1 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
3 * config/rs6000/rs6000.md (type): Add new value "halfmul".
4 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
5 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
6 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
7 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
8 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu):
10 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
11 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
12 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
13 * config/rs6000/titan.md: Delete nonsensical comment.
14 (titan_imul): Add type imul3.
15 (titan_mulhw): Remove type imul3; add type halfmul.
17 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
19 * config/rs6000/rs6000.md (type): Reorder, reformat.
21 2014-05-23 Martin Jambor <mjambor@suse.cz>
23 PR tree-optimization/53787
24 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
25 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
26 analysis_done, update all uses.
27 * ipa-prop.c: Include domwalk.h
28 (param_analysis_info): Removed.
29 (param_aa_status): New type.
30 (ipa_bb_info): Likewise.
31 (func_body_info): Likewise.
32 (ipa_get_bb_info): New function.
33 (aa_overwalked): Likewise.
34 (find_dominating_aa_status): Likewise.
35 (parm_bb_aa_status_for_bb): Likewise.
36 (parm_preserved_before_stmt_p): Changed to use new param AA info.
37 (load_from_unmodified_param): Accept func_body_info as a parameter
38 instead of parms_ainfo.
39 (parm_ref_data_preserved_p): Changed to use new param AA info.
40 (parm_ref_data_pass_through_p): Likewise.
41 (ipa_load_from_parm_agg_1): Likewise. Update callers.
42 (compute_complex_assign_jump_func): Changed to use new param AA info.
43 (compute_complex_ancestor_jump_func): Likewise.
44 (ipa_compute_jump_functions_for_edge): Likewise.
45 (ipa_compute_jump_functions): Removed.
46 (ipa_compute_jump_functions_for_bb): New function.
47 (ipa_analyze_indirect_call_uses): Likewise, moved variable
49 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
50 and info, moved variable declarations down.
51 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
53 (ipa_analyze_stmt_uses): Likewise.
54 (ipa_analyze_params_uses): Removed.
55 (ipa_analyze_params_uses_in_bb): New function.
56 (ipa_analyze_controlled_uses): Likewise.
57 (free_ipa_bb_info): Likewise.
58 (analysis_dom_walker): New class.
59 (ipa_analyze_node): Handle node-specific forbidden analysis,
60 initialize and free func_body_info, use dominator walker.
61 (ipcp_modif_dom_walker): New class.
62 (ipcp_transform_function): Create and free func_body_info, use
63 ipcp_modif_dom_walker, moved a lot of functionality there.
65 2014-05-23 Marek Polacek <polacek@redhat.com>
66 Jakub Jelinek <jakub@redhat.com>
68 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
69 * gcc.c (sanitize_spec_function): Likewise.
70 * convert.c (convert_to_integer): Include "ubsan.h". Add
71 floating-point to integer instrumentation.
72 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
73 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
75 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
76 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
77 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
78 * ubsan.c: Include "realmpfr.h" and "dfp.h".
79 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
80 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
81 float/double/long double.
82 (ubsan_instrument_float_cast): New function.
83 * ubsan.h (ubsan_instrument_float_cast): Declare.
85 2014-05-23 Jiong Wang <jiong.wang@arm.com>
87 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
89 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
90 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
91 Adjust for tailcalling through registers.
92 * config/aarch64/aarch64.h (enum reg_class): New caller save
94 (REG_CLASS_NAMES): Likewise.
95 (REG_CLASS_CONTENTS): Likewise.
96 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
97 Allow tailcalling without decls.
99 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
101 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
102 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
104 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
105 gsi, and variables v_* to v*.
107 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
109 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
111 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
113 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
114 * omp-low.c: Update accordingly.
116 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
117 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
118 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
119 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
120 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
121 GF_OMP_TARGET_KIND_UPDATE.
123 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
124 Explicitly enumerate the expected region types.
126 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
129 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
130 documentation; the old documentation didn't clearly state the
131 constraints on the contents of the pointed-to storage.
133 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
135 Fix bootstrap error on ia64
136 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
137 Return default value.
139 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
141 PR tree-optimization/54733
142 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
144 (find_bswap_or_nop_load): New.
145 (find_bswap_1): Renamed to ...
146 (find_bswap_or_nop_1): This. Also add support for memory source.
147 (find_bswap): Renamed to ...
148 (find_bswap_or_nop): This. Also add support for memory source and
149 detection of bitwise operations equivalent to load in host endianness.
150 (execute_optimize_bswap): Likewise. Also move its leading comment back
151 in place and split statement transformation into ...
152 (bswap_replace): This.
154 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
156 PR rtl-optimization/61215
157 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
158 simplify_gen_subreg until final substitution.
160 2014-05-23 Alan Modra <amodra@gmail.com>
163 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
164 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
165 Use "Y" constraint rather than "m".
167 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
169 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
171 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
172 New function declaration.
173 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
174 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
175 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
176 (aarch64_init_builtins) : Initialize builtins
177 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
178 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
179 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
180 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
181 and __builtins_aarch64_set_fpsr.
182 (aarch64_atomic_assign_expand_fenv): New function.
183 * config/aarch64/aarch64.md (set_fpcr): New pattern.
184 (get_fpcr) : Likewise.
185 (set_fpsr) : Likewise.
186 (get_fpsr) : Likewise.
187 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
188 and UNSPECV_SET_FPSR.
189 * doc/extend.texi (AARCH64 Built-in Functions) : Document
190 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
191 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
193 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
195 PR rtl-optimization/60969
196 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
197 constraints. Set up mem cost for NO_REGS case.
199 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
201 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
203 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
205 * config/darwin.c: Include "lto-section-names.h".
206 (LTO_SEGMENT_NAME): Don't define.
207 * config/i386/winnt.c: Include "lto-section-names.h".
208 * lto-streamer.c: Include "lto-section-names.h".
209 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
210 * lto-wrapper.c: Include "lto-section-names.h".
211 (LTO_SECTION_NAME_PREFIX): Don't define.
212 * lto-section-names.h: New file.
213 * cgraphunit.c: Include "lto-section-names.h".
215 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
217 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
219 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
222 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
224 2014-05-22 Nick Clifton <nickc@redhat.com>
226 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
229 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
231 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
232 -> (T)A transformation to integer types.
234 2014-05-22 Teresa Johnson <tejohnson@google.com>
236 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
237 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
238 (gcov_rewrite): Use gcov_nonruntime_assert.
240 (gcov_write_words): Ditto.
241 (gcov_write_length): Ditto.
242 (gcov_read_words): Use gcov_nonruntime_assert, and remove
243 gcc_assert from IN_LIBGCOV code.
244 (gcov_read_summary): Use gcov_error to flag profile corruption.
245 (gcov_sync): Use gcov_nonruntime_assert.
246 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
247 (gcov_histo_index): Use gcov_nonruntime_assert.
248 (static void gcov_histogram_merge): Ditto.
249 (compute_working_sets): Ditto.
250 * gcov-io.h (gcov_nonruntime_assert): Define.
251 (gcov_error): Define for !IN_LIBGCOV
253 2014-05-22 Richard Biener <rguenther@suse.de>
255 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
256 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
257 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
258 and deallocation site.
259 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
260 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
261 passing through the incoming points-to set.
262 (handle_lhs_call): Use flags argument instead of recomputing it.
263 (find_func_aliases_for_call): Call handle_lhs_call with proper
266 2014-05-22 Jakub Jelinek <jakub@redhat.com>
268 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
269 all padding bits in REAL_VALUE_TYPE are cleared.
271 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
273 Cleanup and improve multipass_dfa_lookahead_guard
274 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
275 (core2i7_first_cycle_multipass_begin,)
276 (core2i7_first_cycle_multipass_issue,)
277 (core2i7_first_cycle_multipass_backtrack): Update signature.
279 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
280 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
281 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
283 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
284 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
286 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
288 * doc/tm.texi: Regenerate.
290 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
291 * haifa-sched.c (ready_try): Make signed to allow negative values.
292 (rebug_ready_list_1): Update.
293 (choose_ready): Simplify.
294 (sched_extend_ready_list): Update.
296 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
298 Remove IA64 speculation tweaking flags
299 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
300 speculation tuning flags.
301 (msched-prefer-non-data-spec-insns,)
302 (msched-prefer-non-control-spec-insns): Obsolete options.
303 * haifa-sched.c (choose_ready): Remove handling of
304 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
305 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
306 and PREFER_NON_DATA_SPEC.
307 * sel-sched.c (process_spec_exprs): Remove handling of
308 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
310 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
312 Improve scheduling debug output
313 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
314 (advance_one_cycle): Update.
315 (schedule_insn, queue_to_ready): Add debug printouts.
316 (debug_ready_list_1): New static function.
317 (debug_ready_list): Update.
318 (max_issue): Add debug printouts.
319 (dump_insn_stream): New static function.
320 (schedule_block): Use it. Also better indent printouts.
322 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
324 Fix sched_insn debug counter
325 * haifa-sched.c (schedule_insn): Update.
326 (struct haifa_saved_data): Add nonscheduled_insns_begin.
327 (save_backtrack_point, restore_backtrack_point): Update.
328 (first_nonscheduled_insn): New static function.
329 (queue_to_ready, choose_ready): Use it.
330 (schedule_block): Init nonscheduled_insns_begin.
331 (sched_emit_insn): Update.
334 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
336 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
338 (aarch64_secondary_reload) : LikeWise.
339 (aarch64_class_max_nregs) : Remove CORE_REGS.
340 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
341 (REG_CLASS_NAMES) : Likewise.
342 (REG_CLASS_CONTENTS) : LikeWise.
343 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
345 2014-05-21 Guozhi Wei <carrot@google.com>
348 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
350 (vqdmulhq_n_s16): Likewise.
352 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
354 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
356 2014-05-21 Marek Polacek <polacek@redhat.com>
359 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
361 2014-05-21 Martin Jambor <mjambor@suse.cz>
363 * doc/invoke.texi (Optimize Options): Document parameters
364 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
365 ipa-cp-array-index-hint-bonus.
367 2014-05-21 Mark Wielaard <mjw@redhat.com>
370 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
371 version >= 3 or not strict DWARF.
372 * langhooks.h (struct lang_hooks_for_types): Add
373 enum_underlying_base_type.
374 * langhooks.c (lhd_enum_underlying_base_type): New function.
375 * gcc/langhooks.h (struct lang_hooks_for_types): Add
376 enum_underlying_base_type.
377 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
378 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
379 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
381 2014-05-21 Richard Biener <rguenther@suse.de>
383 * doc/invoke.texi (-flto-partition=): Document one and
386 2014-05-21 John Marino <gnugcc@marino.st>
388 * config.gcc (*-*-dragonfly*): New target.
389 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
390 * configure: Regenerate.
391 * config/dragonfly-stdint.h: New.
392 * config/dragonfly.h: New.
393 * config/dragonfly.opt: New.
394 * config/i386/dragonfly.h: New.
395 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
397 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
399 * tree.def (VOID_CST): New.
400 * tree-core.h (TI_VOID): New.
401 * tree.h (void_node): New.
402 * tree.c (tree_node_structure_for_code, tree_code_size)
403 (iterative_hash_expr): Handle VOID_CST.
404 (build_common_tree_nodes): Initialize void_node.
406 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
408 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
410 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
412 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
415 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
416 flag_reorder_blocks_and_partition.
417 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
419 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
422 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
424 (*addc_r_t): Add new insn_and_split.
426 2014-05-21 Jakub Jelinek <jakub@redhat.com>
429 * omp-low.c (handle_simd_reference): New function.
430 (lower_rec_input_clauses): Use it. Defer adding reference
431 initialization even for reduction without placeholder if in simd,
432 handle it properly later on.
434 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
436 PR tree-optimization/60899
437 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
438 assume all static symbols will have definition wile parsing and
439 check the do have definition later in compilation; check that
440 variable referring symbol will be output before concluding that
441 reference is safe; be conservative for referring local statics;
442 be more precise about when comdat is output in other partition.
444 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
447 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
448 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
449 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
451 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
453 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
454 set writeonly flag only for vars actually written to.
456 2014-05-20 Dehao Chen <dehao@google.com>
458 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
459 and callee count to get clone count.
460 * tree-inline.c (expand_call_inline): Use callee count instead of bb
463 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
465 PR rtl-optimization/61243
466 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
468 2014-05-20 Xinliang David Li <davidxl@google.com>
470 * cgraphunit.c (walk_polymorphic_call_targets): Add
471 dbgcnt and fopt-info support.
472 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
473 * ipa-devirt.c (ipa_devirt): Ditto.
474 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
476 * ipa.c (walk_polymorphic_call_targets): Ditto.
477 * gimple-fold.c (fold_gimple_assign): Ditto.
478 (gimple_fold_call): Ditto.
479 * dbgcnt.def: New counter.
481 2014-05-20 DJ Delorie <dj@redhat.com>
483 * config/msp430/msp430.md (split): Don't allow subregs when
484 splitting SImode adds.
485 (andneghi): Fix subtraction logic.
486 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
488 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
490 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
492 * except.c (switch_to_exception_section, resolve_unique_section,
493 get_named_text_section, default_function_rodata_section,
494 align_variable, get_block_for_decl, default_section_type_flags):
495 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
496 * symtab.c (symtab_add_to_same_comdat_group,
497 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
498 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
500 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
501 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
502 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
503 (c6x_function_in_section_p): Likewise.
504 * config/darwin.c (machopic_select_section): Likewise.
505 * config/arm/arm.c (arm_function_in_section_p): Likewise.
506 * config/mips/mips.c (mips_function_rodata_section): Likewise.
507 * config/mep/mep.c (mep_select_section): LIkewise.
508 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
510 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
512 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
513 EH region of calls to pure functions that can throw an exception.
514 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
515 (copy_reference_ops_from_call): Also copy the EH region of the call if
516 it can throw an exception.
518 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
520 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
521 nested VEC_SELECTs that are inverses of each other.
523 2014-05-20 Richard Biener <rguenther@suse.de>
525 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
527 (extract_and_process_scc_for_name): not here.
528 (cond_dom_walker::before_dom_children): Only process
529 stmts that end the BB in interesting ways.
530 (run_scc_vn): Mark param uses as visited.
532 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
534 * config/arm/arm.md (arith_shiftsi): Do not predicate for
537 2014-05-20 Nick Clifton <nickc@redhat.com>
539 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
540 (msp430_gimplify_va_arg_expr): New function.
541 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
543 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
544 operand 0 in order to prevent confusion about the number of
547 2014-05-20 Richard Biener <rguenther@suse.de>
549 PR tree-optimization/61221
550 * tree-ssa-pre.c (el_to_update): Remove.
551 (eliminate_dom_walker::before_dom_children): Handle released
552 VDEFs by value-numbering them to the associated VUSE. Update
553 stmt immediately for substituted call address.
554 (eliminate): Remove delayed stmt updating code.
555 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
556 possibly late re-numbered vuses.
557 (vn_reference_lookup_2): Adjust.
558 (vn_reference_lookup_pieces): Likewise.
559 (vn_reference_lookup): Likewise.
561 2014-05-20 Richard Biener <rguenther@suse.de>
563 * config.gcc: Remove need_64bit_hwint.
564 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
565 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
567 * config.in: Regenerate.
568 * configure: Likewise.
570 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
572 * doc/extend.texi: Create Label Attributes section,
573 move all label attributes into it and reference it.
575 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
577 * arm.c (thumb1_reorg): When scanning backwards skip anything
578 that's not a proper insn.
