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8906 (call_may_clobber_ref_p_1): Likewise.
8908 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8911 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8914 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8917 * lto-cgraph.c (lto_output_node): Use
8918 symtab_get_symbol_partitioning_class.
8919 (lto_output_varpool_node): likewise.
8920 (symtab_get_symbol_partitioning_class): Move here from
8922 * cgraph.h (symbol_partitioning_class): Likewise.
8923 (symtab_get_symbol_partitioning_class): Declare.
8925 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8927 * ggc.h (ggc_internal_cleared_alloc): New macro.
8928 * vec.h (vec_safe_copy): Handle memory stats.
8929 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8930 * target-globals.c (save_target_globals): Likewise.
8932 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8935 * expr.c (emit_move_resolve_push): Export; be bit more selective
8936 on when to clear alias set.
8937 * expr.h (emit_move_resolve_push): Declare.
8938 * function.h (struct function): Add tail_call_marked.
8939 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8940 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8941 * config/i386/i386.md (TImode move expander): De not call
8943 (FP push expanders): Preserve memory attributes.
8944 * config/i386/sse.md (push<mode>1): Remove.
8945 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8946 (ix86_expand_push): Remove.
8947 * config/i386/mmx.md (push<mode>1): Remove.
8949 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8951 PR rtl-optimization/60030
8952 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8953 lopart with paradoxical subreg before shifting it up by hprec.
8955 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8957 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8958 Remove extra newline at end of file.
8959 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8960 (arm_issue_rate): Handle cortexa57.
8961 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8962 (cortex-a57.cortex-a53): Likewise.
8964 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8967 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8968 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8970 (arm_expand_prologue): Adjust all callers.
8971 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8972 info, registers also at the lowest numbered registers side. Use
8973 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8977 * var-tracking.c (adjust_mems): Before adding a SET to
8978 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8980 2014-02-06 Alan Modra <amodra@gmail.com>
8983 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8984 change SDmode to DDmode when lra_in_progress.
8986 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8989 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8990 free_data_ref on the dr first, and before goto again also set dr
8991 to the next dr. For simd_lane_access, free old datarefs[i] before
8992 overwriting it. For get_vectype_for_scalar_type failure, don't
8993 free_data_ref if simd_lane_access.
8995 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8998 * tree.h (opts_for_fn): New inline function.
8999 (opt_for_fn): Define.
9000 * config/i386/i386.c (ix86_function_regparm): Use
9001 opt_for_fn (decl, optimize) instead of optimize.
9003 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9005 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9006 for SYMBOL_REF in large memory model.
9008 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9010 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9012 (cortex-a57): Likewise.
9013 (cortex-a57.cortex-a53): Likewise.
9015 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9016 Kugan Vivekanandarajah <kuganv@linaro.org>
9018 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9020 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9022 2014-02-06 Richard Biener <rguenther@suse.de>
9024 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9025 set_loop_copy and initialize_original_copy_tables.
9027 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9029 * config/aarch64/aarch64-simd.md
9030 (aarch64_ashr_simddi): Change QI to SI.
9032 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9033 Jakub Jelinek <jakub@redhat.com>
9036 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9039 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9041 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9042 CODE_FOR_altivec_vpku[hw]um to
9043 CODE_FOR_altivec_vpku[hw]um_direct.
9044 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9045 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9046 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9047 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9049 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9051 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9052 generation for -maltivec=be.
9053 (altivec_vsumsws): Simplify redundant test.
9055 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9057 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9058 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9059 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9060 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9061 gen_altivec_vpkuwum.
9062 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9064 (altivec_vpks<VI_char>ss): Likewise.
9065 (altivec_vpks<VI_char>us): Likewise.
9066 (altivec_vpku<VI_char>us): Likewise.
9067 (altivec_vpku<VI_char>um): Likewise.
9068 (altivec_vpku<VI_char>um_direct): New (copy of
9069 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9071 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9072 target is little endian and -maltivec=be is not specified.
9073 (*altivec_vupkhs<VU_char>_direct): New (copy of
9074 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9075 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9076 target is little endian and -maltivec=be is not specified.
9077 (*altivec_vupkls<VU_char>_direct): New (copy of
9078 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9079 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9080 little endian and -maltivec=be is not specified.
9081 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9082 little endian and -maltivec=be is not specified.
9084 2014-02-05 Richard Henderson <rth@redhat.com>
9087 * combine-stack-adj.c: Revert r206943.
9088 * sched-int.h (struct deps_desc): Add last_args_size.
9089 * sched-deps.c (init_deps): Initialize it.
9090 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9091 contain REG_ARGS_SIZE notes.
9093 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9095 * lto-cgraph.c (asm_nodes_output): Make global.
9096 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9097 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9098 (driver_handle_option): Handle OPT_fwpa.
9100 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9103 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9104 a comment typo and formatting issue. If odr_hash hasn't been
9105 created, return vNULL and set *completep to false.
9108 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9109 bb with no successors.
9111 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9114 * doc/invoke.texi (-march): Clarify documentation for ARM.
9118 2014-02-05 Richard Biener <rguenther@suse.de>
9120 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9121 when not vectorizing because of too many alias checks.
9122 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9123 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9125 2014-02-05 Nick Clifton <nickc@redhat.com>
9127 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9128 accept extended registers in any mode when compiling for the MN10300.
9130 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9132 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9133 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9134 sanitization attributes.
9135 (can_inline_edge_p): Likewise.
9136 (sanitize_attrs_match_for_inline_p): New function.
9138 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9140 * ipa-prop.c (detect_type_change): Shor circuit testing of
9141 type changes on THIS pointer.
9143 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9146 * config/pa/pa.c (legitimize_tls_address): Return original address
9147 if not passed a SYMBOL_REF rtx.
9148 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9150 (pa_emit_move_sequence): Simplify TLS source operands.
9151 (pa_legitimate_constant_p): Reject all TLS constants.
9152 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9153 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9155 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9157 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9158 groups when we know they are controlled by LTO.
9159 * varasm.c (default_binds_local_p_1): If object is in other partition,
9160 it will be resolved locally.
9162 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9164 * config/host-linux.c (linux_gt_pch_use_address): Don't
9165 use SSIZE_MAX because it is not always defined.
9167 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9170 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9171 threshold for pseudo splitting.
9172 (update_ebb_live_info): Process call argument hard registers and
9173 hard registers from insn definition too.
9174 (max_small_class_regs_num): New constant.
9175 (inherit_in_ebb): Update live hard regs through EBBs. Update
9176 reloads_num only for small register classes. Don't split for
9179 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9182 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9185 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9187 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9190 2014-02-04 Marek Polacek <polacek@redhat.com>
9192 * gdbinit.in (pel): Define.
9194 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9196 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9199 2014-02-04 Richard Biener <rguenther@suse.de>
9202 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9203 in function context local.
9204 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9205 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9206 similar to LTO_imported_decl_ref.
9208 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9210 PR tree-optimization/60002
9211 * cgraphclones.c (build_function_decl_skip_args): Clear
9214 PR tree-optimization/60023
9215 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9216 false to gsi_replace.
9217 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9218 has been in some EH region and vec_stmt could throw, add
9219 vec_stmt into the same EH region.
9220 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9221 has no lhs, ignore it.
9222 * internal-fn.c (expand_MASK_LOAD): Likewise.
9225 * tree-inline.c (copy_forbidden): Fail for
9226 __attribute__((optimize (0))) functions.
9229 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9230 is set, copy one less argument.
9231 (expand_simd_clones): Don't subtract clone_info->inbranch
9232 from simd_clone_struct_alloc argument.
9234 PR rtl-optimization/57915
9235 * recog.c (simplify_while_replacing): If all unary/binary/relational
9236 operation arguments are constant, attempt to simplify those.
9239 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9240 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9242 2014-02-04 Richard Biener <rguenther@suse.de>
9244 PR tree-optimization/60012
9245 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9246 TBAA disambiguation to all DDRs.
9248 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9251 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9252 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9254 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9257 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9259 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9261 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9262 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9264 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9267 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9268 to figure out targets of polymorphic calls with known decl.
9269 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9270 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9271 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9272 (get_polymorphic_call_info): ... here.
9273 (get_polymorphic_call_info_from_invariant): New function.
9275 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9277 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9278 lookup via vtable pointer; check for type consistency
9279 and turn inconsitent facts into UNREACHABLE.
9280 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9281 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9282 type inconsistent querries; return UNREACHABLE instead.
9284 2014-02-03 Richard Henderson <rth@twiddle.net>
9287 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9288 already processed this node.
9289 (normalize_one_pred_1): Pass along mark_set.
9290 (normalize_one_pred): Create and destroy a pointer_set_t.
9291 (normalize_one_pred_chain): Likewise.
9293 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9295 PR gcov-profile/58602
9296 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9298 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9301 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9302 -fno-devirtualize; try to devirtualize by the knowledge of
9303 virtual table pointer given by aggregate propagation.
9304 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9305 (ipa_print_node_jump_functions): Dump also offset that
9306 is relevant for polymorphic calls.
9307 (determine_known_aggregate_parts): Add arg_type parameter; use it
9308 instead of determining the type from pointer type.
9309 (ipa_compute_jump_functions_for_edge): Update call of
9310 determine_known_aggregate_parts.
9311 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9312 (gimple_get_virt_method_for_binfo): ... here; simplify using
9313 vtable_pointer_value_to_vtable.
9314 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9315 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9316 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9317 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9319 (vtable_pointer_value_to_binfo): ... here.
9320 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9322 2014-02-03 Teresa Johnson <tejohnson@google.com>
9324 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9325 redef of outer loop index variable.
9327 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9331 * doc/extend.texi (Function Attributes): Typo.
9333 2014-02-03 Cong Hou <congh@google.com>
9335 PR tree-optimization/60000
9336 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9337 if the vectorized statement is a store. A store statement can only
9338 appear at the end of pattern statements.