580 2014-05-19 Richard Biener <rguenther@suse.de>
582 PR tree-optimization/61221
583 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
584 Do nothing for unreachable blocks.
585 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
586 Improve unreachability detection.
588 2014-05-19 Richard Biener <rguenther@suse.de>
590 PR tree-optimization/61209
591 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
593 2014-05-19 Nick Clifton <nickc@redhat.com>
595 * except.c (init_eh): Fix computation of builtin setjmp buffer
596 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
598 2014-05-19 Richard Biener <rguenther@suse.de>
600 PR tree-optimization/61184
601 * tree-vrp.c (is_negative_overflow_infinity): Use
602 TREE_OVERFLOW_P and do that check first.
603 (is_positive_overflow_infinity): Likewise.
604 (is_overflow_infinity): Likewise.
605 (vrp_operand_equal_p): Properly treat operands with
606 differing overflow as not equal.
608 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
610 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
611 shift simplification where it was intended.
613 2014-05-19 Christian Bruel <christian.bruel@st.com>
616 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
618 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
621 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
624 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
626 * reg-notes.def (CROSSING_JUMP): Likewise.
627 * rtl.h (rtx_def): Update comment for jump flag.
628 (CROSSING_JUMP_P): Define.
629 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
630 of a REG_CROSSING_JUMP note.
631 * cfghooks.c (tidy_fallthru_edges): Likewise.
632 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
633 * emit-rtl.c (try_split): Likewise.
634 * haifa-sched.c (sched_create_recovery_edges): Likewise.
635 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
636 * jump.c (redirect_jump_2): Likewise.
637 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
638 (relax_delay_slots): Likewise.
639 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
641 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
642 * config/sh/sh.md (jump_compact): Likewise.
643 * bb-reorder.c (rotate_loop): Likewise.
644 (pass_duplicate_computed_gotos::execute): Likewise.
645 (add_reg_crossing_jump_notes): Rename to...
646 (update_crossing_jump_flags): ...this.
647 (pass_partition_blocks::execute): Update accordingly.
649 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
651 * tree.h: Remove extraneous template <>.
653 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
655 * ipa.c (symtab_remove_unreachable_nodes): Remove
656 symbol from comdat group if its body was eliminated.
657 (comdat_can_be_unshared_p_1): Static symbols can always
659 * symtab.c (symtab_remove_from_same_comdat_group): Break out
661 (symtab_unregister_node): ... this one.
662 (verify_symtab_base): More strict checking of comdats.
663 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
665 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
667 * tree-pass.h (make_pass_ipa_comdats): New pass.
668 * timevar.def (TV_IPA_COMDATS): New timevar.
669 * passes.def (pass_ipa_comdats): Add.
670 * Makefile.in (OBJS): Add ipa-comdats.o
671 * ipa-comdats.c: New file.
673 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
675 * ipa.c (update_visibility_by_resolution_info): New function.
676 (function_and_variable_visibility): Use it.
678 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
680 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
682 (FOR_EACH_DEFINED_SYMBOL): New macro.
683 (varpool_first_static_initializer, varpool_next_static_initializer,
684 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
685 (symtab_in_same_comdat_p): Correctly deal with inline functions.
687 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
689 * ggc-page.c (ggc_handle_finalizers): Add comment.
691 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
693 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
694 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
695 (ggc_internal_cleared_alloc): Likewise.
696 * ggc-page.c (finalizer): New class.
697 (vec_finalizer): Likewise.
698 (globals::finalizers): New member.
699 (globals::vec_finalizers): Likewise.
700 (ggc_internal_alloc): Record the finalizer if any for the block being
702 (ggc_handle_finalizers): New function.
703 (ggc_collect): Call ggc_handle_finalizers.
704 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
706 (ggc_internal_cleared_alloc): Likewise.
707 (finalize): New function.
708 (need_finalization_p): Likewise.
709 (ggc_alloc): Install the type's destructor as the finalizer if it
711 (ggc_cleared_alloc): Likewise.
712 (ggc_vec_alloc): Likewise.
713 (ggc_cleared_vec_alloc): Likewise.
715 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
717 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
719 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
721 * alias.c (record_alias_subset): Adjust.
722 * bitmap.c (bitmap_element_allocate): Likewise.
723 (bitmap_gc_alloc_stat): Likewise.
724 * cfg.c (init_flow): Likewise.
725 (alloc_block): Likewise.
726 (unchecked_make_edge): Likewise.
727 * cfgloop.c (alloc_loop): Likewise.
728 (flow_loops_find): Likewise.
729 (rescan_loop_exit): Likewise.
730 * cfgrtl.c (init_rtl_bb_info): Likewise.
731 * cgraph.c (insert_new_cgraph_node_version): Likewise.
732 (cgraph_allocate_node): Likewise.
733 (cgraph_create_edge_1): Likewise.
734 (cgraph_allocate_init_indirect_info): Likewise.
735 * cgraphclones.c (cgraph_clone_edge): Likewise.
736 * cgraphunit.c (add_asm_node): Likewise.
737 (init_lowered_empty_function): Likewise.
738 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
739 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
740 (alpha_use_linkage): Likewise.
741 * config/arc/arc.c (arc_init_machine_status): Likewise.
742 * config/arm/arm.c (arm_init_machine_status): Likewise.
743 * config/avr/avr.c (avr_init_machine_status): Likewise.
744 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
745 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
746 * config/cris/cris.c (cris_init_machine_status): Likewise.
747 * config/darwin.c (machopic_indirection_name): Likewise.
748 (darwin_build_constant_cfstring): Likewise.
749 (darwin_enter_string_into_cfstring_table): Likewise.
750 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
751 * config/frv/frv.c (frv_init_machine_status): Likewise.
752 * config/i386/i386.c (get_dllimport_decl): Likewise.
753 (ix86_init_machine_status): Likewise.
754 (assign_386_stack_local): Likewise.
755 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
756 (i386_pe_maybe_record_exported_symbol): Likewise.
757 (i386_pe_record_stub): Likewise.
758 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
759 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
760 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
761 (m32c_note_pragma_address): Likewise.
762 * config/mep/mep.c (mep_init_machine_status): Likewise.
763 (mep_note_pragma_flag): Likewise.
764 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
765 (mips16_local_alias): Likewise.
766 (mips_init_machine_status): Likewise.
767 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
768 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
769 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
770 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
771 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
772 * config/pa/pa.c (pa_init_machine_status): Likewise.
773 (pa_get_deferred_plabel): Likewise.
774 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
775 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
776 (rs6000_init_machine_status): Likewise.
777 (output_toc): Likewise.
778 * config/s390/s390.c (s390_init_machine_status): Likewise.
779 * config/score/score.c (score_output_external): Likewise.
780 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
781 * config/spu/spu.c (spu_init_machine_status): Likewise.
782 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
783 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
784 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
785 * coverage.c (coverage_end_function): Likewise.
786 * dbxout.c (dbxout_init): Likewise.
787 * doc/gty.texi: Don't mention variable_size attribute.
788 * dwarf2cfi.c (new_cfi): Adjust.
789 (new_cfi_row): Likewise.
790 (copy_cfi_row): Likewise.
791 (create_cie_data): Likewise.
792 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
793 (new_loc_descr): Likewise.
794 (find_AT_string_in_table): Likewise.
795 (add_addr_table_entry): Likewise.
797 (add_var_loc_to_decl): Likewise.
798 (clone_die): Likewise.
799 (clone_as_declaration): Likewise.
800 (break_out_comdat_types): Likewise.
801 (new_loc_list): Likewise.
802 (add_loc_descr_to_each): Likewise.
803 (add_location_or_const_value_attribute): Likewise.
804 (add_linkage_name): Likewise.
805 (lookup_filename): Likewise.
806 (dwarf2out_var_location): Likewise.
807 (new_line_info_table): Likewise.
808 (dwarf2out_init): Likewise.
809 (mem_loc_descriptor): Likewise.
810 (loc_descriptor): Likewise.
811 (add_const_value_attribute): Likewise.
812 (tree_add_const_value_attribute): Likewise.
813 (comp_dir_string): Likewise.
814 (dwarf2out_vms_debug_main_pointer): Likewise.
815 (string_cst_pool_decl): Likewise.
816 * emit-rtl.c (set_mem_attrs): Likewise.
817 (get_reg_attrs): Likewise.
818 (start_sequence): Likewise.
819 (init_emit): Likewise.
820 (init_emit_regs): Likewise.
821 * except.c (init_eh_for_function): Likewise.
822 (gen_eh_region): Likewise.
823 (gen_eh_region_catch): Likewise.
824 (gen_eh_landing_pad): Likewise.
825 (add_call_site): Likewise.
826 * function.c (add_frame_space): Likewise.
827 (insert_temp_slot_address): Likewise.
828 (assign_stack_temp_for_type): Likewise.
829 (get_hard_reg_initial_val): Likewise.
830 (allocate_struct_function): Likewise.
831 (prepare_function_start): Likewise.
832 (types_used_by_var_decl_insert): Likewise.
833 * gengtype.c (variable_size_p): Remove function.
834 (enum alloc_quantity): Remove enum.
835 (write_typed_alloc_def): Remove function.
836 (write_typed_struct_alloc_def): Likewise.
837 (write_typed_typedef_alloc_def): Likewise.
838 (write_typed_alloc_defns): Likewise.
840 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
841 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
842 * ggc.h (ggc_alloc): new function.
843 (ggc_cleared_alloc): Likewise.
844 (ggc_vec_alloc): Template on type of vector element, and remove
845 element size argument.
846 (ggc_cleared_vec_alloc): Likewise.
847 * gimple.c (gimple_build_omp_for): Adjust.
848 (gimple_copy): Likewise.
849 * ipa-cp.c (get_replacement_map): Likewise.
850 (find_aggregate_values_for_callers_subset): Likewise.
851 (known_aggs_to_agg_replacement_list): Likewise.
852 * ipa-devirt.c (get_odr_type): Likewise.
853 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
854 (read_agg_replacement_chain): Likewise.
855 * loop-iv.c (get_simple_loop_desc): Likewise.
856 * lto-cgraph.c (input_node_opt_summary): Likewise.
857 * lto-section-in.c (lto_new_in_decl_state): Likewise.
858 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
859 (input_eh_region): Likewise.
860 (input_eh_lp): Likewise.
861 (input_cfg): Likewise.
862 * optabs.c (set_optab_libfunc): Likewise.
863 (init_tree_optimization_optabs): Likewise.
864 (set_conv_libfunc): Likewise.
865 * passes.c (do_per_function_toporder): Likewise.
866 * rtl.h: Don't use variable_size gty attribute.
867 * sese.c (if_region_set_false_region): Adjust.
868 * stringpool.c (gt_pch_save_stringpool): Likewise.
869 * target-globals.c (save_target_globals): Likewise.
870 * toplev.c (general_init): Likewise.
871 * trans-mem.c (record_tm_replacement): Likewise.
872 (split_bb_make_tm_edge): Likewise.
873 * tree-cfg.c (move_sese_region_to_fn): Likewise.
874 * tree-data-ref.h (lambda_vector_new): Likewise.
875 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
876 * tree-iterator.c (tsi_link_before): Likewise.
877 (tsi_link_after): Likewise.
878 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
879 * tree-ssa-loop-niter.c (record_estimate): Likewise.
880 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
881 * tree-ssa-operands.h: Don't use variable_size gty attribute.
882 * tree-ssa.c (init_tree_ssa): Adjust.
883 * tree-ssanames.c (set_range_info): Likewise.
884 (get_ptr_info): Likewise.
885 (duplicate_ssa_name_ptr_info): Likewise.
886 (duplicate_ssa_name_range_info): Likewise.
887 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
888 (unpack_ts_fixed_cst_value_fields): Likewise.
889 * tree.c (build_fixed): Likewise.
890 (build_real): Likewise.
891 (build_string): Likewise.
892 (decl_priority_info): Likewise.
893 (decl_debug_expr_insert): Likewise.
894 (decl_value_expr_insert): Likewise.
895 (decl_debug_args_insert): Likewise.
896 (type_hash_add): Likewise.
897 (build_omp_clause): Likewise.
898 * ubsan.c (decl_for_type_insert): Likewise.
899 * varasm.c (get_unnamed_section): Likewise.
900 (get_noswitch_section): Likewise.
901 (get_section): Likewise.
902 (get_block_for_section): Likewise.
903 (create_block_symbol): Likewise.
904 (build_constant_desc): Likewise.
905 (create_constant_pool): Likewise.
906 (force_const_mem): Likewise.
907 (record_tm_clone_pair): Likewise.
908 * varpool.c (varpool_create_empty_node): Likewise.
910 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
912 * dwarf2out.c (tree_add_const_value_attribute): Call
913 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
914 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
915 instead of ggc_internal_<x>alloc_stat.
916 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
917 (ggc_realloc): Likewise.
918 * ggc-none.c (ggc_internal_alloc): Likewise.
919 (ggc_internal_cleared_alloc): Likewise.
920 * ggc-page.c: Likewise.
921 * ggc.h (ggc_internal_alloc_stat): Likewise.
922 (ggc_internal_alloc): Remove macro.
923 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
924 (ggc_internal_cleared_alloc): Remove macro.
925 (GGC_RESIZEVEC): Adjust.
926 (ggc_resizevar): Remove macro.
927 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
928 (ggc_internal_cleared_vec_alloc_stat): Likewise.
929 (ggc_internal_vec_cleared_alloc): Remove macro.
930 (ggc_alloc_atomic_stat): Drop _stat suffix.
931 (ggc_alloc_atomic): Remove macro.
932 (ggc_alloc_cleared_atomic): Remove macro.
933 (ggc_alloc_string_stat): Drop _stat suffix.
934 (ggc_alloc_string): Remove macro.
935 (ggc_alloc_rtx_def_stat): Adjust.
936 (ggc_alloc_tree_node_stat): Likewise.
937 (ggc_alloc_cleared_tree_node_stat): Likewise.
938 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
939 (ggc_alloc_cleared_simd_clone_stat): Likewise.
940 * gimple.c (gimple_build_omp_for): Likewise.
941 (gimple_copy): Likewise.
942 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
943 * toplev.c (realloc_for_line_map): Adjust.
944 * tree-data-ref.h (lambda_vector_new): Likewise.
945 * tree-phinodes.c (allocate_phi_node): Likewise.
946 * tree.c (grow_tree_vec_stat): Likewise.
947 * vec.h (va_gc::reserve): Adjust.
949 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
951 * config/microblaze/microblaze.c (break_handler): New Declaration.
952 (microblaze_break_function_p,microblaze_is_break_handler): New.
953 (compute_frame_size): Use microblaze_break_function_p.
954 Add the test of break_handler.
955 (microblaze_function_prologue) : Add the test of variable
956 break_handler. Check the fnname by BREAK_HANDLER_NAME.
957 (microblaze_function_epilogue) : Add the test of break_handler.
958 (microblaze_globalize_label) : Add the test of break_handler.
959 Check the name by BREAK_HANDLER_NAME.
961 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
963 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
964 microblaze_is_break_handler test.
965 (call_internal1,call_value_intern): Use microblaze_break_function_p.
968 * config/microblaze/microblaze-protos.h
969 (microblaze_break_function_p,microblaze_is_break_handler):
972 * doc/extend.texi (MicroBlaze break_handler Functions): Document
973 new MicroBlaze break_handler functions.
975 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
977 * doc/extend.texi (Size of an asm): Move node text according
978 to its @menu entry position.
980 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
982 PR tree-optimization/61140
983 PR tree-optimization/61150
984 PR tree-optimization/61197
985 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
987 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
989 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
991 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
993 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
994 __SIZEOF_INT128__ is defined.
996 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
998 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
999 (rs6000_delegitimize_address): Use it.
1001 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1003 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1004 inplace argument. Store the new address in the original MEM when true.
1005 * emit-rtl.c (change_address_1): Likewise.
1006 (adjust_address_1, adjust_automodify_address_1, offset_address):
1008 * rtl.h (plus_constant): Add an inplace argument.
1009 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1010 when true. Avoid generating (plus X (const_int 0)).