9340 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9342 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9343 (ix86_option_override_internal): Default long double to 64-bit for
9344 32-bit Bionic and to 128-bit for 64-bit Bionic.
9346 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9347 TARGET_LONG_DOUBLE_128 is true.
9348 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9350 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9351 (mlong-double-64): Negate -mlong-double-128.
9352 (mlong-double-128): New option.
9354 * config/i386/i386-c.c (ix86_target_macros): Define
9355 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9357 * doc/invoke.texi: Document -mlong-double-128.
9359 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9361 PR rtl-optimization/60024
9362 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9364 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9366 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9368 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9370 PR rtl-optimization/57662
9371 * sel-sched.c (code_motion_path_driver): Do not mark already not
9372 existing blocks in the visiting bitmap.
9374 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9376 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9377 on the insn being emitted.
9379 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9380 Will Deacon <will.deacon@arm.com>
9382 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9384 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9386 * config/arm/arm-tables.opt: Regenerate.
9388 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9390 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9391 for vector types other than V16QImode.
9392 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9393 define_expand, and call altivec_expand_vec_perm_le when producing
9394 code with little endian element order.
9395 (*altivec_vperm_<mode>_internal): New insn having previous
9396 behavior of altivec_vperm_<mode>.
9397 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9398 altivec_expand_vec_perm_le when producing code with little endian
9400 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9401 behavior of altivec_vperm_<mode>_uns.
9403 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9405 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9406 (altivec_vsumsws): Add handling for -maltivec=be with a little
9408 (altivec_vsumsws_direct): New.
9409 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9410 gen_altivec_vsumsws.
9412 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9414 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9415 vtable_pointer_value_to_binfo): New functions.
9416 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9417 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9419 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9421 * config/nios2/nios2.md (load_got_register): Initialize GOT
9422 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9423 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9425 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9427 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9428 preserverd by passthrough, do not propagate the type.
9430 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9432 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9433 (mips_atomic_assign_expand_fenv): New function.
9434 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9436 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9438 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9439 (__builtin_mips_set_fcsr): Likewise.
9440 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9441 MIPS_USI_FTYPE_VOID.
9442 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9443 (mips16_expand_set_fcsr): Likewise.
9444 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9445 (mips16_set_fcsr_stub): Likewise.
9446 (mips16_get_fcsr_one_only_stub): New class.
9447 (mips16_set_fcsr_one_only_stub): Likewise.
9448 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9449 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9450 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9451 (hard_float): New availability predicate.
9452 (mips_builtins): Add get_fcsr and set_fcsr.
9453 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9454 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9455 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9456 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9457 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9460 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9462 * config/mips/mips.c (mips_one_only_stub): New class.
9463 (mips_need_mips16_rdhwr_p): Replace with...
9464 (mips16_rdhwr_stub): ...this new variable.
9465 (mips16_stub_call_address): New function.
9466 (mips16_rdhwr_one_only_stub): New class.
9467 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9468 (mips_output_mips16_rdhwr): Delete.
9469 (mips_finish_stub): New function.
9470 (mips_code_end): Use it to handle rdhwr stubs.
9472 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9475 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9476 when calculating size of integer atomic types.
9478 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9480 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9482 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9484 PR tree-optimization/60003
9485 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9486 * profile.c (branch_prob): Use gimple_call_builtin_p
9487 to check for BUILT_IN_SETJMP_RECEIVER.
9488 * tree-inline.c (copy_bb): Call notice_special_calls.
9490 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9493 * lra-constraints.c (process_alt_operands): Update reload_sum only
9496 2014-01-31 Richard Henderson <rth@redhat.com>
9499 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9500 until after else_eh is processed.
9502 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9504 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9505 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9506 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9507 in smmintrin.h, remove them.
9508 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9509 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9510 * config/i386/i386.md (ROUND_SAE): Fix value.
9511 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9512 (const48_operand): New.
9513 * config/i386/subst.md (round), (round_expand): Use
9514 const_4_or_8_to_11_operand.
9515 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9517 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9519 * config/i386/constraints.md (Yk): Swap meaning with k.
9520 * config/i386/i386.md (movhi_internal): Change Yk to k.
9521 (movqi_internal): Ditto.
9522 (*k<logic><mode>): Ditto.
9525 (kandn<mode>): Ditto.
9526 (*<code>hi_1): Ditto.
9527 (*<code>qi_1): Ditto.
9528 (kxnor<mode>): Ditto.
9529 (kortestzhi): Ditto.
9530 (kortestchi): Ditto.
9532 (*one_cmplhi2_1): Ditto.
9533 (*one_cmplqi2_1): Ditto.
9534 * config/i386/sse.md (): Change k to Yk.
9535 (avx512f_load<mode>_mask): Ditto.
9536 (avx512f_blendm<mode>): Ditto.
9537 (avx512f_store<mode>_mask): Ditto.
9538 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9539 (avx512f_storedqu<mode>_mask): Ditto.
9540 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9542 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9543 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9544 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9545 (avx512f_maskcmp<mode>3): Ditto.
9546 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9547 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9548 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9549 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9550 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9551 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9552 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9553 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9554 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9555 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9556 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9557 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9558 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9559 (vec_extract_lo_<mode>_maskm): Ditto.
9560 (vec_extract_hi_<mode>_maskm): Ditto.
9561 (avx512f_vternlog<mode>_mask): Ditto.
9562 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9563 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9564 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9565 (avx512f_<code>v8div16qi2_mask): Ditto.
9566 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9567 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9568 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9569 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9570 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9571 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9572 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9573 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9574 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9575 (avx512cd_maskb_vec_dupv8di): Ditto.
9576 (avx512cd_maskw_vec_dupv16si): Ditto.
9577 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9578 (avx512f_vpermi2var<mode>3_mask): Ditto.
9579 (avx512f_vpermi2var<mode>3_mask): Ditto.
9580 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9581 (*avx512f_gathersi<mode>): Ditto.
9582 (*avx512f_gathersi<mode>_2): Ditto.
9583 (*avx512f_gatherdi<mode>): Ditto.
9584 (*avx512f_gatherdi<mode>_2): Ditto.
9585 (*avx512f_scattersi<mode>): Ditto.
9586 (*avx512f_scatterdi<mode>): Ditto.
9587 (avx512f_compress<mode>_mask): Ditto.
9588 (avx512f_compressstore<mode>_mask): Ditto.
9589 (avx512f_expand<mode>_mask): Ditto.
9590 * config/i386/subst.md (mask): Change k to Yk.
9591 (mask_scalar_merge): Ditto.
9594 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9596 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9598 2014-01-31 Richard Biener <rguenther@suse.de>
9601 * builtins.c (fold_builtin_memory_op): Make sure to not
9602 use a floating-point mode or a boolean or enumeral type for
9605 2014-01-30 DJ Delorie <dj@redhat.com>
9607 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9608 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9609 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9610 whenever main() has an epilogue.
9612 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9614 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9615 unused variable "field".
9616 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9617 (vsx_mergeh_<mode>): Likewise.
9618 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9619 (altivec_vmrghh): Likewise.
9620 (altivec_vmrghw): Likewise.
9621 (altivec_vmrglb): Likewise.
9622 (altivec_vmrglh): Likewise.
9623 (altivec_vmrglw): Likewise.
9624 (altivec_vspltb): Add missing uses.
9625 (altivec_vsplth): Likewise.
9626 (altivec_vspltw): Likewise.
9627 (altivec_vspltsf): Likewise.
9629 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9632 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9633 frame related instructions.
9635 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9637 PR rtl-optimization/59959
9638 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9639 any reload of register whose subreg is invalid.
9641 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9643 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9644 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9645 Add missing return type - void.
9647 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9649 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9650 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9651 remove element index adjustment for endian (now handled in vsx.md
9653 (altivec_expand_vec_perm_const): Use
9654 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9655 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9656 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9657 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9658 define_expand and a new define_insn *altivec_vspltb_internal;
9659 adjust for -maltivec=be on a little endian target.
9660 (altivec_vspltb_direct): New.
9661 (altivec_vsplth): Divide into a define_expand and a new
9662 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9663 little endian target.
9664 (altivec_vsplth_direct): New.
9665 (altivec_vspltw): Divide into a define_expand and a new
9666 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9667 little endian target.
9668 (altivec_vspltw_direct): New.
9669 (altivec_vspltsf): Divide into a define_expand and a new
9670 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9671 a little endian target.
9673 2014-01-30 Richard Biener <rguenther@suse.de>
9675 PR tree-optimization/59993
9676 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9677 can propagate form the earlier stmt and avoid the transform
9678 when the intermediate result is needed.
9680 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9682 * README.Portability: Fix typo.
9684 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9686 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9687 comparison_operator with ordered_comparison_operator.
9689 2014-01-30 Nick Clifton <nickc@redhat.com>
9691 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9692 Rename to mn10300_store_multiple_regs.
9693 * config/mn10300/mn10300.c: Likewise.
9694 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9695 store_multiple_regs.
9696 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9697 Call mn10300_store_multiple_regs.
9699 2014-01-30 Nick Clifton <nickc@redhat.com>
9700 DJ Delorie <dj@redhat.com>
9702 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9703 %fp 2 to keep registers after it properly word-aligned.
9704 (rl78_alloc_physical_registers_umul): Handle the case where both
9705 input operands are the same.
9707 2014-01-30 Richard Biener <rguenther@suse.de>
9709 PR tree-optimization/59903
9710 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9713 2014-01-30 Jason Merrill <jason@redhat.com>
9716 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9719 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9721 2014-01-30 Richard Biener <rguenther@suse.de>
9723 PR tree-optimization/59951
9724 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9726 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9729 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9730 SFmode to DFmode case.
9732 2014-01-29 DJ Delorie <dj@redhat.com>
9734 * config/msp430/msp430.opt (-minrt): New.
9735 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9737 (ENDFILE_SPEC): Likewise.
9739 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9741 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9742 (estimate_function_body_sizes): Use it.
9744 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9747 * dwarf2out.c (is_cxx_auto): New.
9748 (is_base_type): Use it.
9749 (gen_type_die_with_usage): Likewise.
9751 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9753 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9754 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9755 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9756 -maltivec=be with LE targets.
9757 (vsx_mergeh_<mode>): Likewise.
9758 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9759 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9760 (altivec_vmrghb): Replace with define_expand and new
9761 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9762 (altivec_vmrghb_direct): New define_insn.
9763 (altivec_vmrghh): Replace with define_expand and new
9764 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9765 (altivec_vmrghh_direct): New define_insn.
9766 (altivec_vmrghw): Replace with define_expand and new
9767 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9768 (altivec_vmrghw_direct): New define_insn.
9769 (*altivec_vmrghsf): Adjust for endianness.
9770 (altivec_vmrglb): Replace with define_expand and new
9771 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9772 (altivec_vmrglb_direct): New define_insn.
9773 (altivec_vmrglh): Replace with define_expand and new
9774 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9775 (altivec_vmrglh_direct): New define_insn.
9776 (altivec_vmrglw): Replace with define_expand and new
9777 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9778 (altivec_vmrglw_direct): New define_insn.
9779 (*altivec_vmrglsf): Adjust for endianness.
9780 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9781 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9782 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9783 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9784 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9785 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9786 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9787 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9789 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9791 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9792 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9795 2014-01-29 Richard Biener <rguenther@suse.de>
9797 PR tree-optimization/58742
9798 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9799 associate_pointerplus_align.
9800 (associate_pointerplus_diff): New function.
9801 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9802 and associate_pointerplus_diff.
9804 2014-01-29 Richard Biener <rguenther@suse.de>
9806 * lto-streamer.h (LTO_major_version): Bump to 3.
9807 (LTO_minor_version): Reset to 0.
9809 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9811 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9812 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9813 (arm_file_start): Generate correct asm header for armv7ve.
9814 * config/arm/bpabi.h: Add multilib support for armv7ve.
9815 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9816 and cortex-a15 to armv7ve.
9817 * config/arm/t-aprofile: Add multilib support for armv7ve.
9818 * doc/invoke.texi: Document -march=armv7ve.
9820 2014-01-29 Richard Biener <rguenther@suse.de>
9822 PR tree-optimization/58742
9823 * tree-ssa-forwprop.c (associate_plusminus): Return true
9824 if we changed sth, defer EH cleanup to ...
9825 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9826 (simplify_mult): New function.
9828 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9831 PR tree-optimization/59920
9832 * tree.c (build_common_builtin_nodes): Remove
9833 __builtin_setjmp_dispatcher initialization.
9834 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9835 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9836 instead of gsi_after_labels + manually skipping debug stmts.
9837 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9838 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9839 * tree-inline.c (copy_edges_for_bb): Remove
9840 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9841 argument. Ignore computed_goto_p stmts. Don't call
9842 make_abnormal_goto_edges. If a call might need abnormal edges
9843 for non-local gotos, see if it already has an edge to
9844 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9845 with true argument, don't do anything then, otherwise add
9846 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9847 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9849 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9850 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9851 (lower_stmt): Don't set data->calls_builtin_setjmp.
9852 (lower_builtin_setjmp): Adjust comment.
9853 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9854 * tree-cfg.c (found_computed_goto): Remove.
9855 (factor_computed_gotos): Remove.
9856 (make_goto_expr_edges): Return bool, true for computed gotos.
9857 Don't call make_abnormal_goto_edges.
9858 (build_gimple_cfg): Don't set found_computed_goto, don't call
9859 factor_computed_gotos.
9860 (computed_goto_p): No longer static.
9861 (make_blocks): Don't set found_computed_goto.
9862 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9863 (make_edges): If make_goto_expr_edges returns true, push bb
9864 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9865 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9866 vector. Record mapping between bbs and OpenMP regions if there
9867 are any, adjust make_gimple_omp_edges caller. Call
9868 handle_abnormal_edges.
9869 (make_abnormal_goto_edges): Remove.
9870 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9871 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9872 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9873 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9874 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9875 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9876 filling *region also set *region_idx to (*region)->entry->index.
9879 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9880 For REGs set ORIGINAL_REGNO.
9882 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9884 * doc/md.texi: Mention that a target shouldn't implement
9885 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9888 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9890 PR tree-optimization/59594
9891 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9892 a copy of the datarefs vector rather than the vector itself.
9894 2014-01-28 Jason Merrill <jason@redhat.com>
9897 * dwarf2out.c (auto_die): New static.
9898 (gen_type_die_with_usage): Handle C++1y 'auto'.
9899 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9902 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9905 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9906 (SPEC_X32): Likewise.
9907 (SPEC_64): Likewise.
9908 * config/i386/i386.c (ix86_option_override_internal): Turn off
9909 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9911 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9912 * config/i386/i386.h (TARGET_16BIT): New macro.
9913 (TARGET_16BIT_P): Likewise.
9914 * config/i386/i386.opt: Add m16.
9915 * doc/invoke.texi: Document -m16.
9917 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9919 PR preprocessor/59935
9920 * input.c (location_get_source_line): Bail out on when line number
9921 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9923 2014-01-28 Richard Biener <rguenther@suse.de>
9925 PR tree-optimization/58742
9926 * tree-ssa-forwprop.c (associate_plusminus): Handle
9927 pointer subtraction of the form (T)(P + A) - (T)P.
9929 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9931 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9934 2014-01-28 Richard Biener <rguenther@suse.de>
9937 2014-01-28 Richard Biener <rguenther@suse.de>
9939 PR rtl-optimization/45364
9940 PR rtl-optimization/59890
9941 * var-tracking.c (local_get_addr_clear_given_value): Handle
9942 already cleared slot.
9943 (val_reset): Handle not allocated local_get_addr_cache.
9944 (vt_find_locations): Use post-order on the inverted CFG.
9946 2014-01-28 Richard Biener <rguenther@suse.de>
9948 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9950 2014-01-28 Richard Biener <rguenther@suse.de>
9952 PR rtl-optimization/45364
9953 PR rtl-optimization/59890
9954 * var-tracking.c (local_get_addr_clear_given_value): Handle
9955 already cleared slot.
9956 (val_reset): Handle not allocated local_get_addr_cache.
9957 (vt_find_locations): Use post-order on the inverted CFG.
9959 2014-01-28 Alan Modra <amodra@gmail.com>
9961 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9962 * configure.ac <recursive call for build != host>: Define
9963 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9964 and LD_FOR_BUILD too.
9965 * configure: Regenerate.
9967 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9969 * config/i386/i386.c (get_builtin_code_for_version): Separate
9970 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9971 Broadwell from Haswell.
9973 2014-01-27 Steve Ellcey <sellcey@mips.com>
9975 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9976 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9977 * config/mips/mips.c (mips_option_override): Change setting
9979 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9980 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9981 Change from Mask to Var.
9983 2014-01-27 Jeff Law <law@redhat.com>
9985 * ipa-inline.c (inline_small_functions): Fix typo.
9987 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9989 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9990 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9991 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9992 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9993 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9994 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9995 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9996 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9997 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9998 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9999 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10000 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10001 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10002 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10003 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10004 (_mm512_storeu_epi64): Ditto.
10005 (_mm512_cmpge_epi32_mask): Ditto.
10006 (_mm512_cmpge_epu32_mask): Ditto.
10007 (_mm512_cmpge_epi64_mask): Ditto.
10008 (_mm512_cmpge_epu64_mask): Ditto.
10009 (_mm512_cmple_epi32_mask): Ditto.
10010 (_mm512_cmple_epu32_mask): Ditto.
10011 (_mm512_cmple_epi64_mask): Ditto.
10012 (_mm512_cmple_epu64_mask): Ditto.
10013 (_mm512_cmplt_epi32_mask): Ditto.
10014 (_mm512_cmplt_epu32_mask): Ditto.
10015 (_mm512_cmplt_epi64_mask): Ditto.
10016 (_mm512_cmplt_epu64_mask): Ditto.
10017 (_mm512_cmpneq_epi32_mask): Ditto.
10018 (_mm512_cmpneq_epu32_mask): Ditto.
10019 (_mm512_cmpneq_epi64_mask): Ditto.
10020 (_mm512_cmpneq_epu64_mask): Ditto.
10021 (_mm512_expand_pd): Ditto.
10022 (_mm512_expand_ps): Ditto.
10023 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10024 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10025 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10026 * config/i386/i386.c (ix86_builtins): Add
10027 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10028 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10029 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10030 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10031 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10032 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10033 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10034 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10035 IX86_BUILTIN_PMOVUSQW512_MEM.
10036 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10037 __builtin_ia32_pmovsqd512mem_mask,
10038 __builtin_ia32_pmovqd512mem_mask,
10039 __builtin_ia32_pmovusqw512mem_mask,
10040 __builtin_ia32_pmovsqw512mem_mask,
10041 __builtin_ia32_pmovqw512mem_mask,
10042 __builtin_ia32_pmovusdw512mem_mask,
10043 __builtin_ia32_pmovsdw512mem_mask,
10044 __builtin_ia32_pmovdw512mem_mask,
10045 __builtin_ia32_pmovqb512mem_mask,
10046 __builtin_ia32_pmovusqb512mem_mask,
10047 __builtin_ia32_pmovsqb512mem_mask,
10048 __builtin_ia32_pmovusdb512mem_mask,
10049 __builtin_ia32_pmovsdb512mem_mask,
10050 __builtin_ia32_pmovdb512mem_mask.
10051 (bdesc_args): Add __builtin_ia32_expanddf512,
10052 __builtin_ia32_expandsf512.