1011 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1012 in-place. Pass true to plus_constant.
1013 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1015 2014-05-16 Dehao Chen <dehao@google.com>
1017 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1019 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1022 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1024 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1026 2014-05-16 Dehao Chen <dehao@google.com>
1028 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1029 optimize_function_for_size_p.
1030 * regs.h (REG_FREQ_FROM_BB): Likewise.
1032 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1035 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1036 negt_reg_operand cases.
1037 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1039 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1041 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1043 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1046 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1049 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1051 * tree-cfg.c (dump_function_to_file): Dump the return type of
1052 functions, in a line to itself before the function body, mimicking
1053 the layout of a C function.
1055 2014-05-16 Dehao Chen <dehao@google.com>
1057 * cfghooks.c (make_forwarder_block): Use direct computation to
1058 get fall-through edge's count and frequency.
1060 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1062 * config/arc/arc.c (arc_init): Fix typo in error message.
1063 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1064 (split_stack_prologue_scratch_regno): Likewise.
1065 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1066 word from error message.
1068 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1070 * ira-costs.c: Fix typo in comment.
1072 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1074 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1076 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1078 * varpool.c (dump_varpool_node): Dump write-only flag.
1079 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1081 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1082 write-only variables.
1083 * ipa.c (process_references): New function.
1084 (set_readonly_bit): New function.
1085 (set_writeonly_bit): New function.
1086 (clear_addressable_bit): New function.
1087 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1088 fix handling of aliases.
1089 * cgraph.h (struct varpool_node): Add writeonly flag.
1091 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1093 PR rtl-optimization/60969
1094 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1095 Calculate costs for this case.
1097 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1099 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1100 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1102 2014-05-16 Richard Biener <rguenther@suse.de>
1104 PR tree-optimization/61194
1105 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1106 bool patterns ending in a COND_EXPR.
1108 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1110 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1112 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1114 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1115 where we were unable to cost an RTX.
1117 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1119 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1122 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1123 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1125 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1127 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1128 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1130 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1131 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1133 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1134 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1136 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1139 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1140 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1142 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1145 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1146 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1148 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1149 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1151 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1152 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1154 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1157 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1158 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1160 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1161 ZERO_EXTEND and SIGN_EXTEND better.
1163 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1164 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1166 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1169 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1170 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1172 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1173 costs when costing loads and stores to memory.
1175 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1176 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1178 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1181 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1183 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1185 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1186 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1188 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1190 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1192 (aarch64_rtx_mult_cost): New.
1193 (aarch64_rtx_costs): Use it, refactor as appropriate.
1195 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1196 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1198 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1199 emit instructions, return number of instructions which would
1201 (aarch64_add_constant): Update call to aarch64_build_constant.
1202 (aarch64_output_mi_thunk): Likewise.
1203 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1206 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1208 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1209 (TARGET_RTX_COSTS): Call it.
1211 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1213 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1214 (cortexa57_vector_cost): Likewise.
1215 (cortexa57_tunings): Use them.
1217 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1219 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1220 (cpu_addrcost_table): Use it.
1221 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1222 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1225 2014-05-16 Richard Biener <rguenther@suse.de>
1227 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1228 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1229 (visit_phi): Ignore edges marked as not executable.
1230 (class cond_dom_walker): New.
1231 (cond_dom_walker::before_dom_children): Value-number
1232 control statements and mark successor edges as not
1233 executable if possible.
1234 (run_scc_vn): First walk all control statements in
1235 dominator order, marking edges as not executable.
1236 * tree-inline.c (copy_edges_for_bb): Be not confused
1237 about random edge flags.
1239 2014-05-16 Richard Biener <rguenther@suse.de>
1241 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1243 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1246 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1247 (__TM_simple_begin): Use it.
1248 (__TM_begin): Likewise.
1250 2014-05-15 Martin Jambor <mjambor@suse.cz>
1253 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1254 type_preserved flag when the indirect edge is polymorphic.
1256 2014-05-15 Martin Jambor <mjambor@suse.cz>
1258 PR tree-optimization/61090
1259 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1260 build_ref_for_model.
1262 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1264 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1265 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1267 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1269 PR tree-optimization/61158
1270 * fold-const.c (fold_binary_loc): If X is zero-extended and
1271 shiftc >= prec, make sure zerobits is all ones instead of
1272 invoking undefined behavior.
1274 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1276 * regcprop.h: New file.
1277 * regcprop.c (skip_debug_insn_p): New decl.
1278 (replace_oldest_value_reg): Check skip_debug_insn_p.
1279 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1280 * shrink-wrap.c: Include regcprop.h.
1281 (prepare_shrink_wrap): Call
1282 copyprop_hardreg_forward_bb_without_debug_insn.
1284 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1286 * shrink-wrap.h: Update comment.
1287 * shrink-wrap.c: Update comment.
1288 (next_block_for_reg): Rename to live_edge_for_reg.
1289 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1290 (move_insn_for_shrink_wrap): Split live_edge.
1291 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1293 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1295 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1297 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1298 * config/sparc/sparc.md (fptype_ut699): New attribute.
1299 (in_branch_delay): Return false if -mfix-ut699 is specified and
1300 fptype_ut699 is set to single.
1301 (truncdfsf2): Add fptype_ut699 attribute.
1302 (fix_truncdfsi2): Likewise.
1303 (floatsisf2): Change fptype attribute.
1304 (fix_truncsfsi2): Likewise.
1305 (negtf2_notv9): Delete.
1306 (negtf2_v9): Likewise.
1307 (negtf2_hq): New instruction.
1308 (negtf2): New instruction and splitter.
1309 (negdf2_notv9): Rewrite.
1310 (abstf2_notv9): Delete.
1311 (abstf2_hq_v9): Likewise.
1312 (abstf2_v9): Likewise.
1313 (abstf2_hq): New instruction.
1314 (abstf2): New instruction and splitter.
1315 (absdf2_notv9): Rewrite.
1317 2014-05-14 Cary Coutant <ccoutant@google.com>
1320 * opts.c (common_handle_option): Don't special-case "-g".
1321 (set_debug_level): Default to at least level 2 with "-g".
1323 2014-05-14 DJ Delorie <dj@redhat.com>
1325 * config/msp430/msp430.c (msp430_builtin): Add
1326 MSP430_BUILTIN_DELAY_CYCLES.
1327 (msp430_init_builtins): Register void __delay_cycles(long long).
1328 (msp430_builtin_decl): Add it.
1329 (cg_magic_constant): New.
1330 (msp430_expand_delay_cycles): New.
1331 (msp430_expand_builtin): Call it.
1332 (msp430_print_operand_raw): Change integer printing from "int" to
1334 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1335 (delay_cycles_start): New.
1336 (delay_cycles_end): New.
1337 (delay_cycles_32): New.
1338 (delay_cycles_32x): New.
1339 (delay_cycles_16): New.
1340 (delay_cycles_16x): New.
1341 (delay_cycles_2): New.
1342 (delay_cycles_1): New.
1343 * doc/extend.texi: Document __delay_cycles().
1345 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1347 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1348 length attribute computation.
1350 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1353 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1355 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1358 * config/sparc/sparc.md: Fix types of low and high in DI constant
1359 splitter. Use gen_int_mode in some other splitters.
1361 2014-05-14 Martin Jambor <mjambor@suse.cz>
1364 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1366 2014-05-14 James Norris <jnorris@codesourcery.com>
1368 * omp-low.c (expand_parallel_call): Remove shadow variable.
1369 (expand_omp_taskreg): Likewise.
1371 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1373 * common/config/i386/i386-common.c
1374 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1375 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1376 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1377 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1378 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1379 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1380 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1381 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1382 xsavecintrin.h, xsavesintrin.h.
1383 (x86_64-*-*): Ditto.
1384 * config/i386/clflushoptintrin.h: New.
1385 * config/i386/xsavecintrin.h: Ditto.
1386 * config/i386/xsavesintrin.h: Ditto.
1387 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1388 (bit_XSAVES): Ditto.
1389 (bit_XSAVES): Ditto.
1390 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1391 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1393 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1394 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1395 OPTION_MASK_ISA_XSAVES.
1396 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1398 (PTA_CLFLUSHOPT) Define.
1399 (PTA_XSAVEC): Ditto.
1400 (PTA_XSAVES): Ditto.
1401 (ix86_option_override_internal): Handle new options.
1402 (ix86_valid_target_attribute_inner_p): Ditto.
1403 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1404 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1405 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1406 (bdesc_special_args): Add __builtin_ia32_xsaves,
1407 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1408 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1409 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1410 (ix86_expand_builtin): Handle new builtins.
1411 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1412 (TARGET_CLFLUSHOPT_P): Ditto.
1413 (TARGET_XSAVEC): Ditto.
1414 (TARGET_XSAVEC_P): Ditto.
1415 (TARGET_XSAVES): Ditto.
1416 (TARGET_XSAVES_P): Ditto.
1417 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1418 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1419 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1421 (ANY_XRSTOR64): Ditto.
1423 (xrstor): Change into <xrstor>.
1424 (xrstor_rex64): Change into <xrstor>_rex64.
1425 (xrstor64): Change into <xrstor>64
1427 * config/i386/i386.opt (mclflushopt): New.
1430 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1432 * doc/invoke.texi: Document new options.
1434 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1436 PR rtl-optimization/60866
1437 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1438 Default it to -1. Pass it down to init_simplejump_data.
1439 (init_simplejump_data): New parameter old_seqno. Pass it down
1440 to get_seqno_for_a_jump.
1441 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1442 initializing new jump seqno as a last resort. Add comment.
1443 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1444 jump and pass it down to sel_init_new_insn.
1445 (sel_redirect_edge_and_branch_force): Likewise.
1447 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1449 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1450 shifted values to avoid build warning.
1452 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1454 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1455 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1456 (cfg_layout_merge_blocks): Likewise.
1457 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1459 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1461 PR rtl-optimization/60901
1462 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1463 bb predecessor belongs to the same scheduling region. Adjust comment.
1465 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1467 * doc/sourcebuild.texi: (dfp_hw): Document.
1468 (p8vector_hw): Likewise.
1469 (powerpc_eabi_ok): Likewise.
1470 (powerpc_elfv2): Likewise.
1471 (powerpc_htm_ok): Likewise.
1472 (ppc_recip_hw): Likewise.
1475 2014-05-13 Cary Coutant <ccoutant@google.com>
1477 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1479 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1481 * gengtype-parse.c (require3): Eliminate in favor of...
1483 (require_template_declaration): Update to support optional single *
1486 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1487 (create_user_defined_type): Handle a single level of explicit
1488 pointerness within template arguments.
1489 (struct write_types_data): Add field "kind".
1490 (filter_type_name): Handle "*" character.
1491 (write_user_func_for_structure_ptr): Require a write_types_data
1492 rather than just a prefix string, so that we can look up the kind
1493 of the wtd and use it as an index into wrote_user_func_for_ptr,
1494 ensuring that such functions are written at most once. Support
1495 subclasses by invoking the marking function of the ultimate base class.
1496 (write_user_func_for_structure_body): Require a write_types_data
1497 rather than just a prefix string, so that we can pass this to
1498 write_user_func_for_structure_ptr.
1499 (write_func_for_structure): Likewise.
1500 (ggc_wtd): Add initializer of new "kind" field.
1501 (pch_wtd): Likewise.
1503 * gengtype.h (enum write_types_kinds): New.
1504 (struct type): Add field wrote_user_func_for_ptr to the "s"
1507 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1509 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1510 instead of const_binop.
1511 (fold_binary_loc): Likewise.
1513 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1515 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1516 calculation to match get_ref_base_and_extent.
1518 2014-05-13 Catherine Moore <clm@codesourcery.com>
1519 Sandra Loosemore <sandra@codesourcery.com>
1521 * configure.ac: Fix assembly for explicit JALR relocation check.
1522 * configure: Regenerate.
1524 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1526 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1527 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1528 Remove associated type declarations and initialisations.
1529 (arm_expand_neon_builtin): Likewise.
1530 (neon_emit_pair_result_insn): Delete.
1531 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1532 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1533 (neon_vzip<mode>): Likewise.
1534 (neon_vuzp<mode>): Likewise.
1536 2014-05-13 Richard Biener <rguenther@suse.de>
1539 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1540 it needs revisiting whether the call still may be tail-called.
1542 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1544 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1545 * rtl.h (block_symbol): Reduce number of fields to 2.
1546 (rtx_def): Add u2.symbol_ref_flags.
1547 (SYMBOL_REF_FLAGS): Use it.
1548 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1549 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1550 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1551 Lower index of SYMBOL_REF_DATA.
1552 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1553 Print SYMBOL_REF_FLAGS at the same time.
1554 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1556 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1558 * rtl.def (VAR_LOCATION): Remove "i" field.
1559 * rtl.h (rtx_def): Add u2.var_location_status.
1560 (PAT_VAR_LOCATION_STATUS): Use it.
1561 (gen_rtx_VAR_LOCATION): Declare.
1562 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1563 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1564 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1566 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1568 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1569 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1571 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1573 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1574 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1575 * rtl.h (rtx_def): Add insn_uid to u2 field.
1576 (RTX_FLAG_CHECK8): Delete in favor of...
1577 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1578 (INSN_DELETED_P): Update accordingly.
1579 (INSN_UID): Use u2.insn_uid.
1580 (INSN_CHAIN_CODE_P): Define.
1581 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1582 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1583 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1584 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1585 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1586 indices accordingly.
1587 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1588 Update indices for insn-chain rtxes.
1589 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1590 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1591 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1592 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1593 * combine.c (try_combine): Likewise.
1594 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1596 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1598 * rtl.def (REG): Remove middle field.
1599 * rtl.h (rtx_def): Add orignal_regno to u2.
1600 (ORIGINAL_REGNO): Use it instead of field 1.
1601 (REG_ATTRS): Lower field index accordingly.
1602 * gengtype.c (adjust_field_rtx_def): Remove handling of
1603 ORIGINAL_REGNO. Move REG_ATTRS index down.
1604 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1605 code that prints the REGNO.
1607 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1609 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1612 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1614 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1616 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1618 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1619 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1621 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1623 * config/aarch64/aarch64-protos.h
1624 (aarch64_hard_regno_caller_save_mode): New prototype.
1625 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1627 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1629 2014-05-13 Christian Bruel <christian.bruel@st.com>
1631 * target.def (mode_switching): New hook vector.
1632 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1633 (mode_exit, modepriority_to_mode): Likewise.
1634 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1635 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1636 * target.h: Include tm.h and hard-reg-set.h.
1637 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1638 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1639 * doc/tm.texi Regenerate.
1640 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1641 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1642 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1643 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1644 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1645 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1646 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1647 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1648 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1649 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1650 (ix86_emit_mode_set): Hookify.
1651 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1653 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1654 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1655 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1656 (epiphany_mode_priority_to_mode): Remove declaration.
1657 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1658 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1659 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1661 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1662 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1663 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1665 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1668 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1669 is const0_rtx, return immediately. Don't test count == 0 when
1672 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1674 * Makefile.in: add shrink-wrap.o.
1675 * config/i386/i386.c: include "shrink-wrap.h"
1676 * function.c: Likewise.
1677 (requires_stack_frame_p, next_block_for_reg,
1678 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1679 dup_block_and_redirect): Move to shrink-wrap.c
1680 (thread_prologue_and_epilogue_insns): Extract three code segments
1681 as functions in shrink-wrap.c
1682 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1684 * shrink-wrap.c: New file.
1685 * shrink-wrap.h: New file.
1687 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1689 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1690 reference to Solaris.
1692 2014-05-12 Mike Stump <mikestump@comcast.net>
1695 * genattrtab.c (filename): Add.
1696 (convert_set_attr_alternative): Improve error message.