10053 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10054 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10055 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10056 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10057 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10058 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10059 (avx512f_<code>v8div16qi2_mask_store): This.
10060 (avx512f_expand<mode>): New.
10062 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10064 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10066 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10067 (_mm512_prefetch_i32scatter_pd): Ditto.
10068 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10069 (_mm512_prefetch_i64scatter_pd): Ditto.
10070 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10071 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10072 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10073 (_mm512_prefetch_i32scatter_ps): Ditto.
10074 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10075 (_mm512_prefetch_i64scatter_ps): Ditto.
10076 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10077 * config/i386/i386-builtin-types.def: Define
10078 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10079 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10080 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10081 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10082 IX86_BUILTIN_SCATTERPFQPD.
10083 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10084 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10085 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10086 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10087 __builtin_ia32_scatterpfqps.
10088 (ix86_expand_builtin): Expand new built-ins.
10089 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10090 fix memory access data type.
10091 (*avx512pf_gatherpf<mode>_mask): Ditto.
10092 (*avx512pf_gatherpf<mode>): Ditto.
10093 (avx512pf_scatterpf<mode>): Ditto.
10094 (*avx512pf_scatterpf<mode>_mask): Ditto.
10095 (*avx512pf_scatterpf<mode>): Ditto.
10096 (GATHER_SCATTER_SF_MEM_MODE): New.
10097 (avx512pf_gatherpf<mode>df): Ditto.
10098 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10099 (*avx512pf_scatterpf<mode>df): Ditto.
10101 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10104 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10105 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10108 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10110 * common/config/arm/arm-common.c
10111 (arm_rewrite_mcpu): Handle multiple names.
10113 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10115 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10117 * gimple-builder.h (create_gimple_tmp): Delete.
10119 2014-01-27 Christian Bruel <christian.bruel@st.com>
10121 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10124 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10126 * config/pa/pa.md (call): Generate indirect long calls to non-local
10127 functions when outputing 32-bit code.
10128 (call_value): Likewise except for special call to buggy powf function.
10130 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10131 portable runtime and PIC indirect calls.
10132 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10133 and PIC call sequences. Use ldo instead of blr to set return register
10134 in PIC call sequence.
10136 2014-01-25 Walter Lee <walt@tilera.com>
10138 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10139 avoid clobbering a live register.
10141 2014-01-25 Walter Lee <walt@tilera.com>
10143 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10144 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10145 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10146 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10148 2014-01-25 Walter Lee <walt@tilera.com>
10150 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10151 arguments on even registers.
10152 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10154 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10155 (BIGGEST_ALIGNMENT): Ditto.
10156 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10158 2014-01-25 Walter Lee <walt@tilera.com>
10160 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10161 insns before bundling.
10162 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10164 2014-01-25 Walter Lee <walt@tilera.com>
10166 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10167 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10168 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10170 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10172 * config/mips/constraints.md (kl): Delete.
10173 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10174 define expands, using...
10175 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10176 instructions for MIPS16.
10177 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10178 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10180 2014-01-25 Walter Lee <walt@tilera.com>
10182 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10186 2014-01-25 Walter Lee <walt@tilera.com>
10188 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10189 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10191 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10193 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10196 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10198 * print-rtl.c (in_call_function_usage): New var.
10199 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10200 EXPR_LIST mode as mode and not as reg note name.
10202 PR middle-end/59561
10203 * cfgloopmanip.c (copy_loop_info): If
10204 loop->warned_aggressive_loop_optimizations, make sure
10205 the flag is set in target loop too.
10207 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10209 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10211 * builtins.def: Likewise.
10212 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10213 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10214 * ira.c (ira_setup_eliminable_regset): Likewise.
10215 * omp-low.c (gate_expand_omp): Likewise.
10216 (execute_lower_omp): Likewise.
10217 (diagnose_sb_0): Likewise.
10218 (gate_diagnose_omp_blocks): Likewise.
10219 (simd_clone_clauses_extract): Likewise.
10222 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10224 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10225 correction for little endian...
10226 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10229 2014-01-24 Jeff Law <law@redhat.com>
10231 PR tree-optimization/59919
10232 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10233 for non-returning calls.
10235 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10237 * common/config/aarch64/aarch64-common.c
10238 (aarch64_rewrite_mcpu): Handle multiple names.
10239 * config/aarch64/aarch64.h
10240 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10242 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10244 * input.c (add_file_to_cache_tab): Handle the case where fopen
10247 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10250 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10251 from push operand if code of push isn't PRE_DEC.
10253 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10256 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10257 -mquad-memory-atomic. Update -mquad-memory documentation to say
10258 it is only used for non-atomic loads/stores.
10260 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10261 -mquad-memory or -mquad-memory-atomic switches.
10263 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10264 -mquad-memory-atomic to ISA 2.07 support.
10266 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10267 to separate support of normal quad word memory operations (ldq, stq)
10268 from the atomic quad word memory operations.
10270 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10271 support to separate non-atomic quad word operations from atomic
10272 quad word operations. Disable non-atomic quad word operations in
10273 little endian mode so that we don't have to swap words after the
10274 load and before the store.
10275 (quad_load_store_p): Add comment about atomic quad word support.
10276 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10277 options printed with -mdebug=reg.
10279 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10280 -mquad-memory-atomic as the test for whether we have quad word
10281 atomic instructions.
10282 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10283 or -mp8-vector are used, allow byte/half-word atomic operations.
10285 * config/rs6000/sync.md (load_lockedti): Insure that the address
10286 is a proper indexed or indirect address for the lqarx instruction.
10287 On little endian systems, swap the hi/lo registers after the lqarx
10289 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10290 insure the address is valid for the lqarx instruction.
10291 (store_conditionalti): Insure that the address is a proper indexed
10292 or indirect address for the stqcrx. instruction. On little endian
10293 systems, swap the hi/lo registers before doing the stqcrx.
10295 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10296 insure the address is valid for the stqcrx. instruction.
10298 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10299 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10300 type of quad memory support is available.
10302 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10304 PR regression/59915
10305 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10306 there is a danger of looping.
10308 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10310 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10311 force flag_ira_loop_pressure if set via command line.
10313 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10315 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10316 (ashr_simd): New builtin handling DI mode.
10317 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10318 (aarch64_sshr_simddi): New match pattern.
10319 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10320 (vshrd_n_s64): Likewise.
10321 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10323 2014-01-23 Nick Clifton <nickc@redhat.com>
10325 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10326 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10327 favour of mcu specific scripts.
10328 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10331 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10332 Alex Velenko <Alex.Velenko@arm.com>
10334 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10335 (vaddv_s16): Likewise.
10336 (vaddv_s32): Likewise.
10337 (vaddv_u8): Likewise.
10338 (vaddv_u16): Likewise.
10339 (vaddv_u32): Likewise.
10340 (vaddvq_s8): Likewise.
10341 (vaddvq_s16): Likewise.
10342 (vaddvq_s32): Likewise.
10343 (vaddvq_s64): Likewise.
10344 (vaddvq_u8): Likewise.
10345 (vaddvq_u16): Likewise.
10346 (vaddvq_u32): Likewise.
10347 (vaddvq_u64): Likewise.
10348 (vaddv_f32): Likewise.
10349 (vaddvq_f32): Likewise.
10350 (vaddvq_f64): Likewise.
10351 (vmaxv_f32): Likewise.
10352 (vmaxv_s8): Likewise.
10353 (vmaxv_s16): Likewise.
10354 (vmaxv_s32): Likewise.
10355 (vmaxv_u8): Likewise.
10356 (vmaxv_u16): Likewise.
10357 (vmaxv_u32): Likewise.
10358 (vmaxvq_f32): Likewise.
10359 (vmaxvq_f64): Likewise.
10360 (vmaxvq_s8): Likewise.
10361 (vmaxvq_s16): Likewise.
10362 (vmaxvq_s32): Likewise.
10363 (vmaxvq_u8): Likewise.
10364 (vmaxvq_u16): Likewise.
10365 (vmaxvq_u32): Likewise.
10366 (vmaxnmv_f32): Likewise.
10367 (vmaxnmvq_f32): Likewise.
10368 (vmaxnmvq_f64): Likewise.
10369 (vminv_f32): Likewise.
10370 (vminv_s8): Likewise.
10371 (vminv_s16): Likewise.
10372 (vminv_s32): Likewise.
10373 (vminv_u8): Likewise.
10374 (vminv_u16): Likewise.
10375 (vminv_u32): Likewise.
10376 (vminvq_f32): Likewise.
10377 (vminvq_f64): Likewise.
10378 (vminvq_s8): Likewise.
10379 (vminvq_s16): Likewise.
10380 (vminvq_s32): Likewise.
10381 (vminvq_u8): Likewise.
10382 (vminvq_u16): Likewise.
10383 (vminvq_u32): Likewise.
10384 (vminnmv_f32): Likewise.
10385 (vminnmvq_f32): Likewise.
10386 (vminnmvq_f64): Likewise.
10388 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10390 * config/aarch64/aarch64-simd.md
10391 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10392 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10393 (*aarch64_mul3_elt<mode>): Likewise.
10394 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10395 (*aarch64_mul3_elt_to_64v2df): Likewise.
10396 (*aarch64_mla_elt<mode>): Likewise.
10397 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10398 (*aarch64_mls_elt<mode>): Likewise.
10399 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10400 (*aarch64_fma4_elt<mode>): Likewise.
10401 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10402 (*aarch64_fma4_elt_to_64v2df): Likewise.
10403 (*aarch64_fnma4_elt<mode>): Likewise.
10404 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10405 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10406 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10407 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10408 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10409 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10410 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10411 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10412 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10414 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10416 * config/aarch64/aarch64-simd.md
10417 (aarch64_be_checked_get_lane<mode>): New define_expand.
10418 * config/aarch64/aarch64-simd-builtins.def
10419 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10420 New builtin definition.
10421 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10422 Use new safe be builtin.