1697 (check_defs): Restore read_md_filename for error messages.
1698 (gen_insn): Save filename.
1700 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1702 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1703 -fno-local-ivars and -fivar-visibility.
1704 * c-family/c.opt: Make -Wshadow also implicitly enable
1707 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1709 * doc/tm.texi: Remove reference to deleted macro.
1710 * doc/tm.texi.in: Likewise.
1712 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1715 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1718 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1720 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1721 * config/arm/aout.h (License): Same.
1722 * config/arm/bpabi.h (License): Same.
1723 * config/arm/elf.h (License): Same.
1724 * config/arm/linux-elf.h (License): Same.
1725 * config/arm/linux-gas.h (License): Same.
1726 * config/arm/netbsd-elf.h (License): Same.
1727 * config/arm/uclinux-eabi.h (License): Same.
1728 * config/arm/uclinux-elf.h (License): Same.
1729 * config/arm/vxworks.h (License): Same.
1731 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1733 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1734 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1735 number of operands to 3.
1736 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1737 * tree-nested.c (convert_nonlocal_omp_clauses,
1738 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1739 * gimplify.c (gimplify_scan_omp_clauses): Handle
1740 OMP_CLAUSE_LINEAR_STMT.
1741 * omp-low.c (lower_rec_input_clauses): Fix typo.
1742 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1743 cast between Fortran boolean_type_node and C _Bool if
1746 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1748 PR tree-optimization/61136
1749 * wide-int.h (multiple_of_p): Define a version that doesn't return
1751 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1752 integer_zerop/const_binop pair.
1753 (multiple_of_p): Likewise, converting both operands to widest_int
1756 2014-05-09 Teresa Johnson <tejohnson@google.com>
1758 * cgraphunit.c (analyze_functions): Use correct dump file.
1760 2014-05-09 Florian Weimer <fweimer@redhat.com>
1762 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1764 (stack_protect_return_slot_p): New function.
1765 (expand_used_vars): Call stack_protect_decl_p and
1766 stack_protect_return_slot_p for -fstack-protector-strong.
1768 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1771 * fold-const.c (fold_binary_loc): Changed width of mask.
1773 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1775 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1776 unsigned int initializers for regno_in, regno_out.
1778 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1781 * config/avr/avr.md (cc): Add new attribute set_vzn.
1782 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1783 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1784 with INC, DEC or NEG.
1785 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1786 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1787 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1789 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1792 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1794 * wide-int.cc (UTItype): Define.
1795 (UDWtype): Define for appropriate W_TYPE_SIZE.
1797 2014-05-09 Richard Biener <rguenther@suse.de>
1799 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1800 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1801 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1802 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1803 ssa_propagate): Adjust.
1805 2014-05-08 Jeff Law <law@redhat.com>
1807 PR tree-optimization/61009
1808 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1809 tri-state rather than a boolean. When a block is too big to
1810 thread through, inform caller via negative return value.
1811 (thread_across_edge): If a block was too big for normal threading,
1812 then it's too big for a joiner too, so remove temporary equivalences
1813 and return immediately.
1815 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1816 Matthias Klose <doko@ubuntu.com>
1819 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1821 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1824 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1826 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1829 * config/alpha/alpha.c: Include gimple-iterator.h.
1830 (alpha_gimple_fold_builtin): New function. Move
1831 ALPHA_BUILTIN_UMULH folding from ...
1832 (alpha_fold_builtin): ... here.
1833 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1835 2014-05-08 Wei Mi <wmi@google.com>
1838 * config/i386/i386.c (ix86_compute_frame_layout): Update
1839 preferred_stack_boundary for call, expanded from tls descriptor.
1840 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1841 to depend on SP register.
1842 (*tls_local_dynamic_base_32_gnu): Ditto.
1843 (*tls_local_dynamic_32_once): Ditto.
1844 (tls_global_dynamic_64_<mode>): Set
1845 ix86_tls_descriptor_calls_expanded_in_cfun.
1846 (tls_local_dynamic_base_64_<mode>): Ditto.
1847 (tls_global_dynamic_32): Set
1848 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1849 to depend on SP register.
1850 (tls_local_dynamic_base_32): Ditto.
1852 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1854 * config/arm/arm_neon.h: Update comment.
1855 * config/arm/neon-docgen.ml: Delete.
1856 * config/arm/neon-gen.ml: Delete.
1857 * doc/arm-neon-intrinsics.texi: Update comment.
1859 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1861 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1863 (vand, vorr, veor, vorn, vbic): Remove.
1864 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1866 (neon_vsub_unspec): Likewise.
1867 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1869 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1871 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1872 (vadd_s16): Likewise.
1873 (vadd_s32): Likewise.
1874 (vadd_f32): Likewise.
1875 (vadd_u8): Likewise.
1876 (vadd_u16): Likewise.
1877 (vadd_u32): Likewise.
1878 (vadd_s64): Likewise.
1879 (vadd_u64): Likewise.
1880 (vaddq_s8): Likewise.
1881 (vaddq_s16): Likewise.
1882 (vaddq_s32): Likewise.
1883 (vaddq_s64): Likewise.
1884 (vaddq_f32): Likewise.
1885 (vaddq_u8): Likewise.
1886 (vaddq_u16): Likewise.
1887 (vaddq_u32): Likewise.
1888 (vaddq_u64): Likewise.
1889 (vmul_s8): Likewise.
1890 (vmul_s16): Likewise.
1891 (vmul_s32): Likewise.
1892 (vmul_f32): Likewise.
1893 (vmul_u8): Likewise.
1894 (vmul_u16): Likewise.
1895 (vmul_u32): Likewise.
1896 (vmul_p8): Likewise.
1897 (vmulq_s8): Likewise.
1898 (vmulq_s16): Likewise.
1899 (vmulq_s32): Likewise.
1900 (vmulq_f32): Likewise.
1901 (vmulq_u8): Likewise.
1902 (vmulq_u16): Likewise.
1903 (vmulq_u32): Likewise.
1904 (vsub_s8): Likewise.
1905 (vsub_s16): Likewise.
1906 (vsub_s32): Likewise.
1907 (vsub_f32): Likewise.
1908 (vsub_u8): Likewise.
1909 (vsub_u16): Likewise.
1910 (vsub_u32): Likewise.
1911 (vsub_s64): Likewise.
1912 (vsub_u64): Likewise.
1913 (vsubq_s8): Likewise.
1914 (vsubq_s16): Likewise.
1915 (vsubq_s32): Likewise.
1916 (vsubq_s64): Likewise.
1917 (vsubq_f32): Likewise.
1918 (vsubq_u8): Likewise.
1919 (vsubq_u16): Likewise.
1920 (vsubq_u32): Likewise.
1921 (vsubq_u64): Likewise.
1922 (vand_s8): Likewise.
1923 (vand_s16): Likewise.
1924 (vand_s32): Likewise.
1925 (vand_u8): Likewise.
1926 (vand_u16): Likewise.
1927 (vand_u32): Likewise.
1928 (vand_s64): Likewise.
1929 (vand_u64): Likewise.
1930 (vandq_s8): Likewise.
1931 (vandq_s16): Likewise.
1932 (vandq_s32): Likewise.
1933 (vandq_s64): Likewise.
1934 (vandq_u8): Likewise.
1935 (vandq_u16): Likewise.
1936 (vandq_u32): Likewise.
1937 (vandq_u64): Likewise.
1938 (vorr_s8): Likewise.
1939 (vorr_s16): Likewise.
1940 (vorr_s32): Likewise.
1941 (vorr_u8): Likewise.
1942 (vorr_u16): Likewise.
1943 (vorr_u32): Likewise.
1944 (vorr_s64): Likewise.
1945 (vorr_u64): Likewise.
1946 (vorrq_s8): Likewise.
1947 (vorrq_s16): Likewise.
1948 (vorrq_s32): Likewise.
1949 (vorrq_s64): Likewise.
1950 (vorrq_u8): Likewise.
1951 (vorrq_u16): Likewise.
1952 (vorrq_u32): Likewise.
1953 (vorrq_u64): Likewise.
1954 (veor_s8): Likewise.
1955 (veor_s16): Likewise.
1956 (veor_s32): Likewise.
1957 (veor_u8): Likewise.
1958 (veor_u16): Likewise.
1959 (veor_u32): Likewise.
1960 (veor_s64): Likewise.
1961 (veor_u64): Likewise.
1962 (veorq_s8): Likewise.
1963 (veorq_s16): Likewise.
1964 (veorq_s32): Likewise.
1965 (veorq_s64): Likewise.
1966 (veorq_u8): Likewise.
1967 (veorq_u16): Likewise.
1968 (veorq_u32): Likewise.
1969 (veorq_u64): Likewise.
1970 (vbic_s8): Likewise.
1971 (vbic_s16): Likewise.
1972 (vbic_s32): Likewise.
1973 (vbic_u8): Likewise.
1974 (vbic_u16): Likewise.
1975 (vbic_u32): Likewise.
1976 (vbic_s64): Likewise.
1977 (vbic_u64): Likewise.
1978 (vbicq_s8): Likewise.
1979 (vbicq_s16): Likewise.
1980 (vbicq_s32): Likewise.
1981 (vbicq_s64): Likewise.
1982 (vbicq_u8): Likewise.
1983 (vbicq_u16): Likewise.
1984 (vbicq_u32): Likewise.
1985 (vbicq_u64): Likewise.
1986 (vorn_s8): Likewise.
1987 (vorn_s16): Likewise.
1988 (vorn_s32): Likewise.
1989 (vorn_u8): Likewise.
1990 (vorn_u16): Likewise.
1991 (vorn_u32): Likewise.
1992 (vorn_s64): Likewise.
1993 (vorn_u64): Likewise.
1994 (vornq_s8): Likewise.
1995 (vornq_s16): Likewise.
1996 (vornq_s32): Likewise.
1997 (vornq_s64): Likewise.
1998 (vornq_u8): Likewise.
1999 (vornq_u16): Likewise.
2000 (vornq_u32): Likewise.
2001 (vornq_u64): Likewise.
2003 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2005 * wide-int.cc (UTItype): Define.
2006 (UDWtype): Define for appropriate W_TYPE_SIZE.
2008 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2010 PR tree-optimization/59100
2011 * tree-ssa-phiopt.c: Include tree-inline.h.
2012 (neutral_element_p, absorbing_element_p): New functions.
2013 (value_replacement): Handle conditional binary operations with a
2014 neutral or absorbing element.
2016 2014-05-08 Richard Biener <rguenther@suse.de>
2018 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2019 folding the expression.
2020 (valueize_expr): Remove.
2021 (visit_reference_op_load): Do not valueize the result of
2023 (simplify_binary_expression): Likewise.
2024 (simplify_unary_expression): Likewise.
2026 2014-05-08 Richard Biener <rguenther@suse.de>
2028 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2029 looking at TYPE_ARG_TYPES.
2031 2014-05-08 Richard Biener <rguenther@suse.de>
2033 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2034 pointer propagation special-case.
2036 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2038 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2039 core part of address expressions.
2041 2014-05-08 Alan Modra <amodra@gmail.com>
2044 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2045 loads and stores when -mno-strict-align at any alignment.
2046 (expand_block_clear): Similarly. Also correct calculation of
2049 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2052 * tree-complex.c (expand_complex_move): Keep line info when expanding
2054 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2055 of complex expression. Use new argument to display correct location
2056 for values coming from phi statement.
2057 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2058 (warn_uninitialized_phi): Pass location of phi argument to
2060 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2061 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2063 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2065 * config/rs6000/predicates.md (indexed_address_mem): New.
2066 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2067 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2068 fpstore_ux, fpstore_u.
2069 (sign_extend, indexed, update): New.
2070 (cell_micro): Adjust.
2071 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2072 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2073 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2074 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2075 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2076 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2077 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2078 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2079 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2080 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2081 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2082 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2083 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2084 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2085 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2087 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2088 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2089 *vsx_extract_<mode>_store): Adjust.
2090 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2091 is_cracked_insn, insn_must_be_first_in_group,
2092 insn_must_be_last_in_group): Adjust.
2094 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2096 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2097 ppc440-fpstore): Adjust.
2098 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2099 ppc476-fpstore): Adjust.
2100 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2101 ppc601-fpstore): Adjust.
2102 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2104 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2106 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2107 ppc7450-fpstore): Adjust.
2108 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2109 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2110 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2112 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2113 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2114 cell-fpstore, cell-fpstore-update): Adjust.
2115 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2116 ppce300c3_store, ppce300c3_fpstore): Adjust.
2117 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2118 e500mc_fpstore): Adjust.
2119 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2120 e500mc64_store, e500mc64_fpstore): Adjust.
2121 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2122 e5500_fpstore): Adjust.
2123 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2124 e6500_fpstore): Adjust.
2125 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2127 * config/rs6000/power4.md (power4-load, power4-load-ext,
2128 power4-load-ext-update, power4-load-ext-update-indexed,
2129 power4-load-update-indexed, power4-load-update, power4-fpload,
2130 power4-fpload-update, power4-store, power4-store-update,
2131 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2133 * config/rs6000/power5.md (power5-load, power5-load-ext,
2134 power5-load-ext-update, power5-load-ext-update-indexed,
2135 power5-load-update-indexed, power5-load-update, power5-fpload,
2136 power5-fpload-update, power5-store, power5-store-update,
2137 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2139 * config/rs6000/power6.md (power6-load, power6-load-ext,
2140 power6-load-update, power6-load-update-indexed,
2141 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2142 power6-fpload-update, power6-store, power6-store-update,
2143 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2145 * config/rs6000/power7.md (power7-load, power7-load-ext,
2146 power7-load-update, power7-load-update-indexed,
2147 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2148 power7-fpload-update, power7-store, power7-store-update,
2149 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2151 * config/rs6000/power8.md (power8-load, power8-load-update,
2152 power8-load-ext, power8-load-ext-update, power8-fpload,
2153 power8-fpload-update, power8-store, power8-store-update-indexed,
2154 power8-fpstore, power8-fpstore-update): Adjust.
2155 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2157 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2158 titan_lsu_store, titan_lsu_fpstore): Adjust.
2159 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2161 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2164 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2165 unrolled byte insns. Emit address increments after move insns.
2167 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2169 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2170 const_gimple, rather than a gimple.
2171 (gimple_call_builtin_p): Likewise, for the three variants.
2173 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2174 (gimple_call_builtin_p): Likewise, for the three variants.
2176 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2178 PR tree-optimization/61095
2179 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2181 2014-05-07 Richard Biener <rguenther@suse.de>
2183 PR tree-optimization/61034
2184 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2185 (maybe_skip_until): Use translate to take into account
2186 lattices when trying to do disambiguations.
2187 (get_continuation_for_phi_1): Likewise.
2188 (get_continuation_for_phi): Adjust for added translate arguments.
2189 (walk_non_aliased_vuses): Likewise.
2190 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2191 (walk_non_aliased_vuses): Likewise.
2192 (call_may_clobber_ref_p_1): Declare.
2193 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2194 calls. Stop early if we are only supposed to disambiguate.
2195 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2197 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2199 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2200 Emit an error when the function has arguments.
2202 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2204 * cfgloop.h (unswitch_loops): Remove.
2205 * doc/passes.texi: Remove references to loop-unswitch.c
2206 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2208 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2210 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2211 check for loads group of length 3.
2212 (vect_permute_load_chain): New permutations for loads group of
2214 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2215 of vec_perm_shuffle for the new permutations.
2217 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2219 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2220 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2221 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2222 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2223 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2224 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2225 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2226 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2228 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2230 * loop-unswitch.c: Delete.
2232 2014-05-07 Richard Biener <rguenther@suse.de>
2234 * config.gcc: Always set need_64bit_hwint to yes.
2236 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2238 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2239 of using optimize_size.