10424 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10426 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10428 (aarch64_be_st1<mode>): Likewise.
10429 (aarch_ld1<VALL:mode>): Define_expand modified.
10430 (aarch_st1<VALL:mode>): Likewise.
10431 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10432 (UNSPEC_ST1): Likewise.
10434 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10436 * config/microblaze/microblaze.md: Add trap insn and attribute
10438 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10440 PR preprocessor/58580
10441 * input.h (location_get_source_line): Take an additional line_size
10443 (void diagnostics_file_cache_fini): Declare new function.
10444 * input.c (struct fcache): New type.
10445 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10446 New static constants.
10447 (diagnostic_file_cache_init, total_lines_num)
10448 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10449 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10450 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10451 (get_next_line, read_next_line, goto_next_line, read_line_num):
10452 New static function definitions.
10453 (diagnostic_file_cache_fini): New function.
10454 (location_get_source_line): Take an additional output line_len
10455 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10457 * diagnostic.c (diagnostic_finish): Call
10458 diagnostic_file_cache_fini.
10459 (adjust_line): Take an additional input parameter for the length
10460 of the line, rather than calculating it with strlen.
10461 (diagnostic_show_locus): Adjust the use of
10462 location_get_source_line and adjust_line with respect to their new
10463 signature. While displaying a line now, do not stop at the first
10464 null byte. Rather, display the zero byte as a space and keep
10465 going until we reach the size of the line.
10466 * Makefile.in: Add vec.o to OBJS-libcommon
10468 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10469 Ilya Tocar <ilya.tocar@intel.com>
10471 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10472 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10473 (__builtin_ia32_kmov16): Ditto.
10474 * config/i386/i386.md (UNSPEC_KMOV): New.
10477 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10479 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10480 (_mm512_storeu_si512): Ditto.
10482 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10485 * rtl.h (get_referenced_operands): Declare.
10486 * recog.c (get_referenced_operands): New function.
10487 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10488 operands have been referenced when recording LO_SUM references.
10490 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10492 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10494 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10496 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10497 Enable for generic and recent AMD targets.
10499 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10501 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10502 ARG_SIZE note when adjustment was eliminated.
10504 2014-01-22 Jeff Law <law@redhat.com>
10506 PR tree-optimization/59597
10507 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10508 in file. Accept new argument REGISTERING and use it to modify
10509 dump output appropriately.
10510 (register_jump_thread): Corresponding changes.
10511 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10512 thread paths involving joiner blocks. Add code to dump cancelled
10513 jump threading paths.
10515 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10517 PR rtl-optimization/59477
10518 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10519 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10521 2014-01-22 Tom Tromey <tromey@redhat.com>
10523 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10525 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10527 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10529 PR rtl-optimization/59896
10530 * lra-constraints.c (process_alt_operands): Check unused note for
10531 matched operands of insn with no output reloads.
10533 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10535 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10536 (mips_move_from_gpr_cost): Likewise.
10538 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10540 PR rtl-optimization/59858
10541 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10542 ira_class_hard_regs_num.
10543 (process_alt_operands): Increase reject for dying matched operand.
10545 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10548 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10549 smaller than size, perform several stores or loads and stores
10550 at dst + count - size to store or copy all of size bytes, rather
10551 than just last modesize bytes.
10553 2014-01-20 DJ Delorie <dj@redhat.com>
10555 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10556 that CLOBBERs are REGs before propogating their values.
10558 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10560 PR middle-end/59789
10561 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10562 (cgraph_inline_failed_type): New function.
10563 * cgraph.h (DEFCIFCODE): Add type.
10564 (cgraph_inline_failed_type_t): New enum.
10565 (cgraph_inline_failed_type): New prototype.
10566 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10567 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10568 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10569 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10570 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10571 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10572 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10573 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10574 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10575 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10576 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10577 OPTIMIZATION_MISMATCH.
10578 * tree-inline.c (expand_call_inline): Emit errors during
10579 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10581 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10584 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10585 mode attribute in insn output.
10587 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10589 * output.h (output_constant): Delete.
10590 * varasm.c (output_constant): Make private.
10592 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10594 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10596 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10598 PR middle-end/59860
10599 * tree.h (fold_builtin_strcat): New prototype.
10600 * builtins.c (fold_builtin_strcat): No longer static. Add len
10601 argument, if non-NULL, don't call c_strlen. Optimize
10602 directly into __builtin_memcpy instead of __builtin_strcpy.
10603 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10604 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10606 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10608 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10609 for SImode_address_operand operands, having only a REG argument.
10611 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10613 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10614 loader name using mbig-endian.
10615 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10617 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10619 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10620 (-mtune): Likewise.
10623 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10625 * config/aarch64/aarch64-protos.h
10626 (aarch64_cannot_change_mode_class_ptr): Declare.
10627 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10628 aarch64_cannot_change_mode_class_ptr): New.
10629 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10630 backend hook aarch64_cannot_change_mode_class.
10632 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10634 * common/config/aarch64/aarch64-common.c
10635 (aarch64_handle_option): Don't handle any option order logic here.
10636 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10637 selected_cpu, warn on architecture version mismatch.
10638 (aarch64_override_options): Fix parsing order for option strings.
10640 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10641 Iain Sandoe <iain@codesourcery.com>
10644 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10645 warning. Amend comment to reflect current functionality.
10647 2014-01-20 Richard Biener <rguenther@suse.de>
10649 PR middle-end/59860
10650 * builtins.c (fold_builtin_strcat): Remove case better handled
10651 by tree-ssa-strlen.c.
10653 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10655 * config/aarch64/aarch64.opt
10656 (mcpu, march, mtune): Make case-insensitive.
10658 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10661 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10662 if operands[1] is a REG or ZERO_EXTEND of a REG.
10664 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10666 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10668 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10670 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10671 long non-pic millicode calls.
10673 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10675 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10677 2014-01-19 Kito Cheng <kito@0xlab.org>
10679 * builtins.c (expand_movstr): Check movstr expand done or fail.
10681 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10682 H.J. Lu <hongjiu.lu@intel.com>
10685 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10686 to DImode for zero-extended addresses.
10688 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10690 PR rtl-optimization/57763
10691 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10692 on the new indirect jump_insn and increment LABEL_NUSES (label).
10694 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10698 * config.gcc (x86_archs): New variable.
10699 (x86_64_archs): Likewise.
10700 (x86_cpus): Likewise.
10701 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10702 --with-arch/--with-cpu= options.
10703 Support --with-arch=/--with-cpu={nehalem,westmere,
10704 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10706 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10708 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10709 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10711 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10713 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10715 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10718 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10719 clear cpp_get_options (parse_in)->warn_unused_macros for
10720 ix86_target_macros_internal with cpp_define.
10722 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10724 * jump.c (delete_related_insns): Keep (use (insn))s.
10725 * reorg.c (redundant_insn): Check for barriers too.
10727 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10729 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10731 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10733 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10734 call to $$dyncall when TARGET_LONG_CALLS is true.
10736 2014-01-17 Jeff Law <law@redhat.com>
10738 * ree.c (combine_set_extension): Temporarily disable test for
10739 changing number of hard registers.
10741 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10743 PR middle-end/58125
10744 * ipa-inline-analysis.c (inline_free_summary):
10745 Do not free summary of aliases.
10747 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10749 PR middle-end/59706
10750 * gimplify.c (gimplify_expr): Use create_tmp_var
10751 instead of create_tmp_var_raw. If cond doesn't have
10752 integral type, don't add the IFN_ANNOTATE builtin at all.
10754 2014-01-17 Martin Jambor <mjambor@suse.cz>
10757 * ipa-cp.c (prev_edge_clone): New variable.
10758 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10759 Also resize prev_edge_clone vector.
10760 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10761 (ipcp_edge_removal_hook): New function.
10762 (ipcp_driver): Register ipcp_edge_removal_hook.
10764 2014-01-17 Andrew Pinski <apinski@cavium.com>
10765 Steve Ellcey <sellcey@mips.com>
10768 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10771 2014-01-17 Jeff Law <law@redhat.com>
10773 PR middle-end/57904
10774 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10775 so that pass_ccp runs first.
10777 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10779 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10780 (ix86_adjust_cost): Use !TARGET_XXX.
10781 (do_reorder_for_imul): Likewise.
10782 (swap_top_of_ready_list): Likewise.
10783 (ix86_sched_reorder): Likewise.
10785 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10787 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10788 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10789 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10790 (intel_memset): New. Duplicate slm_memset.
10791 (intel_cost): New. Duplicate slm_cost.
10792 (m_INTEL): New macro.
10793 (processor_target_table): Add "intel".
10794 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10795 with PROCESSOR_INTEL for "intel".
10796 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10797 PROCESSOR_SILVERMONT.
10798 (ix86_issue_rate): Likewise.
10799 (ix86_adjust_cost): Likewise.
10800 (ia32_multipass_dfa_lookahead): Likewise.
10801 (swap_top_of_ready_list): Likewise.
10802 (ix86_sched_reorder): Likewise.
10803 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10804 instead of TARGET_OPT_AGU.
10805 * config/i386/i386.h (TARGET_INTEL): New.
10806 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10807 (processor_type): Add PROCESSOR_INTEL.
10808 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10809 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10811 2014-01-17 Marek Polacek <polacek@redhat.com>
10814 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10817 2014-01-17 Richard Biener <rguenther@suse.de>
10819 PR tree-optimization/46590
10820 * opts.c (default_options_table): Add entries for
10821 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10822 all enabled at -O1 but not for -Og.
10823 * common.opt (fbranch-count-reg): Remove Init(1).
10824 (fmove-loop-invariants): Likewise.
10825 (ftree-pta): Likewise.
10827 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10829 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10830 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10831 decls to at least the GCC 4.8 used alignments.
10834 * tree-nested.c (convert_nonlocal_reference_stmt,
10835 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10836 of GIMPLE_BIND stmts, adjust associated decls.