2241 2014-05-06 Mike Stump <mikestump@comcast.net>
2243 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2245 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2247 * config/i386/sse.md (*mov<mode>_internal)
2248 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2249 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2250 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2251 (*<code><mode>3, *andnot<mode>3<mask_name>)
2252 (<mask_codefor><code><mode>3<mask_name>): Only consider
2253 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2255 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2258 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2260 * lra-constraints.c (valid_address_p): Move earlier in file.
2261 Add a constraint argument to the address_info version.
2262 (satisfies_memory_constraint_p): New function.
2263 (satisfies_address_constraint_p): Likewise.
2264 (process_alt_operands, curr_insn_transform): Use them.
2265 (process_address): Pass the constraint to valid_address_p when
2266 checking address operands.
2268 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2270 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2271 to their respective blocks. Fix inadvertent use of "node".
2273 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2275 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2277 (init_emit_once): ...here.
2278 * rtl.h (init_derived_machine_modes): Declare.
2279 * toplev.c (do_compile): Call it even if no_backend.
2281 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2282 Mike Stump <mikestump@comcast.net>
2283 Richard Sandiford <rdsandiford@googlemail.com>
2284 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2286 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2287 (rtx_equal_for_memref_p): Update comment.
2288 (adjust_offset_for_component_ref): Use wide-int interfaces.
2289 * builtins.c (get_object_alignment_2): Likewise.
2290 (c_readstr): Likewise.
2291 (target_char_cast): Add comment.
2292 (determine_block_size): Use wide-int interfaces.
2293 (expand_builtin_signbit): Likewise.
2294 (fold_builtin_int_roundingfn): Likewise.
2295 (fold_builtin_bitop): Likewise.
2296 (fold_builtin_bswap): Likewise.
2297 (fold_builtin_logarithm): Use signop.
2298 (fold_builtin_pow): Likewise.
2299 (fold_builtin_memory_op): Use wide-int interfaces.
2300 (fold_builtin_object_size): Likewise.
2301 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2302 nb_iterations_estimate.
2303 (record_niter_bound): Use wide-int interfaces.
2304 (get_estimated_loop_iterations_int): Likewise.
2305 (get_estimated_loop_iterations): Likewise.
2306 (get_max_loop_iterations): Likewise.
2307 * cfgloop.h: Include wide-int.h.
2308 (struct nb_iter_bound): Change bound to widest_int.
2309 (struct loop): Change nb_iterations_upper_bound and
2310 nb_iterations_estimate to widest_int.
2311 (record_niter_bound): Switch to use widest_int.
2312 (get_estimated_loop_iterations): Likewise.
2313 (get_max_loop_iterations): Likewise.
2314 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2315 update for wide-int.
2316 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2317 * combine.c (try_combine): Likewise.
2318 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2319 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2321 (aarch64_float_const_representable_p): Likewise.
2322 * config/arc/arc.c: Include wide-int.h.
2323 (arc_can_use_doloop_p): Use wide-int interfaces.
2324 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2325 (vfp3_const_double_index): Likewise.
2326 * config/avr/avr.c (avr_out_round): Likewise.
2327 (avr_fold_builtin): Likewise.
2328 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2329 (bfin_can_use_doloop_p): Likewise.
2330 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2331 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2332 * config/i386/i386.c: Include wide-int.h.
2333 (ix86_data_alignment): Use wide-int interfaces.
2334 (ix86_local_alignment): Likewise.
2335 (ix86_emit_swsqrtsf): Update real_from_integer.
2336 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2337 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2338 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2339 (zero_constant): Likewise.
2340 (input_operand): Likewise.
2341 (splat_input_operand): Likewise.
2342 (non_logical_cint_operand): Change const_double to const_wide_int.
2343 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2344 (easy_altivec_constant): Remove comment.
2345 (paired_expand_vector_init): Use CONSTANT_P.
2346 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2347 (rs6000_emit_move): Update checks.
2348 (rs6000_aggregate_candidate): Use wide-int interfaces.
2349 (rs6000_expand_ternop_builtin): Likewise.
2350 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2351 (rs6000_assemble_integer): Likewise.
2352 (rs6000_hash_constant): Likewise.
2353 (output_toc): Likewise.
2354 (rs6000_rtx_costs): Likewise.
2355 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2356 * config/rs6000/rs6000-c.c: Include wide-int.h.
2357 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2358 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2359 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2360 Handle CONST_WIDE_INT.
2361 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2362 Use tree_fits_uhwi_p.
2363 * config/sparc/sparc.c: Include wide-int.h.
2364 (sparc_fold_builtin): Use wide-int interfaces.
2365 * config/vax/vax.c: Include wide-int.h.
2366 (vax_float_literal): Use real_from_integer.
2367 * coretypes.h (struct hwivec_def): New.
2369 (const_hwivec): New.
2370 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2371 (equiv_constant): Handle CONST_WIDE_INT.
2372 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2373 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2374 * dbxout.c (stabstr_U): Use wide-int interfaces.
2375 (dbxout_type): Update to use cst_fits_shwi_p.
2376 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2377 (TARGET_SUPPORTS_WIDE_INT): Add default.
2378 * dfp.c: Include wide-int.h.
2379 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2380 decimal_real_to_integer.
2381 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2382 decimal_real_to_integer.
2383 * doc/generic.texi (Constant expressions): Update for wide_int.
2384 * doc/rtl.texi (const_double): Likewise.
2385 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2386 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2387 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2388 (REAL_VALUE_FROM_INT): Remove.
2389 (TARGET_SUPPORTS_WIDE_INT): New.
2390 * doc/tm.texi: Regenerate.
2391 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2392 * double-int.h: Include wide-int.h.
2393 (struct wi::int_traits): New.
2394 * dwarf2out.c (get_full_len): New.
2395 (dw_val_equal_p): Add case dw_val_class_wide_int.
2396 (size_of_loc_descr): Likewise.
2397 (output_loc_operands): Likewise.
2398 (insert_double): Remove.
2399 (insert_wide_int): New.
2401 (print_die): Add case dw_val_class_wide_int.
2402 (attr_checksum): Likewise.
2403 (attr_checksum_ordered): Likewise.
2404 (same_dw_val_p): Likewise.
2405 (size_of_die): Likewise.
2406 (value_format): Likewise.
2407 (output_die): Likewise.
2408 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2410 (clz_loc_descriptor): Use wide-int interfaces.
2411 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2412 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2413 (round_up_to_align): Use wide-int interfaces.
2414 (field_byte_offset): Likewise.
2415 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2416 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2417 CONST_DOUBLE handling. Use wide-int interfaces.
2418 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2419 (gen_enumeration_type_die): Use add_AT_wide.
2420 (hash_loc_operands): Add case dw_val_class_wide_int.
2421 (compare_loc_operands): Likewise.
2422 * dwarf2out.h: Include wide-int.h.
2423 (wide_int_ptr): New.
2424 (enum dw_val_class): Add dw_val_class_wide_int.
2425 (struct dw_val_struct): Add val_wide.
2426 * emit-rtl.c (const_wide_int_htab): New.
2427 (const_wide_int_htab_hash): New.
2428 (const_wide_int_htab_eq): New.
2429 (lookup_const_wide_int): New.
2430 (const_double_htab_hash): Use wide-int interfaces.
2431 (const_double_htab_eq): Likewise.
2432 (rtx_to_double_int): Conditionally compile for wide-int.
2433 (immed_double_int_const): Rename to immed_wide_int_const and
2434 update for wide-int.
2435 (immed_double_const): Conditionally compile for wide-int.
2436 (init_emit_once): Use wide-int interfaces.
2437 * explow.c (plus_constant): Likewise.
2438 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2439 (lshift_value): Use wide-int interfaces.
2440 (expand_mult): Likewise.
2441 (choose_multiplier): Likewise.
2442 (expand_smod_pow2): Likewise.
2443 (make_tree): Likewise.
2444 * expr.c (convert_modes): Consolidate handling of constants.
2445 Use wide-int interfaces.
2446 (emit_group_load_1): Add note.
2447 (store_expr): Update comment.
2448 (get_inner_reference): Use wide-int interfaces.
2449 (expand_constructor): Update comment.
2450 (expand_expr_real_2): Use wide-int interfaces.
2451 (expand_expr_real_1): Likewise.
2452 (reduce_to_bit_field_precision): Likewise.
2453 (const_vector_from_tree): Likewise.
2454 * final.c: Include wide-int-print.h.
2455 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2456 * fixed-value.c: Include wide-int.h.
2457 (fixed_from_string): Use wide-int interfaces.
2458 (fixed_to_decimal): Likewise.
2459 (fixed_convert_from_real): Likewise.
2460 (real_convert_from_fixed): Likewise.
2461 * fold-const.h (mem_ref_offset): Return an offset_int.
2462 (div_if_zero_remainder): Remove code parameter.
2463 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2464 Use wide-int interfaces.
2465 (may_negate_without_overflow_p): Use wide-int interfaces.
2466 (negate_expr_p): Likewise.
2467 (fold_negate_expr): Likewise.
2468 (int_const_binop_1): Likewise.
2469 (const_binop): Likewise.
2470 (fold_convert_const_int_from_int): Likewise.
2471 (fold_convert_const_int_from_real): Likewise.
2472 (fold_convert_const_int_from_fixed): Likewise.
2473 (fold_convert_const_fixed_from_int): Likewise.
2474 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2475 (sign_bit_p): Use wide-int interfaces.
2476 (make_range_step): Likewise.
2477 (build_range_check): Likewise. Pass an integer of the correct type
2478 instead of using integer_one_node.
2479 (range_predecessor): Pass an integer of the correct type instead
2480 of using integer_one_node.
2481 (range_successor): Likewise.
2482 (merge_ranges): Likewise.
2483 (unextend): Use wide-int interfaces.
2484 (extract_muldiv_1): Likewise.
2485 (fold_div_compare): Likewise.
2486 (fold_single_bit_test): Likewise.
2487 (fold_sign_changed_comparison): Likewise.
2488 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2489 (fold_plusminus_mult_expr): Use wide-int interfaces.
2490 (native_encode_int): Likewise.
2491 (native_interpret_int): Likewise.
2492 (fold_unary_loc): Likewise.
2493 (pointer_may_wrap_p): Likewise.
2494 (size_low_cst): Likewise.
2495 (mask_with_tz): Likewise.
2496 (fold_binary_loc): Likewise.
2497 (fold_ternary_loc): Likewise.
2498 (multiple_of_p): Likewise.
2499 (tree_call_nonnegative_warnv_p): Update calls to
2500 tree_int_cst_min_precision and real_from_integer.
2501 (fold_negate_const): Use wide-int interfaces.
2502 (fold_abs_const): Likewise.
2503 (fold_relational_const): Use tree_int_cst_lt.
2504 (round_up_loc): Use wide-int interfaces.
2505 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2506 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2507 * gengtype.c: Remove include of double-int.h.
2508 (do_typedef): Use wide-int interfaces.
2509 (open_base_files): Add wide-int.h.
2510 (main): Add offset_int and widest_int typedefs.
2511 * gengtype-lex.l: Handle "^".
2512 (CXX_KEYWORD): Add "static".
2513 * gengtype-parse.c (require3): New.
2514 (require_template_declaration): Handle constant template arguments
2515 and nested templates.
2516 * gengtype-state.c: Don't include "double-int.h".
2517 * genpreds.c (write_one_predicate_function): Update comment.
2518 (write_tm_constrs_h): Add check for hval and lval use in
2520 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2521 (add_to_sequence): Likewise.
2522 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2523 and const_double_operand.
2524 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2526 * gimple-fold.c (get_base_constructor): Likewise.
2527 (fold_array_ctor_reference): Likewise.
2528 (fold_nonarray_ctor_reference): Likewise.
2529 (fold_const_aggregate_ref_1): Likewise.
2530 (gimple_val_nonnegative_real_p): Likewise.
2531 (gimple_fold_indirect_ref): Likewise.
2532 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2533 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2534 (struct slsr_cand_d): Change index to be widest_int.
2535 (struct incr_info_d): Change incr to be widest_int.
2536 (alloc_cand_and_find_basis): Use wide-int interfaces.
2537 (slsr_process_phi): Likewise.
2538 (backtrace_base_for_ref): Likewise. Return a widest_int.
2539 (restructure_reference): Take a widest_int instead of a double_int.
2540 (slsr_process_ref): Use wide-int interfaces.
2541 (create_mul_ssa_cand): Likewise.
2542 (create_mul_imm_cand): Likewise.
2543 (create_add_ssa_cand): Likewise.
2544 (create_add_imm_cand): Take a widest_int instead of a double_int.
2545 (slsr_process_add): Use wide-int interfaces.
2546 (slsr_process_cast): Likewise.
2547 (slsr_process_copy): Likewise.
2548 (dump_candidate): Likewise.
2549 (dump_incr_vec): Likewise.
2550 (replace_ref): Likewise.
2551 (cand_increment): Likewise. Return a widest_int.
2552 (cand_abs_increment): Likewise.
2553 (replace_mult_candidate): Take a widest_int instead of a double_int.
2554 (replace_unconditional_candidate): Use wide-int interfaces.
2555 (incr_vec_index): Take a widest_int instead of a double_int.
2556 (create_add_on_incoming_edge): Likewise.
2557 (create_phi_basis): Use wide-int interfaces.
2558 (replace_conditional_candidate): Likewise.
2559 (record_increment): Take a widest_int instead of a double_int.
2560 (record_phi_increments): Use wide-int interfaces.
2561 (phi_incr_cost): Take a widest_int instead of a double_int.
2562 (lowest_cost_path): Likewise.
2563 (total_savings): Likewise.
2564 (analyze_increments): Use wide-int interfaces.
2565 (ncd_with_phi): Take a widest_int instead of a double_int.
2566 (ncd_of_cand_and_phis): Likewise.
2567 (nearest_common_dominator_for_cands): Likewise.
2568 (insert_initializers): Use wide-int interfaces.
2569 (all_phi_incrs_profitable): Likewise.
2570 (replace_one_candidate): Likewise.
2571 (replace_profitable_candidates): Likewise.
2572 * godump.c: Include wide-int-print.h.
2573 (go_output_typedef): Use wide-int interfaces.
2574 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2575 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2576 (build_loop_iteration_domains): Likewise.
2577 * hooks.h: Include wide-int.h rather than double-int.h.
2578 (hook_bool_dint_dint_uint_bool_true): Delete.
2579 (hook_bool_wint_wint_uint_bool_true): Declare.
2580 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2581 (hook_bool_wint_wint_uint_bool_true): New.
2582 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2584 (ubsan_expand_si_overflow_mul_check): Likewise.
2585 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2586 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2587 (get_ancestor_addr_info): Likewise.
2588 (ipa_modify_call_arguments): Likewise.
2589 * loop-doloop.c (doloop_modify): Likewise.
2590 (doloop_optimize): Likewise.
2591 * loop-iv.c (iv_number_of_iterations): Likewise.
2592 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2593 (unroll_loop_constant_iterations): Likewise.
2594 (decide_unroll_runtime_iterations): Likewise.
2595 (unroll_loop_runtime_iterations): Likewise.
2596 (decide_peel_simple): Likewise.
2597 (decide_unroll_stupid): Likewise.
2598 * lto-streamer-in.c (streamer_read_wi): Add.
2599 (input_cfg): Use wide-int interfaces.
2600 (lto_input_tree_1): Likewise.
2601 * lto-streamer-out.c (streamer_write_wi): Add.
2602 (hash_tree): Use wide-int interfaces.
2603 (output_cfg): Likewise.
2604 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2605 (GTFILES): Add wide-int.h and signop.h.
2606 (TAGS): Look for .cc files too.
2607 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2608 * optabs.c (expand_subword_shift): Likewise.
2609 (expand_doubleword_shift): Likewise.
2610 (expand_absneg_bit): Likewise.
2611 (expand_copysign_absneg): Likewise.
2612 (expand_copysign_bit): Likewise.
2613 * postreload.c (reload_cse_simplify_set): Likewise.
2614 * predict.c (predict_iv_comparison): Likewise.
2615 * pretty-print.h: Include wide-int-print.h.
2617 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2618 * print-tree.c: Include wide-int-print.h.
2619 (print_node_brief): Use wide-int interfaces.
2620 (print_node): Likewise.
2621 * read-rtl.c (validate_const_wide_int): New.
2622 (read_rtx_code): Add CONST_WIDE_INT case.
2623 * real.c: Include wide-int.h.
2624 (real_to_integer2): Delete.
2625 (real_to_integer): New function, returning a wide_int.
2626 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2627 (ten_to_ptwo): Update call to real_from_integer.
2628 (real_digit): Likewise.
2629 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2630 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2631 (REAL_VALUE_TO_INT): Delete.
2632 (real_to_integer): Declare a wide-int form.
2633 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2634 * recog.c (const_int_operand): Improve comment.
2635 (const_scalar_int_operand): New.
2636 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2637 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2638 (split_double): Likewise.
2639 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2640 (rtx_size): Likewise.
2641 (rtx_alloc_stat_v): New.
2642 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2643 (cwi_output_hex): New.
2644 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2645 (cwi_check_failed_bounds): New.
2646 * rtl.def (CONST_WIDE_INT): New.
2647 * rtl.h: Include <utility> and wide-int.h.
2648 (struct hwivec_def): New.
2649 (CWI_GET_NUM_ELEM): New.
2650 (CWI_PUT_NUM_ELEM): New.
2651 (struct rtx_def): Add num_elem and hwiv.
2652 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2653 (CASE_CONST_UNIQUE): Likewise.
2654 (CASE_CONST_ANY): Likewise.
2655 (CONST_SCALAR_INT_P): Likewise.
2656 (CONST_WIDE_INT_P): New.
2658 (HWIVEC_CHECK): New.
2659 (cwi_check_failed_bounds): New.
2661 (HWIVEC_CHECK): New.
2662 (CONST_WIDE_INT_VEC) New.
2663 (CONST_WIDE_INT_NUNITS) New.
2664 (CONST_WIDE_INT_ELT) New.
2665 (rtx_mode_t): New type.
2666 (wi::int_traits <rtx_mode_t>): New.
2668 (wi::min_value): New.
2669 (wi::max_value): New.
2671 (const_wide_int_alloc): New.
2672 (immed_wide_int_const): New.
2673 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2674 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2675 * signop.h: New file.
2676 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2677 (simplify_const_unary_operation): Use wide-int interfaces.
2678 (simplify_binary_operation_1): Likewise.
2679 (simplify_const_binary_operation): Likewise.
2680 (simplify_const_relational_operation): Likewise.
2681 (simplify_immed_subreg): Likewise.
2682 * stmt.c (expand_case): Likewise.
2683 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2684 signop rather than a bool.
2685 * stor-layout.c (layout_type): Use wide-int interfaces.
2686 (initialize_sizetypes): Update calls to
2687 set_min_and_max_values_for_integral_type.
2688 (set_min_and_max_values_for_integral_type): Take a signop rather
2689 than a bool. Use wide-int interfaces.
2690 (fixup_signed_type): Update accordingly. Remove
2691 HOST_BITS_PER_DOUBLE_INT limit.
2692 (fixup_unsigned_type): Likewise.
2693 * system.h (STATIC_CONSTANT_P): New.
2694 (STATIC_ASSERT): New.
2695 * target.def (can_use_doloop_p): Take widest_ints rather than
2697 * target.h: Include wide-int.h rather than double-int.h.
2698 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2700 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2701 rather than INT_CST_LT_UNSIGNED.
2702 (can_use_doloop_if_innermost): Take widest_ints rather than
2704 * tree-affine.c: Include wide-int-print.h.
2705 (double_int_ext_for_comb): Delete.
2706 (wide_int_ext_for_comb): New.
2707 (aff_combination_zero): Use wide-int interfaces.
2708 (aff_combination_const): Take a widest_int instead of a double_int.
2709 (aff_combination_elt): Use wide-int interfaces.
2710 (aff_combination_scale): Take a widest_int instead of a double_int.
2711 (aff_combination_add_elt): Likewise.
2712 (aff_combination_add_cst): Likewise.
2713 (aff_combination_add): Use wide-int interfaces.
2714 (aff_combination_convert): Likewise.
2715 (tree_to_aff_combination): Likewise.
2716 (add_elt_to_tree): Take a widest_int instead of a double_int.
2717 (aff_combination_to_tree): Use wide-int interfaces.
2718 (aff_combination_remove_elt): Likewise.
2719 (aff_combination_add_product): Take a widest_int instead of
2721 (aff_combination_mult): Use wide-int interfaces.
2722 (aff_combination_expand): Likewise.
2723 (double_int_constant_multiple_p): Delete.
2724 (wide_int_constant_multiple_p): New.
2725 (aff_combination_constant_multiple_p): Take a widest_int pointer
2726 instead of a double_int pointer.
2727 (print_aff): Use wide-int interfaces.
2728 (get_inner_reference_aff): Take a widest_int pointer
2729 instead of a double_int pointer.
2730 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2731 * tree-affine.h: Include wide-int.h.
2732 (struct aff_comb_elt): Change type of coef to widest_int.
2733 (struct affine_tree_combination): Change type of offset to widest_int.
2734 (double_int_ext_for_comb): Delete.
2735 (wide_int_ext_for_comb): New.
2736 (aff_combination_const): Use widest_int instead of double_int.
2737 (aff_combination_scale): Likewise.
2738 (aff_combination_add_elt): Likewise.
2739 (aff_combination_constant_multiple_p): Likewise.
2740 (get_inner_reference_aff): Likewise.
2741 (aff_comb_cannot_overlap_p): Likewise.
2742 (aff_combination_zero_p): Use wide-int interfaces.
2743 * tree.c: Include tree.h.
2744 (init_ttree): Use make_int_cst.
2745 (tree_code_size): Removed code for INTEGER_CST case.
2746 (tree_size): Add INTEGER_CST case.
2747 (make_node_stat): Update comment.
2748 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2749 (build_int_cst_type): Use wide-int interfaces.
2750 (double_int_to_tree): Likewise.
2751 (double_int_fits_to_tree_p): Delete.
2752 (force_fit_type_double): Delete.
2753 (force_fit_type): New.
2754 (int_cst_hash_hash): Use wide-int interfaces.
2755 (int_cst_hash_eq): Likewise.
2756 (build_int_cst_wide): Delete.
2757 (wide_int_to_tree): New.
2758 (cache_integer_cst): Use wide-int interfaces.
2759 (build_low_bits_mask): Likewise.
2760 (cst_and_fits_in_hwi): Likewise.
2761 (real_value_from_int_cst): Likewise.
2762 (make_int_cst_stat): New.
2763 (integer_zerop): Use wide_int interfaces.
2764 (integer_onep): Likewise.
2765 (integer_all_onesp): Likewise.
2766 (integer_pow2p): Likewise.
2767 (integer_nonzerop): Likewise.
2768 (tree_log2): Likewise.
2769 (tree_floor_log2): Likewise.
2770 (tree_ctz): Likewise.
2771 (int_size_in_bytes): Likewise.
2772 (mem_ref_offset): Return an offset_int rather than a double_int.
2773 (build_type_attribute_qual_variant): Use wide_int interfaces.
2774 (type_hash_eq): Likewise
2775 (tree_int_cst_equal): Likewise.
2776 (tree_int_cst_lt): Delete.
2777 (tree_int_cst_compare): Likewise.
2778 (tree_fits_shwi_p): Use wide_int interfaces.
2779 (tree_fits_uhwi_p): Likewise.
2780 (tree_int_cst_sign_bit): Likewise.
2781 (tree_int_cst_sgn): Likewise.
2782 (tree_int_cst_min_precision): Take a signop rather than a bool.
2783 (simple_cst_equal): Use wide_int interfaces.
2784 (compare_tree_int): Likewise.
2785 (iterative_hash_expr): Likewise.
2786 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2788 (get_type_static_bounds): Use wide_int interfaces.
2789 (tree_int_cst_elt_check_failed): New.
2790 (build_common_tree_nodes): Reordered to set prec before filling in
2792 (int_cst_value): Check cst_and_fits_in_hwi.
2793 (widest_int_cst_value): Use wide_int interfaces.
2794 (upper_bound_in_type): Likewise.
2795 (lower_bound_in_type): Likewise.
2796 (num_ending_zeros): Likewise.
2797 (drop_tree_overflow): Likewise.
2798 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2799 (gen_conditions_for_pow_cst_base): Likewise.
2800 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2801 (group_case_labels_stmt): Use wide-int interfaces.
2802 (verify_gimple_assign_binary): Likewise.
2803 (print_loop): Likewise.
2804 * tree-chrec.c (tree_fold_binomial): Likewise.
2805 * tree-core.h (struct tree_base): Add int_length.
2806 (struct tree_int_cst): Change rep of value.
2807 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2808 (dr_may_alias_p): Likewise.
2809 (max_stmt_executions_tree): Likewise.
2810 * tree.def (INTEGER_CST): Update comment.
2811 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2812 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2813 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2814 (dequeue_and_dump): Use wide-int interfaces.
2815 * tree.h: Include wide-int.h.
2816 (NULL_TREE): Moved to earlier loc in file.
2817 (TREE_INT_CST_ELT_CHECK): New.
2818 (tree_int_cst_elt_check_failed): New.
2820 (TREE_INT_CST): Delete.
2821 (TREE_INT_CST_LOW): Use wide-int interfaces.
2822 (TREE_INT_CST_HIGH): Delete.
2823 (TREE_INT_CST_NUNITS): New.
2824 (TREE_INT_CST_EXT_NUNITS): Likewise.
2825 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2826 (TREE_INT_CST_ELT): Likewise.
2827 (INT_CST_LT): Delete.
2828 (tree_int_cst_elt_check): New (two forms).
2829 (type_code_size): Update comment.
2830 (make_int_cst_stat, make_int_cst): New.
2831 (tree_to_double_int): Delete.
2832 (double_int_fits_to_tree_p): Delete.
2833 (force_fit_type_double): Delete.
2834 (build_int_cstu): Replace with out-of-line function.
2835 (build_int_cst_wide): Delete.
2836 (tree_int_cst_lt): Define inline.
2837 (tree_int_cst_le): New.
2838 (tree_int_cst_compare): Define inline.
2839 (tree_int_cst_min_precision): Take a signop rather than a bool.
2840 (wi::int_traits <const_tree>): New.
2841 (wi::int_traits <tree>): New.
2842 (wi::extended_tree): New.
2843 (wi::int_traits <wi::extended_tree>): New.
2844 (wi::to_widest): New.
2845 (wi::to_offset): New.
2846 (wi::fits_to_tree_p): New.
2847 (wi::min_value): New.
2848 (wi::max_value): New.
2849 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2850 (copy_tree_body_r): Likewise.
2851 * tree-object-size.c (compute_object_offset): Likewise.
2852 (addr_object_size): Likewise.
2853 * tree-predcom.c: Include wide-int-print.h.
2854 (struct dref_d): Change type of offset to widest_int.
2855 (dump_dref): Call wide-int printer.
2856 (aff_combination_dr_offset): Use wide-int interfaces.
2857 (determine_offset): Take a widest_int pointer rather than a
2859 (split_data_refs_to_components): Use wide-int interfaces.
2860 (suitable_component_p): Likewise.
2861 (order_drefs): Likewise.
2862 (add_ref_to_chain): Likewise.
2863 (valid_initializer_p): Likewise.
2864 (determine_roots_comp): Likewise.
2865 * tree-pretty-print.c: Include wide-int-print.h.
2866 (dump_generic_node): Use wide-int interfaces.
2867 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2868 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2869 (move_fixed_address_to_symbol): Likewise.
2870 (move_hint_to_base): Likewise.
2871 (move_pointer_to_base): Likewise.
2872 (move_variant_to_index): Likewise.
2873 (most_expensive_mult_to_index): Likewise.
2874 (addr_to_parts): Likewise.
2875 (copy_ref_info): Likewise.
2876 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2877 (indirect_refs_may_alias_p): Likewise.
2878 (stmt_kills_ref_p_1): Likewise.
2879 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2880 * tree-ssa-ccp.c: Update comment at top of file. Include
2882 (struct prop_value_d): Change type of mask to widest_int.
2883 (extend_mask): New function.
2884 (dump_lattice_value): Use wide-int interfaces.
2885 (get_default_value): Likewise.
2886 (set_constant_value): Likewise.
2887 (set_value_varying): Likewise.
2888 (valid_lattice_transition): Likewise.
2889 (set_lattice_value): Likewise.
2890 (value_to_double_int): Delete.
2891 (value_to_wide_int): New.
2892 (get_value_from_alignment): Use wide-int interfaces.
2893 (get_value_for_expr): Likewise.
2894 (do_dbg_cnt): Likewise.
2895 (ccp_finalize): Likewise.
2896 (ccp_lattice_meet): Likewise.
2897 (bit_value_unop_1): Use widest_ints rather than double_ints.
2898 (bit_value_binop_1): Likewise.
2899 (bit_value_unop): Use wide-int interfaces.
2900 (bit_value_binop): Likewise.
2901 (bit_value_assume_aligned): Likewise.
2902 (evaluate_stmt): Likewise.
2903 (ccp_fold_stmt): Likewise.
2904 (visit_cond_stmt): Likewise.
2905 (ccp_visit_stmt): Likewise.
2906 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2907 (constant_pointer_difference): Likewise.
2908 (associate_pointerplus): Likewise.
2909 (combine_conversions): Likewise.
2910 * tree-ssa-loop.h: Include wide-int.h.
2911 (struct tree_niter_desc): Change type of max to widest_int.
2912 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2913 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2914 (remove_redundant_iv_tests): Likewise.
2915 (canonicalize_loop_induction_variables): Likewise.
2916 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2917 (constant_multiple_of): Take a widest_int pointer instead of
2918 a double_int pointer.
2919 (get_computation_aff): Use wide-int interfaces.
2920 (ptr_difference_cost): Likewise.
2921 (difference_cost): Likewise.
2922 (get_loop_invariant_expr_id): Likewise.
2923 (get_computation_cost_at): Likewise.
2924 (iv_elimination_compare_lt): Likewise.
2925 (may_eliminate_iv): Likewise.
2926 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2927 instead of double_int.
2928 (max_loop_iterations): Likewise.
2929 (max_stmt_executions): Likewise.
2930 (estimated_stmt_executions): Likewise.
2931 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2932 (split_to_var_and_offset): Use wide-int interfaces.
2933 (determine_value_range): Likewise.
2934 (bound_difference_of_offsetted_base): Likewise.
2935 (bounds_add): Take a widest_int instead of a double_int.
2936 (number_of_iterations_ne_max): Use wide-int interfaces.
2937 (number_of_iterations_ne): Likewise.
2938 (number_of_iterations_lt_to_ne): Likewise.
2939 (assert_loop_rolls_lt): Likewise.
2940 (number_of_iterations_lt): Likewise.
2941 (number_of_iterations_le): Likewise.
2942 (number_of_iterations_cond): Likewise.
2943 (number_of_iterations_exit): Likewise.
2944 (finite_loop_p): Likewise.
2945 (derive_constant_upper_bound_assign): Likewise.
2946 (derive_constant_upper_bound): Return a widest_int.
2947 (derive_constant_upper_bound_ops): Likewise.
2948 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2949 (record_estimate): Take a widest_int rather than a double_int.