10838 2014-01-17 Richard Biener <rguenther@suse.de>
10840 PR tree-optimization/46590
10841 * vec.h (vec<>::bseach): New member function implementing
10842 binary search according to C89 bsearch.
10843 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10844 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10845 bitmap pointer again. Make accesses_in_loop a flat array.
10846 (mem_ref_obstack): New global.
10847 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10848 (mark_ref_stored): Likewise.
10849 (ref_indep_loop_p_2): Likewise.
10850 (set_ref_stored_in_loop): New helper function.
10851 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10852 (memref_free): Adjust.
10853 (record_mem_ref_loc): Simplify.
10854 (gather_mem_refs_stmt): Adjust.
10855 (sort_locs_in_loop_postorder_cmp): New function.
10856 (analyze_memory_references): Sort accesses_in_loop after
10857 loop postorder number.
10858 (find_ref_loc_in_loop_cmp): New function.
10859 (for_all_locs_in_loop): Find relevant cluster of locs in
10860 accesses_in_loop and iterate without recursion.
10861 (execute_sm): Avoid uninit warning.
10862 (struct ref_always_accessed): Simplify.
10863 (ref_always_accessed::operator ()): Likewise.
10864 (ref_always_accessed_p): Likewise.
10865 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10866 loop postorder numbers here.
10867 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10870 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10873 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10874 on decls for which assemble_alias has been called.
10876 2014-01-17 Nick Clifton <nickc@redhat.com>
10878 * config/msp430/msp430.opt: (mcpu): New option.
10879 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10880 (msp430_option_override): Parse target_cpu. If the MCU name
10881 matches a generic string, clear target_mcu.
10882 (msp430_attr): Allow numeric interrupt values up to 63.
10883 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10884 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10886 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10888 * config/msp430/msp430.md (popm): Use %J rather than %I.
10889 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10890 (addhi_cy_i): Use immediate_operand for operand 2.
10891 * doc/invoke.texi: Document -mcpu option.
10893 2014-01-17 Richard Biener <rguenther@suse.de>
10895 PR rtl-optimization/38518
10896 * df.h (df_analyze_loop): Declare.
10897 * df-core.c: Include cfgloop.h.
10898 (df_analyze_1): Split out main part of df_analyze.
10899 (df_analyze): Adjust.
10900 (loop_inverted_post_order_compute): New function.
10901 (loop_post_order_compute): Likewise.
10902 (df_analyze_loop): New function avoiding whole-function
10903 postorder computes.
10904 * loop-invariant.c (find_defs): Use df_analyze_loop.
10905 (find_invariants): Adjust.
10906 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10908 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10910 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10911 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10913 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10915 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10916 traversal when removing references.
10918 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10921 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10923 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10925 PR middle-end/56791
10926 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10927 pushing a reload for an autoinc when we had previously reloaded an
10928 inner part of the address.
10930 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10932 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10934 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10935 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10936 when not giving up or versioning for alias only because of
10938 (vect_analyze_data_ref_dependences): Set to true.
10939 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10941 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10942 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10945 PR middle-end/58344
10946 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10949 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10950 operand 0 predicate for gathers, use a new pseudo as subtarget.
10952 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10954 PR middle-end/59609
10955 * lra-constraints.c (process_alt_operands): Add printing debug info.
10956 Check absence of input/output reloads for matched operands too.
10958 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10960 PR rtl-optimization/59835
10961 * ira.c (ira_init_register_move_cost): Increase cost for
10964 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10966 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10968 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10971 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10972 non-register objects. Use gen_(high/low)part more consistently.
10975 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10978 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10979 endian support, remove tests for WORDS_BIG_ENDIAN.
10980 (p8_mfvsrd_3_<mode>): Likewise.
10981 (reload_gpr_from_vsx<mode>): Likewise.
10982 (reload_gpr_from_vsxsf): Likewise.
10983 (p8_mfvsrd_4_disf): Likewise.
10985 2014-01-16 Richard Biener <rguenther@suse.de>
10987 PR rtl-optimization/46590
10988 * lcm.c (compute_antinout_edge): Use postorder iteration.
10989 (compute_laterin): Use inverted postorder iteration.
10991 2014-01-16 Nick Clifton <nickc@redhat.com>
10993 PR middle-end/28865
10994 * varasm.c (output_constant): Return the number of bytes actually
10996 (output_constructor_array_range): Update the field size with the
10997 number of bytes emitted by output_constant.
10998 (output_constructor_regular_field): Likewise. Also do not
10999 complain if the total number of bytes emitted is now greater
11000 than the expected fieldpos.
11001 * output.h (output_constant): Update prototype and descriptive comment.
11003 2014-01-16 Marek Polacek <polacek@redhat.com>
11005 PR middle-end/59827
11006 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11007 it is error_mark_node.
11009 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11011 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11012 VALID_AVX256_REG_OR_OI_MODE.
11014 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11016 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11017 current procedure should be profiled.
11019 2014-01-15 Andrew Pinski <apinski@cavium.com>
11021 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11022 of moving from/to the STACK_REG register class.
11024 2014-01-15 Richard Henderson <rth@redhat.com>
11027 * reginfo.c (global_regs_decl): Globalize.
11028 * rtl.h (global_regs_decl): Declare.
11029 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11030 reserved via global_regs.
11032 2014-01-15 Teresa Johnson <tejohnson@google.com>
11034 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11036 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11038 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11039 and vmulosh rather than call gen_vec_widen_smult_*.
11040 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11041 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11042 (vec_widen_smult_even_v16qi): Likewise.
11043 (vec_widen_umult_even_v8hi): Likewise.
11044 (vec_widen_smult_even_v8hi): Likewise.
11045 (vec_widen_umult_odd_v16qi): Likewise.
11046 (vec_widen_smult_odd_v16qi): Likewise.
11047 (vec_widen_umult_odd_v8hi): Likewise.
11048 (vec_widen_smult_odd_v8hi): Likewise.
11049 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11050 vmuloub rather than call gen_vec_widen_umult_*.
11051 (vec_widen_umult_lo_v16qi): Likewise.
11052 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11053 vmulosb rather than call gen_vec_widen_smult_*.
11054 (vec_widen_smult_lo_v16qi): Likewise.
11055 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11056 rather than call gen_vec_widen_umult_*.
11057 (vec_widen_umult_lo_v8hi): Likewise.
11058 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11059 rather than call gen_vec_widen_smult_*.
11060 (vec_widen_smult_lo_v8hi): Likewise.
11062 2014-01-15 Jeff Law <law@redhat.com>
11064 PR tree-optimization/59747
11065 * ree.c (find_and_remove_re): Properly handle case where a second
11066 eliminated extension requires widening a copy created for elimination
11067 of a prior extension.
11068 (combine_set_extension): Ensure that the number of hard regs needed
11069 for a destination register does not change when we widen it.
11071 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11073 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11074 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11075 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11076 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11077 (avr-*-rtems*): Likewise.
11078 (bfin*-rtems*): Likewise.
11079 (moxie-*-rtems*): Likewise.
11080 (h8300-*-rtems*): Likewise.
11081 (i[34567]86-*-rtems*): Likewise.
11082 (lm32-*-rtems*): Likewise.
11083 (m32r-*-rtems*): Likewise.
11084 (m68k-*-rtems*): Likewise.
11085 (microblaze*-*-rtems*): Likewise.
11086 (mips*-*-rtems*): Likewise.
11087 (powerpc-*-rtems*): Likewise.
11088 (sh-*-rtems*): Likewise.
11089 (sparc-*-rtems*): Likewise.
11090 (sparc64-*-rtems*): Likewise.
11091 (v850-*-rtems*): Likewise.
11092 (m32c-*-rtems*): Likewise.
11094 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11096 PR rtl-optimization/59511
11097 * ira.c (ira_init_register_move_cost): Use memory costs for some
11098 cases of register move cost calculations.
11099 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11100 instead of BB frequency.
11101 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11102 * lra-assigns.c (find_hard_regno_for): Ditto.
11104 2014-01-15 Richard Biener <rguenther@suse.de>
11106 PR tree-optimization/59822
11107 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11108 (vectorizable_load): Use it to hoist defs of uses of invariant
11109 loads out of the loop.
11111 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11112 Kugan Vivekanandarajah <kuganv@linaro.org>
11115 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11118 2014-01-15 Richard Biener <rguenther@suse.de>
11120 PR rtl-optimization/59802
11121 * lcm.c (compute_available): Use inverted postorder to seed
11122 the initial worklist.
11124 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11127 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11128 ADDR_REGS for invalid symrefs in non-PIC code.
11130 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11133 * builtins.c (determine_block_size): Initialize *probable_max_size
11134 even if len_rtx is CONST_INT.
11136 2014-01-14 Andrew Pinski <apinski@cavium.com>
11138 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11139 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11140 (cortexa53_tunings): Likewise.
11141 (aarch64_sched_issue_rate): New function.
11142 (TARGET_SCHED_ISSUE_RATE): Define.
11144 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11146 * ira-costs.c (find_costs_and_classes): Add missed
11147 ira_init_register_move_cost_if_necessary.
11149 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11152 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11154 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11157 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11158 to indicate if type is used for function return value. Warn ABI
11159 change if the vector mode isn't available for function return value.
11160 (ix86_function_arg_advance): Pass false to type_natural_mode.
11161 (ix86_function_arg): Likewise.
11162 (ix86_gimplify_va_arg): Likewise.
11163 (function_arg_32): Don't warn ABI change.
11164 (ix86_function_value): Pass true to type_natural_mode.
11165 (ix86_return_in_memory): Likewise.
11166 (ix86_struct_value_rtx): Removed.
11167 (TARGET_STRUCT_VALUE_RTX): Likewise.
11169 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11171 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11172 converting a conditional jump into a conditional return.