2950 (record_nonwrapping_iv): Use wide-int interfaces.
2951 (double_int_cmp): Delete.
2952 (wide_int_cmp): New.
2953 (bound_index): Take a widest_int rather than a double_int.
2954 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2955 (maybe_lower_iteration_bound): Likewise.
2956 (estimate_numbers_of_iterations_loop): Likewise.
2957 (estimated_loop_iterations): Take a widest_int pointer than than
2958 a double_int pointer.
2959 (estimated_loop_iterations_int): Use wide-int interfaces.
2960 (max_loop_iterations): Take a widest_int pointer than than
2961 a double_int pointer.
2962 (max_loop_iterations_int): Use wide-int interfaces.
2963 (max_stmt_executions): Take a widest_int pointer than than
2964 a double_int pointer.
2965 (estimated_stmt_executions): Likewise.
2966 (n_of_executions_at_most): Use wide-int interfaces.
2967 (scev_probably_wraps_p): Likewise.
2968 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2970 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2972 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2973 double_ints. Adjust for trailing_wide_ints <3> representation.
2974 (set_nonzero_bits): Likewise.
2975 (get_range_info): Return wide_ints rather than double_ints.
2976 Adjust for trailing_wide_ints <3> representation.
2977 (get_nonzero_bits): Likewise.
2978 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2980 * tree-ssanames.h (struct range_info_def): Replace min, max and
2981 nonzero_bits with a trailing_wide_ints <3>.
2982 (set_range_info): Use wide_int_refs rather than double_ints.
2983 (set_nonzero_bits): Likewise.
2984 (get_range_info): Return wide_ints rather than double_ints.
2985 (get_nonzero_bits): Likewise.
2986 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2987 * tree-ssa-pre.c (phi_translate_1): Likewise.
2988 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2989 (acceptable_pow_call): Likewise.
2990 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2992 (vn_reference_fold_indirect): Likewise.
2993 (vn_reference_maybe_forwprop_address): Likewise.
2994 (valueize_refs_1): Likewise.
2995 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2996 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2997 tree_int_cst_lt and tree_int_cst_le.
2998 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3000 (streamer_alloc_tree): Likewise.
3001 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3002 (streamer_write_tree_header): Likewise.
3003 (streamer_write_integer_cst): Likewise.
3004 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3005 (build_constructors): Likewise.
3006 (array_value_type): Likewise.
3007 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3008 (vect_check_gather): Likewise.
3009 * tree-vect-generic.c (build_replicated_const): Likewise.
3010 (expand_vector_divmod): Likewise.
3011 * tree-vect-loop.c (vect_transform_loop): Likewise.
3012 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3013 (vect_do_peeling_for_alignment): Likewise.
3014 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3015 * tree-vrp.c: Include wide-int.h.
3016 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3017 (extract_range_from_assert): Use wide-int interfaces.
3018 (vrp_int_const_binop): Likewise.
3019 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3020 double_int pointers.
3021 (ranges_from_anti_range): Use wide-int interfaces.
3022 (quad_int_cmp): Delete.
3023 (quad_int_pair_sort): Likewise.
3024 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3025 (extract_range_from_unary_expr_1): Likewise.
3026 (adjust_range_with_scev): Likewise.
3027 (masked_increment): Take and return wide_ints rather than double_ints.
3028 (register_edge_assert_for_2): Use wide-int interfaces.
3029 (check_array_ref): Likewise.
3030 (search_for_addr_array): Likewise.
3031 (maybe_set_nonzero_bits): Likewise.
3032 (union_ranges): Pass an integer of the correct type instead of
3033 using integer_one_node.
3034 (intersect_ranges): Likewise.
3035 (simplify_truth_ops_using_ranges): Likewise.
3036 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3037 (range_fits_type_p): Likewise.
3038 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3040 (simplify_conversion_using_ranges): Use wide-int interfaces.
3041 (simplify_float_conversion_using_ranges): Likewise.
3042 (vrp_finalize): Likewise.
3043 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3044 (gimple_stringops_transform): Likewise.
3045 * varasm.c (decode_addr_const): Likewise.
3046 (const_hash_1): Likewise.
3047 (const_rtx_hash_1): Likewise
3048 (output_constant): Likewise.
3049 (array_size_for_constructor): Likewise.
3050 (output_constructor_regular_field): Likewise.
3051 (output_constructor_bitfield): Likewise.
3052 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3053 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3055 * gencheck.c: Define BITS_PER_UNIT.
3058 * wide-int-print.cc: New.
3059 * wide-int-print.h: New.
3061 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3063 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3065 2014-05-06 Richard Biener <rguenther@suse.de>
3067 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3068 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3069 (TODO_verify_all): Adjust.
3070 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3071 TODO_verify_stmts and TODO_verify_rtl_sharing.
3072 * bb-reorder.c: Likewise.
3073 * cfgexpand.c: Likewise.
3074 * cprop.c: Likewise.
3076 * function.c: Likewise.
3077 * fwprop.c: Likewise.
3079 * gimple-ssa-isolate-paths.c: Likewise.
3080 * gimple-ssa-strength-reduction.c: Likewise.
3081 * ipa-split.c: Likewise.
3082 * loop-init.c: Likewise.
3083 * loop-unroll.c: Likewise.
3084 * lower-subreg.c: Likewise.
3085 * modulo-sched.c: Likewise.
3086 * postreload-gcse.c: Likewise.
3087 * predict.c: Likewise.
3088 * recog.c: Likewise.
3089 * sched-rgn.c: Likewise.
3090 * store-motion.c: Likewise.
3091 * tracer.c: Likewise.
3092 * trans-mem.c: Likewise.
3093 * tree-call-cdce.c: Likewise.
3094 * tree-cfg.c: Likewise.
3095 * tree-cfgcleanup.c: Likewise.
3096 * tree-complex.c: Likewise.
3097 * tree-eh.c: Likewise.
3098 * tree-emutls.c: Likewise.
3099 * tree-if-conv.c: Likewise.
3100 * tree-into-ssa.c: Likewise.
3101 * tree-loop-distribution.c: Likewise.
3102 * tree-object-size.c: Likewise.
3103 * tree-parloops.c: Likewise.
3104 * tree-pass.h: Likewise.
3105 * tree-sra.c: Likewise.
3106 * tree-ssa-ccp.c: Likewise.
3107 * tree-ssa-copy.c: Likewise.
3108 * tree-ssa-copyrename.c: Likewise.
3109 * tree-ssa-dce.c: Likewise.
3110 * tree-ssa-dom.c: Likewise.
3111 * tree-ssa-dse.c: Likewise.
3112 * tree-ssa-forwprop.c: Likewise.
3113 * tree-ssa-ifcombine.c: Likewise.
3114 * tree-ssa-loop-ch.c: Likewise.
3115 * tree-ssa-loop-ivcanon.c: Likewise.
3116 * tree-ssa-loop.c: Likewise.
3117 * tree-ssa-math-opts.c: Likewise.
3118 * tree-ssa-phiopt.c: Likewise.
3119 * tree-ssa-phiprop.c: Likewise.
3120 * tree-ssa-pre.c: Likewise.
3121 * tree-ssa-reassoc.c: Likewise.
3122 * tree-ssa-sink.c: Likewise.
3123 * tree-ssa-strlen.c: Likewise.
3124 * tree-ssa-tail-merge.c: Likewise.
3125 * tree-ssa-uncprop.c: Likewise.
3126 * tree-switch-conversion.c: Likewise.
3127 * tree-tailcall.c: Likewise.
3128 * tree-vect-generic.c: Likewise.
3129 * tree-vectorizer.c: Likewise.
3130 * tree-vrp.c: Likewise.
3132 * var-tracking.c: Likewise.
3133 * bt-load.c: Likewise.
3134 * cfgcleanup.c: Likewise.
3135 * combine-stack-adj.c: Likewise.
3136 * combine.c: Likewise.
3137 * compare-elim.c: Likewise.
3138 * config/epiphany/resolve-sw-modes.c: Likewise.
3139 * config/i386/i386.c: Likewise.
3140 * config/mips/mips.c: Likewise.
3141 * config/s390/s390.c: Likewise.
3142 * config/sh/sh_treg_combine.cc: Likewise.
3143 * config/sparc/sparc.c: Likewise.
3146 * final.c: Likewise.
3147 * ifcvt.c: Likewise.
3148 * mode-switching.c: Likewise.
3149 * passes.c: Likewise.
3150 * postreload.c: Likewise.
3152 * reg-stack.c: Likewise.
3153 * regcprop.c: Likewise.
3154 * regrename.c: Likewise.
3157 2014-05-06 Richard Biener <rguenther@suse.de>
3160 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3161 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3163 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3166 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3168 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3169 Tom de Vries <tom@codesourcery.com>
3171 * target.def (call_fusage_contains_non_callee_clobbers): New
3173 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3175 (@node Miscellaneous Register Hooks): New node.
3176 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3177 * doc/tm.texi: Regenerate.
3179 2014-05-05 Marek Polacek <polacek@redhat.com>
3182 * opts.c (common_handle_option): Call error_at instead of warning_at.
3184 2014-05-05 Richard Biener <rguenther@suse.de>
3186 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3187 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3188 under the TODO_verify_il umbrella.
3190 2014-05-05 Richard Biener <rguenther@suse.de>
3192 * passes.c (execute_function_todo): Move TODO_verify_flow under
3193 the TODO_verify_ul umbrella.
3195 2014-05-05 Richard Biener <rguenther@suse.de>
3198 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3199 X & CST away from a CST that is the mask of a mode.
3201 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3203 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3204 int argument to enum machine_mode.
3205 (picochip_class_max_nregs): Ditto.
3206 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3207 (picochip_class_max_nregs): Ditto.
3209 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3211 * target.def: Add new target hook.
3212 * doc/tm.texi: Regenerate.
3213 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3214 * targhooks.c (default_keep_leaf_when_profiled): New function.
3216 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3217 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3219 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3221 PR tree-optimization/60363
3222 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3223 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3224 (update_destination_phis): New parameter.
3225 (create_edge_and_update_destination_phis): Ditto.
3226 (ssa_fix_duplicate_block_edges): Pass new arguments.
3227 (thread_single_edge): Ditto.
3229 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3231 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3232 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3233 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3234 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3235 Use RS6000_BTM_HARD_FLOAT.
3236 (BU_MISC_2): Likewise.
3237 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3238 RS6000_BTM_HARD_FLOAT.
3239 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3241 (rs6000_invalid_builtin): Add hard floating builtin support.
3242 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3243 hard float builtins.
3244 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3246 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3248 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3249 Add missing function* argument.
3251 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3253 * lra-constraints.c (valid_address_p): Move earlier in file.
3254 Add a constraint argument to the address_info version.
3255 (satisfies_memory_constraint_p): New function.
3256 (satisfies_address_constraint_p): Likewise.
3257 (process_alt_operands, curr_insn_transform): Use them.
3258 (process_address): Pass the constraint to valid_address_p when
3259 checking address operands.
3261 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3263 * config/mips/mips.c (mips_isa_rev): New variable.
3264 (mips_set_architecture): Set it.
3265 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3267 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3268 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3269 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3270 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3271 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3272 conditions in terms of mips_isa_rev.
3273 (mips_isa_rev): Declare.
3275 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3277 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3278 (prob_unlikely, prob_likely): Make variables const.
3280 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3282 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3284 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3286 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3288 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3290 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3291 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3292 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3294 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3295 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3297 Replace usage of ROUND_REG with sh_round_reg.
3298 Use CEIL instead of ROUND_ADVANCE.
3300 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3303 * config/sh/sh.c: Include stdlib headers before everything else.
3305 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3307 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3308 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3309 (gimplify_adjust_omp_clauses): Simd region is never
3310 directly nested in combined parallel. Instead, for linear
3311 with copyin/copyout, if in combined for simd loop, make decl
3312 firstprivate/lastprivate on OMP_FOR.
3313 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3314 expand_omp_for_static_chunk): When setting endvar, also set
3315 fd->loop.v to the same value.
3317 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3319 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3321 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3323 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3326 2014-05-02 Marek Polacek <polacek@redhat.com>
3328 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3330 2014-05-02 Kito Cheng <kito@0xlab.org>
3332 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3333 to a C expression marco.
3334 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3335 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3336 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3337 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3338 HONOR_REG_ALLOC_ORDER.
3339 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3341 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3343 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3345 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3347 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3349 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3351 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3352 (convert_scalar_cond_reduction): Likewise.
3353 (predicate_scalar_phi): Add recognition and transformation
3354 of simple conditioanl reduction to be vectorizable.
3356 2014-05-01 Marek Polacek <polacek@redhat.com>
3359 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3361 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3363 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3364 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3365 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3366 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3367 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3368 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3369 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3370 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3372 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3374 * config/arc/arc.opt (mlra): Move comment above option name
3375 to avoid mis-parsing as language options.
3377 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3379 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3380 * config/sol2.h: ... here.
3381 * config/sol2-10.h: Remove.
3383 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3384 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3385 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3386 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3387 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3388 * config/sol2.h: ... here.
3389 (SECTION_NAME_FORMAT): Don't redefine.
3390 (STARTFILE_ARCH32_SPEC): Rename to ...
3391 (STARTFILE_ARCH_SPEC): ... this.
3392 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3393 * config/sparc/sol2.h: ... here.
3394 (SECTION_NAME_FORMAT): Don't undef.
3395 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3396 (SUBTARGET_EXTRA_SPECS): Remove.
3397 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3399 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3400 (MD_STARTFILE_PREFIX): Remove.
3401 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3402 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3403 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3404 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3405 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3406 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3407 * config/i386/sol2.h: ... here.
3408 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3409 * config/i386/sol2-bi.h: Remove.
3410 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3411 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3413 * config/i386/t-sol2-64: Rename to ...
3414 * config/i386/t-sol2: ... this.
3415 * config/sparc/t-sol2-64: Rename to ...
3416 * config/sparc/t-sol2: ... this.
3418 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3419 sol2_tm_file_head, sol2_tm_file_tail.
3420 Include ${cpu_type}/sol2.h before sol2.h.
3422 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3423 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3424 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3425 Reflect i386/t-sol2-64 renaming.
3426 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3427 Reflect sparc/t-sol2-64 renaming.
3429 2014-04-30 Richard Biener <rguenther@suse.de>
3431 * passes.c (execute_function_todo): Move TODO_verify_stmts
3432 and TODO_verify_ssa under the TODO_verify_il umbrella.
3433 * tree-ssa.h (verify_ssa): Adjust prototype.
3434 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3435 we should verify SSA operands.
3436 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3437 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3438 whether we should verify whether not throwing stmts have EH info.
3439 * graphite-scop-detection.c (create_sese_edges): Adjust.
3440 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3441 * tree-eh.c (lower_try_finally_switch): Do not add the
3442 default case label twice.
3444 2014-04-30 Marek Polacek <polacek@redhat.com>
3446 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3447 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3448 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3449 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3451 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3453 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3454 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3455 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3456 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3457 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3458 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3459 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3460 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3462 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3464 * tree-cfg.c (dump_function_to_file): Dump the return type of
3465 functions, in a line to itself before the function body, mimicking
3466 the layout of a C function.
3468 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3470 PR tree-optimization/60971
3471 * tree-tailcall.c (process_assignment): Reject conversions which
3474 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3476 * calls.c (initialize_argument_information): Always treat
3477 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3478 (expand_call): Likewise.
3479 (emit_library_call_calue_1): Likewise.
3480 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3481 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3484 2014-04-29 Nick Clifton <nickc@redhat.com>
3486 * config/msp430/msp430.md (umulsidi): Fix typo.
3487 (mulhisi3): Enable even inside interrupt handlers.
3488 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3489 bigger return address pushed in large mode.