11174 2014-01-14 Richard Biener <rguenther@suse.de>
11176 PR tree-optimization/58921
11177 PR tree-optimization/59006
11178 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11179 hoisting invariant stmts.
11180 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11181 invariant loads on the preheader edge if possible.
11183 2014-01-14 Joey Ye <joey.ye@arm.com>
11185 * doc/plugin.texi (Building GCC plugins): Update to C++.
11187 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11189 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11190 (_mm_rcp28_round_ss): Ditto.
11191 (_mm_rsqrt28_round_sd): Ditto.
11192 (_mm_rsqrt28_round_ss): Ditto.
11193 (_mm_rcp28_sd): Ditto.
11194 (_mm_rcp28_ss): Ditto.
11195 (_mm_rsqrt28_sd): Ditto.
11196 (_mm_rsqrt28_ss): Ditto.
11197 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11198 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11199 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11200 (IX86_BUILTIN_RCP28SD): Ditto.
11201 (IX86_BUILTIN_RCP28SS): Ditto.
11202 (IX86_BUILTIN_RSQRT28SD): Ditto.
11203 (IX86_BUILTIN_RSQRT28SS): Ditto.
11204 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11205 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11206 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11207 (ix86_expand_special_args_builtin): Expand new FTYPE.
11208 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11209 (srcp14<mode>): Make insn unary.
11210 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11211 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11212 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11213 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11214 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11215 Fix rounding: make it SAE only.
11216 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11218 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11220 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11221 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11222 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11223 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11224 (round_saeonly_mask_scalar_operand4): Ditto.
11225 (round_saeonly_mask_scalar_op3): Ditto.
11226 (round_saeonly_mask_scalar_op4): Ditto.
11228 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11230 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11231 Implement -maltivec=be for vec_insert and vec_extract.
11233 2014-01-10 DJ Delorie <dj@redhat.com>
11235 * config/msp430/msp430.md (call_internal): Don't allow memory
11236 references with SP as the base register.
11237 (call_value_internal): Likewise.
11238 * config/msp430/constraints.md (Yc): New. For memory references
11239 that don't use SP as a base register.
11241 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11242 "an integer without a # prefix"
11243 * config/msp430/msp430.md (epilogue_helper): Use it.
11245 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11248 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11249 AVX512F gather builtins.
11250 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11251 on gather decls with INTEGER_TYPE masktype.
11252 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11253 directly into the builtin rather than hoisting it before loop.
11255 PR tree-optimization/59387
11256 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11257 (scev_const_prop): If folded_casts and type has undefined overflow,
11258 use force_gimple_operand instead of force_gimple_operand_gsi and
11259 for each added stmt if it is assign with
11260 arith_code_with_undefined_signed_overflow, call
11261 rewrite_to_defined_overflow.
11262 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11263 gimple-fold.h instead.
11264 (arith_code_with_undefined_signed_overflow,
11265 rewrite_to_defined_overflow): Moved to ...
11266 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11267 rewrite_to_defined_overflow): ... here. No longer static.
11268 Include gimplify-me.h.
11269 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11270 rewrite_to_defined_overflow): New prototypes.
11272 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11274 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11276 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11278 * builtins.c (get_object_alignment_2): Minor tweak.
11279 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11281 2014-01-13 Christian Bruel <christian.bruel@st.com>
11283 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11284 optimized non constant lengths.
11286 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11289 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11290 load as __atomic_load_N if possible.
11292 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11294 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11296 (rs6000_expand_builtin): Adjust call.
11298 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11301 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11302 * config/rs6000/rs6000.c: Include target-globals.h.
11303 (rs6000_set_current_function): Instead of doing target_reinit
11304 unconditionally, use save_target_globals_default_opts and
11305 restore_target_globals.
11307 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11309 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11310 (rs6000_expand_builtin): Handle mffs and mtfsf.
11311 (rs6000_init_builtins): Define mffs and mtfsf.
11312 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11313 (rs6000_mffs): New pattern.
11314 (rs6000_mtfsf): New pattern.
11316 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11318 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11319 Start narrowing with START. Apply candidate-use pair
11320 and check overall cost in narrowing.
11321 (iv_ca_prune): Pass new argument.
11323 2014-01-10 Jeff Law <law@redhat.com>
11325 PR middle-end/59743
11326 * ree.c (combine_reaching_defs): Ensure the defining statement
11327 occurs before the extension when optimizing extensions with
11328 different source and destination hard registers.
11330 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11333 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11334 vtables into the type inheritance graph.
11336 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11338 PR rtl-optimization/59754
11339 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11340 modes in the REGNO != REGNO case.
11342 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11344 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11346 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11348 PR tree-optimization/59745
11349 * tree-predcom.c (tree_predictive_commoning_loop): Call
11350 free_affine_expand_cache if giving up because components is NULL.
11352 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11353 GC in payload of target_globals struct instead of allocating them on
11354 the heap and the larger structs separately using GC.
11355 * target-globals.h (struct target_globals): Make regs, hard_regs,
11356 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11357 of GTY((skip)) and change type to void *.
11358 (reset_target_globals): Cast loads from those fields to corresponding
11361 2014-01-10 Steve Ellcey <sellcey@mips.com>
11364 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11365 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11366 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11368 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11371 * aarch64-modes.def (CC_Zmode): New flags mode.
11372 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11373 represents an equality.
11374 (aarch64_get_condition_code): Handle CC_Zmode.
11375 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11377 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11379 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11380 extraction in good case.
11382 2014-01-10 Richard Biener <rguenther@suse.de>
11384 PR tree-optimization/59374
11385 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11386 checking after SLP discovery. Mark stmts not participating
11387 in any SLP instance properly.
11389 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11391 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11392 when handling a SET rtx.
11394 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11396 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11397 (cortex-a57): Likewise.
11398 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11400 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11402 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11403 non-iwmmxt builtins.
11405 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11409 * ipa-devirt.c record_target_from_binfo): Take as argument
11410 stack of binfos and lookup matching one for virtual inheritance.
11411 (possible_polymorphic_call_targets_1): Update.
11413 2014-01-10 Huacai Chen <chenhc@lemote.com>
11415 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11416 kernel strings for Loongson-2E/2F/3A.
11418 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11420 PR middle-end/59670
11421 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11422 is_gimple_call before calling gimple_call_internal_p.
11424 2014-01-09 Steve Ellcey <sellcey@mips.com>
11426 * Makefile.in (TREE_FLOW_H): Remove.
11427 (TREE_SSA_H): Add file names from tree-flow.h.
11428 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11429 * tree.h: Remove tree-flow.h reference.
11430 * hash-table.h: Remove tree-flow.h reference.
11431 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11432 reference with tree-ssa-loop.h.
11434 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11436 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11437 default element-order behavior for -maltivec.
11438 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11439 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11440 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11441 when targeting big endian, at least for now.
11442 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11444 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11446 PR middle-end/47735
11447 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11448 var satisfies use_register_for_decl, just take into account type
11449 alignment, rather than decl alignment.
11451 PR tree-optimization/59622
11452 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11453 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11454 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11455 Don't devirtualize for inplace at all. For targets.length () == 1,
11456 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11458 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11460 * config/i386/i386.md (cpu): Remove the unused btver1.
11462 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11464 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11466 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11469 * tree-core.h (struct target_globals): New forward declaration.
11470 (struct tree_target_option): Add globals field.
11471 * tree.h (TREE_TARGET_GLOBALS): Define.
11472 (prepare_target_option_nodes_for_pch): New prototype.
11473 * target-globals.h (struct target_globals): Define even if
11474 !SWITCHABLE_TARGET.
11475 * tree.c (prepare_target_option_node_for_pch,
11476 prepare_target_option_nodes_for_pch): New functions.
11477 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11478 * config/i386/i386.c: Include target-globals.h.
11479 (ix86_set_current_function): Instead of doing target_reinit
11480 unconditionally, use save_target_globals_default_opts and
11481 restore_target_globals.
11483 2014-01-09 Richard Biener <rguenther@suse.de>
11485 PR tree-optimization/59715
11486 * tree-cfg.h (split_critical_edges): Declare.
11487 * tree-cfg.c (split_critical_edges): Export.
11488 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11490 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11492 * cfgexpand.c (expand_stack_vars): Optionally disable
11493 asan stack protection.
11494 (expand_used_vars): Likewise.
11495 (partition_stack_vars): Likewise.
11496 * asan.c (asan_emit_stack_protection): Optionally disable
11497 after return stack usage.
11498 (instrument_derefs): Optionally disable memory access instrumentation.
11499 (instrument_builtin_call): Likewise.
11500 (instrument_strlen_call): Likewise.
11501 (asan_protect_global): Optionally disable global variables protection.
11502 * doc/invoke.texi: Added doc for new options.
11503 * params.def: Added new options.
11504 * params.h: Likewise.
11506 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11508 PR rtl-optimization/59724
11509 * ifcvt.c (cond_exec_process_if_block): Don't call
11510 flow_find_head_matching_sequence with 0 longest_match.
11511 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11512 non-active insns if !stop_after.
11513 (try_head_merge_bb): Revert 2014-01-07 changes.
11515 2014-01-08 Jeff Law <law@redhat.com>
11517 * ree.c (get_sub_rtx): New function, extracted from...
11518 (merge_def_and_ext): Here.
11519 (combine_reaching_defs): Use get_sub_rtx.
11521 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11523 * cgraph.h (varpool_variable_node): Do not choke on null node.
11525 2014-01-08 Catherine Moore <clm@codesourcery.com>
11527 * config/mips/mips.md (simple_return): Attempt to use JRC
11529 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11531 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11533 PR rtl-optimization/59137
11534 * reorg.c (steal_delay_list_from_target): Call update_block for
11536 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11538 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11540 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11541 two duplicate entries.
11543 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11546 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11548 * config/mips/mips.c (mips_truncated_op_cost): New function.
11549 (mips_rtx_costs): Adjust test for BADDU.
11550 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11552 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11554 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11555 (*baddu_si): ...this new pattern.
11557 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11560 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11562 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11564 PR middle-end/57748
11565 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11567 (expand_expr, expand_normal): Adjust.
11568 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11569 inner_reference_p. Use inner_reference_p to expand inner references.
11570 (store_expr): Adjust.
11571 * cfgexpand.c (expand_call_stmt): Adjust.
11573 2014-01-08 Rong Xu <xur@google.com>
11575 * gcov-io.c (gcov_var): Move from gcov-io.h.
11576 (gcov_position): Ditto.
11577 (gcov_is_error): Ditto.
11578 (gcov_rewrite): Ditto.
11579 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11580 only part to libgcc/libgcov.h.
11582 2014-01-08 Marek Polacek <polacek@redhat.com>
11584 PR middle-end/59669
11585 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11587 2014-01-08 Marek Polacek <polacek@redhat.com>
11590 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11592 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11594 PR rtl-optimization/59649
11595 * stor-layout.c (get_mode_bounds): For BImode return
11596 0 and STORE_FLAG_VALUE.
11598 2014-01-08 Richard Biener <rguenther@suse.de>
11600 PR middle-end/59630
11601 * gimple.h (is_gimple_builtin_call): Remove.
11602 (gimple_builtin_call_types_compatible_p): New.
11603 (gimple_call_builtin_p): New overload.
11604 * gimple.c (is_gimple_builtin_call): Remove.
11605 (validate_call): Rename to ...
11606 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11607 check return types.
11608 (validate_type): New static function.
11609 (gimple_call_builtin_p): New overload and adjust.
11610 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11611 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11612 (gimple_fold_stmt_to_constant_1): Likewise.
11613 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11615 2014-01-08 Richard Biener <rguenther@suse.de>
11617 PR middle-end/59471
11618 * gimplify.c (gimplify_expr): Gimplify register-register type
11619 VIEW_CONVERT_EXPRs to separate stmts.
11621 2014-01-07 Jeff Law <law@redhat.com>
11623 PR middle-end/53623
11624 * ree.c (combine_set_extension): Handle case where source
11625 and destination registers in an extension insn are different.
11626 (combine_reaching_defs): Allow source and destination registers
11627 in extension to be different under limited circumstances.
11628 (add_removable_extension): Remove restriction that the
11629 source and destination registers in the extension are the same.
11630 (find_and_remove_re): Emit a copy from the extension's
11631 destination to its source after the defining insn if
11632 the source and destination registers are different.
11634 PR middle-end/59285
11635 * ifcvt.c (merge_if_block): If we are merging a block with more than
11636 one successor with a block with no successors, remove any BARRIER
11637 after the second block.
11639 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11641 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11643 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11646 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11647 for 14-bit register offsets when INT14_OK_STRICT is false.
11649 2014-01-07 Roland Stigge <stigge@antcom.de>
11650 Michael Meissner <meissner@linux.vnet.ibm.com>
11653 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11654 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11656 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11658 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11661 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11663 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11664 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11665 rtx is const0_rtx or not.
11667 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11670 * target-globals.c (save_target_globals): Remove this_fn_optab
11672 * toplev.c: Include optabs.h.
11673 (target_reinit): Temporarily restore the global options if another
11674 set of options are in force.
11676 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11678 PR rtl-optimization/58668
11679 * cfgcleanup.c (flow_find_cross_jump): Don't count
11680 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11681 to determine what is counted.
11682 (flow_find_head_matching_sequence): Use active_insn_p to determine
11684 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11686 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11687 determine what is counted.
11689 PR tree-optimization/59643
11690 * tree-predcom.c (split_data_refs_to_components): If one dr is
11691 read and one write, determine_offset fails and the write isn't
11692 in the bad component, just put the read into the bad component.
11694 2014-01-07 Mike Stump <mikestump@comcast.net>
11695 Jakub Jelinek <jakub@redhat.com>
11698 * tree-core.h (struct tree_optimization_option): Change optabs
11699 type from unsigned char * to void *.
11700 * optabs.c (init_tree_optimization_optabs): Adjust
11701 TREE_OPTIMIZATION_OPTABS initialization.
11703 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11706 * config/i386/i386.h (struct machine_function): Add
11707 no_drap_save_restore field.
11708 * config/i386/i386.c (ix86_save_reg): Use
11709 !cfun->machine->no_drap_save_restore instead of
11710 crtl->stack_realign_needed.
11711 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11712 this function clears frame_pointer_needed. Set
11713 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11714 and DRAP reg is needed.
11716 2014-01-06 Marek Polacek <polacek@redhat.com>
11719 * doc/implement-c.texi: Mention that other integer types are
11720 permitted as bit-field types in strictly conforming mode.
11722 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11724 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11725 is newly allocated.
11727 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11729 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11731 2014-01-06 Martin Jambor <mjambor@suse.cz>
11734 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11736 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11738 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11742 * var-tracking.c (add_stores): Preserve the value of the source even if
11743 we don't record the store.
11745 2014-01-06 Terry Guo <terry.guo@arm.com>
11747 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11749 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11752 * config/darwin.c (darwin_function_section): Adjust return values to
11753 correspond to optimisation changes made in r206070.
11755 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11757 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11758 from prefetch_block tune setting.
11759 (nocona_cost): Correct size of prefetch block to 64.
11761 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11763 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11764 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11765 used to save the static chain register in the computation of the offset
11766 from which the FP registers need to be restored.
11768 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11770 PR tree-optimization/59519
11771 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11772 ICE if get_current_def (current_new_name) is already non-NULL, as long
11773 as it is a phi result of some other phi in *new_exit_bb that has
11776 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11777 or vmovdqu* for misaligned_operand.
11778 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11779 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11780 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11781 aligned_mem for AVX512F masked aligned load and store builtins and for
11782 non-temporal moves.
11784 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11786 PR tree-optimization/59651
11787 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11788 Address range for negative step should be added by TYPE_SIZE_UNIT.
11790 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11792 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11793 overlapping registers also for registers other than the stack pointer.
11795 2014-01-03 Marek Polacek <polacek@redhat.com>
11798 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11801 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11804 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11807 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11808 (push splitter): Use <P:MODE_SIZE> instead of
11809 GET_MODE_SIZE (<P:MODE>mode).
11810 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11811 (mov -1, reg peephole2): Likewise.
11812 * config/i386/sse.md (*mov<mode>_internal,
11813 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11814 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11815 *<code><mode>3, *andnot<mode>3<mask_name>,
11816 <mask_codefor><code><mode>3<mask_name>): Likewise.
11817 * config/i386/subst.md (mask_mode512bit_condition,
11818 sd_mask_mode512bit_condition): Likewise.
11820 2014-01-02 Xinliang David Li <davidxl@google.com>
11822 PR tree-optimization/59303
11823 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11824 (dump_predicates): Better output format.
11825 (pred_equal_p): New function.
11826 (is_neq_relop_p): Ditto.
11827 (is_neq_zero_form_p): Ditto.
11828 (pred_expr_equal_p): Ditto.
11829 (pred_neg_p): Ditto.
11830 (simplify_pred): Ditto.
11831 (simplify_preds_2): Ditto.
11832 (simplify_preds_3): Ditto.
11833 (simplify_preds_4): Ditto.
11834 (simplify_preds): Ditto.
11835 (push_pred): Ditto.
11836 (push_to_worklist): Ditto.
11837 (get_pred_info_from_cmp): Ditto.
11838 (is_degenerated_phi): Ditto.
11839 (normalize_one_pred_1): Ditto.
11840 (normalize_one_pred): Ditto.
11841 (normalize_one_pred_chain): Ditto.
11842 (normalize_preds): Ditto.
11843 (normalize_cond_1): Remove function.
11844 (normalize_cond): Ditto.
11845 (is_gcond_subset_of): Ditto.
11846 (is_subset_of_any): Ditto.
11847 (is_or_set_subset_of): Ditto.
11848 (is_and_set_subset_of): Ditto.
11849 (is_norm_cond_subset_of): Ditto.
11850 (pred_chain_length_cmp): Ditto.
11851 (convert_control_dep_chain_into_preds): Type change.
11852 (find_predicates): Ditto.
11853 (find_def_preds): Ditto.
11854 (destroy_predicates_vecs): Ditto.
11855 (find_matching_predicates_in_rest_chains): Ditto.
11856 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11857 (is_pred_expr_subset): Ditto.
11858 (is_pred_chain_subset_of): Ditto.
11859 (is_included_in): Ditto.
11860 (is_superset_of): Ditto.
11862 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11864 Update copyright years.
11866 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11868 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11869 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11870 config/arc/arc.md, config/arc/arc.opt,
11871 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11872 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11873 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11874 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11875 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11876 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11877 vtable-verify.c, vtable-verify.h: Use the standard form for the
11880 2014-01-02 Tobias Burnus <burnus@net-b.de>
11882 * gcc.c (process_command): Update copyright notice dates.
11883 * gcov-dump.c: Ditto.
11885 * doc/cpp.texi: Bump @copying's copyright year.
11886 * doc/cppinternals.texi: Ditto.
11887 * doc/gcc.texi: Ditto.
11888 * doc/gccint.texi: Ditto.
11889 * doc/gcov.texi: Ditto.
11890 * doc/install.texi: Ditto.
11891 * doc/invoke.texi: Ditto.
11893 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11895 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11897 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11899 * config/i386/sse.md (*mov<mode>_internal): Guard
11900 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11902 PR rtl-optimization/59647
11903 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11904 new_rtx into UNSIGNED_FLOAT rtxes.
11906 Copyright (C) 2014 Free Software Foundation, Inc.
11908 Copying and distribution of this file, with or without modification,
11909 are permitted in any medium without royalty provided the copyright
11910 notice and this notice are preserved.