3491 2014-04-29 Nick Clifton <nickc@redhat.com>
3493 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3494 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3496 * config/m32r/m32r.c (init_reg_tables): Likewise.
3497 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3498 enum to hold the modes.
3500 2014-04-29 Richard Biener <rguenther@suse.de>
3502 * dominance.c (free_dominance_info): Add overload with
3504 (dom_info_state): Likewise.
3505 (dom_info_available_p): Likewise.
3506 * basic-block.h (free_dominance_info, dom_info_state,
3507 dom_info_available_p): Declare overloads.
3508 * passes.c (execute_function_todo): Verify that verifiers
3509 don't change dominator info state. Drop dominator info
3510 for IPA pass invocations.
3511 * cgraph.c (release_function_body): Restore asserts that
3512 dominator information is released.
3514 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3516 * doc/invoke.texi: Fix typo.
3517 * tree-vrp.c: Fix typos.
3518 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3520 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3522 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3524 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3526 * config/aarch64/aarch64-builtins.c
3527 (aarch64_types_storestruct_lane_qualifiers): New.
3528 (TYPES_STORESTRUCT_LANE): Likewise.
3529 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3530 (st3_lane): Likewise.
3531 (st4_lane): Likewise.
3532 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3533 (vec_store_lanesci_lane<mode>): Likewise.
3534 (vec_store_lanesxi_lane<mode>): Likewise.
3535 (aarch64_st2_lane<VQ:mode>): Likewise.
3536 (aarch64_st3_lane<VQ:mode>): Likewise.
3537 (aarch64_st4_lane<VQ:mode>): Likewise.
3538 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3539 * config/aarch64/arm_neon.h
3540 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3541 use new macro arguments.
3542 (__ST3_LANE_FUNC): Likewise.
3543 (__ST4_LANE_FUNC): Likewise.
3544 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3545 (V_THREE_ELEM): Likewise.
3546 (V_FOUR_ELEM): Likewise.
3548 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3550 * doc/gimple.texi: Replace the description of the now-defunct
3551 union gimple_statement_d with a diagram showing the
3552 gimple_statement_base class hierarchy and its relationships to
3553 the GSS_ and GIMPLE_ enums.
3555 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3557 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3558 * config/aarch64/aarch64.c
3559 (aarch64_cannot_change_mode_class): Weaken conditions.
3560 (aarch64_modes_tieable_p): New.
3561 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3563 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3565 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3566 (loadsync_<mode>): Change mode.
3567 (load_quadpti, store_quadpti): New.
3568 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3569 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3571 2014-04-28 Martin Jambor <mjambor@suse.cz>
3573 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3574 same alias type as the original statement.
3575 (subreplacement_assignment_data): New type.
3576 (handle_unscalarized_data_in_subtree): New type of parameter,
3577 generate new memory accesses with same alias type as the original
3579 (load_assign_lhs_subreplacements): Likewise.
3580 (sra_modify_constructor_assign): Generate new memory accesses with
3581 same alias type as the original statement.
3583 2014-04-28 Richard Biener <rguenther@suse.de>
3585 * tree-pass.h (TODO_verify_il): Define.
3586 (TODO_verify_all): Complete properly.
3587 * passes.c (execute_function_todo): Move existing loop-closed
3588 SSA verification under TODO_verify_il.
3589 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3590 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3591 Fix tree sharing issue.
3593 2014-04-28 Richard Biener <rguenther@suse.de>
3596 * builtins.def (DEF_C11_BUILTIN): Add.
3597 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3598 * coretypes.h (enum function_class): Add function_c11_misc.
3599 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3600 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3601 (call_may_clobber_ref_p_1): Likewise.
3602 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3603 (mark_all_reaching_defs_necessary_1): Likewise.
3604 (propagate_necessity): Likewise.
3605 (eliminate_unnecessary_stmts): Likewise.
3606 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3608 2014-04-28 Richard Biener <rguenther@suse.de>
3610 * tree-vrp.c (vrp_var_may_overflow): Remove.
3611 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3612 with overflow immediately bump to one before that value and
3613 let iteration figure out overflow status.
3615 2014-04-28 Richard Biener <rguenther@suse.de>
3617 * configure.ac: Do valgrind header checks unconditionally.
3618 Add --enable-valgrind-annotations.
3619 * system.h: Guard valgrind header inclusion with
3620 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3621 * alloc-pool.c (pool_alloc, pool_free): Use
3622 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3623 to guard possibly dead code.
3624 * config.in: Regenerated.
3625 * configure: Likewise.
3627 2014-04-28 Jeff Law <law@redhat.com>
3629 PR tree-optimization/60902
3630 * tree-ssa-threadedge.c
3631 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3632 over real defs when invalidating outputs from statements that do not
3633 produce useful outputs for threading.
3635 2014-04-28 Richard Biener <rguenther@suse.de>
3637 PR tree-optimization/60979
3638 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3639 SCOPs that end in a block with a successor with abnormal
3642 2014-04-28 Richard Biener <rguenther@suse.de>
3644 * tree-pass.h (execute_pass_list): Adjust prototype.
3645 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3646 (do_per_function): Change callback signature, push all actual
3647 work to the callbals.
3648 (do_per_function_toporder): Likewise.
3649 (execute_function_dump): Adjust.
3650 (execute_function_todo): Likewise.
3651 (clear_last_verified): Likewise.
3652 (verify_curr_properties): Likewise.
3653 (update_properties_after_pass): Likewise.
3654 (execute_pass_list_1): Split out from ...
3655 (execute_pass_list): ... here. Adjust.
3656 (execute_ipa_pass_list): Likewise.
3657 * cgraphunit.c (cgraph_add_new_function): Adjust.
3658 (analyze_function): Likewise.
3659 (expand_function): Likewise.
3660 * cgraph.c (release_function_body): Free dominance info
3661 here instead of asserting it was magically freed elsewhere.
3663 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3665 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3666 * configure: Regenerate.
3667 * config/sparc/sparc.opt (muser-mode): New option.
3668 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3670 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3671 * doc/invoke.texi (SPARC options): Document -muser-mode.
3673 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3675 * cselib.c (find_slot_memmode): Delete.
3676 (cselib_hasher): Change compare_type to a struct.
3677 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3679 (preserve_constants_and_equivs): Adjust for new compare_type.
3680 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3681 (wrap_constant): Delete.
3682 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3684 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3686 * doc/install.texi (Building with profile feedback): Remove
3689 2014-04-26 Tom de Vries <tom@codesourcery.com>
3691 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3694 2014-04-25 Cary Coutant <ccoutant@google.com>
3697 * dwarf2out.c (should_move_die_to_comdat): A type definition
3698 can contain a subprogram definition, but don't move it to a
3700 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3701 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3703 (clone_tree_hash): Rename to...
3704 (clone_tree_partial): ...this; change callers. Copy
3705 DW_TAG_subprogram DIEs as declarations.
3706 (copy_decls_walk): Don't copy children of a declaration into a
3709 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3712 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3715 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3717 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3718 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3720 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3723 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3725 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3727 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3729 PR tree-optimization/60930
3730 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3731 creating a multiply candidate by folding two constant
3732 multiplicands when the result overflows.
3734 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3736 PR tree-optimization/60960
3737 * tree-vect-generic.c (expand_vector_operation): Only call
3738 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3740 2014-04-25 Tom de Vries <tom@codesourcery.com>
3742 * expr.c (clobber_reg_mode): New function.
3743 * expr.h (clobber_reg): New function.
3745 2014-04-25 Tom de Vries <tom@codesourcery.com>
3747 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3750 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3751 Tom de Vries <tom@codesourcery.com>
3753 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3755 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3756 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3757 new argument to find_all_hard_reg_sets call.
3759 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3761 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3762 Use HOST_WIDE_INT_C for mask literal.
3763 (aarch_rev16_shleft_mask_imm_p): Likewise.
3765 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3768 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3770 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3772 PR preprocessor/56540
3773 * config/i386/i386-c.c (ix86_target_macros): Define
3774 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3776 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3778 * configure.ac (tga_func): Remove.
3779 (LIB_TLS_SPEC): Remove.
3780 * configure: Regenerate.
3781 * config.in: Regenerate.
3782 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3784 2014-04-25 Richard Biener <rguenther@suse.de>
3787 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3788 call stmt use/clobber sets during stmt walk instead of
3789 walking the possibly incomplete set of caller edges.
3791 2014-04-25 Richard Biener <rguenther@suse.de>
3794 * passes.c (apply_ipa_transforms): Inline into only caller ...
3795 (execute_one_pass): ... here. Properly bring in function
3796 bodies for nodes we want to apply IPA transforms to.
3798 2014-04-24 Cong Hou <congh@google.com>
3800 PR tree-optimization/60896
3801 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3802 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3803 (vect_mark_pattern_stmts): Set the def type of all statements in
3804 PATTERN_DEF_SEQ as vect_internal_def.
3806 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3808 * doc/extend.texi (PowerPC Built-in Functions): Document new
3809 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3810 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3812 * config/rs6000/predicates.md (const_0_to_3_operand): New
3813 predicate to match 0..3 integer constants.
3815 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3816 to support adding miscellaneous builtin functions.
3817 (BU_DFP_MISC_2): Likewise.
3818 (BU_P7_MISC_1): Likewise.
3819 (BU_P7_MISC_2): Likewise.
3820 (BU_P8V_MISC_3): Likewise.
3821 (BU_MISC_1): Likewise.
3822 (BU_MISC_2): Likewise.
3823 (DIVWE): Add extended divide builtin functions.
3826 (DIVWEUO): Likewise.
3830 (DIVDEUO): Likewise.
3831 (DXEX): Add decimal floating-point builtin functions.
3834 (DDEDPDQ): Likewise.
3836 (DENBCDQ): Likewise.
3843 (CDTBCD): Add new BCD builtin functions.
3847 (BCDADD_LT): Likewise.
3848 (BCDADD_EQ): Likewise.
3849 (BCDADD_GT): Likewise.
3850 (BCDADD_OV): Likewise.
3852 (BCDSUB_LT): Likewise.
3853 (BCDSUB_EQ): Likewise.
3854 (BCDSUB_GT): Likewise.
3855 (BCDSUB_OV): Likewise.
3856 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3857 (UNPACK_TD): Likewise.
3858 (PACK_TF): Likewise.
3859 (UNPACK_TF): Likewise.
3860 (UNPACK_TF_0): Likewise.
3861 (UNPACK_TF_1): Likewise.
3862 (PACK_V1TI): Likewise.
3863 (UNPACK_V1TI): Likewise.
3865 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3866 support for decimal floating point builtin functions.
3867 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3868 functions that take constant arguments.
3869 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3870 (rs6000_init_builtins): Setup long double, _Decimal64, and
3871 _Decimal128 types for new builtin functions.
3872 (builtin_function_type): Set the unsigned flags appropriately for
3873 the new builtin functions.
3874 (rs6000_opt_masks): Add support for decimal floating point builtin
3877 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3878 floating point builtin functions.
3879 (RS6000_BTM_COMMON): Likewise.
3880 (RS6000_BTI_long_double): Likewise.
3881 (RS6000_BTI_dfloat64): Likewise.
3882 (RS6000_BTI_dfloat128): Likewise.
3883 (long_double_type_internal_node): Likewise.
3884 (dfloat64_type_internal_node): Likewise.
3885 (dfloat128_type_internal_node): Likewise.
3887 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3888 2.07 bcd arithmetic instructions.
3889 (UNSPEC_BCDSUB): Likewise.
3890 (UNSPEC_BCD_OVERFLOW): Likewise.
3891 (UNSPEC_BCD_ADD_SUB): Likewise.
3892 (bcd_add_sub): Likewise.
3893 (BCD_TEST): Likewise.
3894 (bcd<bcd_add_sub>): Likewise.
3895 (bcd<bcd_add_sub>_test): Likewise.
3896 (bcd<bcd_add_sub>_test2): Likewise.
3897 (bcd<bcd_add_sub>_<code>): Likewise.
3898 (peephole2 for combined bcd ops): Likewise.
3900 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3901 decimal floating point builtin functions.
3902 (UNSPEC_DENBCD): Likewise.
3903 (UNSPEC_DXEX): Likewise.
3904 (UNSPEC_DIEX): Likewise.
3905 (UNSPEC_DSCLI): Likewise.
3906 (UNSPEC_DSCRI): Likewise.
3907 (D64_D128): Likewise.
3908 (dfp_suffix): Likewise.
3909 (dfp_ddedpd_<mode>): Likewise.
3910 (dfp_denbcd_<mode>): Likewise.
3911 (dfp_dxex_<mode>): Likewise.
3912 (dfp_diex_<mode>): Likewise.
3913 (dfp_dscli_<mode>): Likewise.
3914 (dfp_dscri_<mode>): Likewise.
3916 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3918 (UNSPEC_CDTBCD): Likewise.
3919 (UNSPEC_CBCDTD): Likewise.
3920 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3921 (UNSPEC_DIVEO): Likewise.
3922 (UNSPEC_DIVEU): Likewise.
3923 (UNSPEC_DIVEUO): Likewise.
3924 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3925 pack/unpack 128-bit types.
3926 (UNSPEC_PACK_128BIT): Likewise.
3927 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3928 (udiv<mode>3): Use idiv_ldiv mode attribute.
3929 (div<mode>3): Likewise.
3930 (addg6s): Add new BCD builtin functions.
3933 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3934 (div_extend): Likewise.
3935 (div<div_extend>_<mode>"): Likewise.
3936 (FP128_64): Add support for new builtin functions to pack/unpack
3938 (unpack<mode>): Likewise.
3939 (unpacktf_0): Likewise.
3940 (unpacktf_1): Likewise.
3941 (unpack<mode>_dm): Likewise.
3942 (unpack<mode>_nodm): Likewise.
3943 (pack<mode>): Likewise.
3944 (unpackv1ti): Likewise.
3945 (packv1ti): Likewise.
3947 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3949 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3952 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3954 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3955 * gimplify.c (omp_is_private): Change last argument's type to int.
3956 Only diagnose lastprivate if the simd argument is 1, only diagnose
3957 linear if the simd argument is 2.
3958 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3959 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3960 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3961 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3962 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3963 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3964 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3965 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3966 * tree-nested.c (convert_nonlocal_omp_clauses,
3967 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3969 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3972 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3975 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3977 * flag-types.h (enum ivar_visibility): Add.
3979 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3981 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3982 function * argument.
3984 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3986 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3988 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3989 Tom de Vries <tom@codesourcery.com>
3991 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3992 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3994 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3995 * emit-rtl.c (try_split): Same.
3997 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3998 Tom de Vries <tom@codesourcery.com>
4000 * common.opt (fuse-caller-save): New option.
4002 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4004 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4005 elements for big-endian.
4007 2014-04-24 Richard Biener <rguenther@suse.de>
4009 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4010 during TER and instead use the sepops interface for expanding
4011 non-GIMPLE_SINGLE_RHS.
4013 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4015 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4016 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4018 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4020 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4021 assembler 64-bit option.
4022 * configure: Regenerate.
4024 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4026 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4027 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4028 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4029 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4030 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4032 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4034 * config/aarch64/aarch64-builtins.c
4035 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4036 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4037 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4038 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4040 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4041 (Vrevsuff): New mode attribute.
4043 2014-04-24 Terry Guo <terry.guo@arm.com>
4045 * config/arm/arm.h (machine_function): Define variable
4046 after_arm_reorg here.
4047 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4048 (arm_split_constant): Update the way to access variable
4051 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4053 2014-04-23 Tom de Vries <tom@codesourcery.com>
4055 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4057 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4059 * is-a.h: Update comments to reflect the following changes to the
4060 "pointerness" of the API, making the template parameter match the
4061 return type, allowing use of is-a.h with typedefs of pointers.
4062 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4063 that the return type matches the parameter to the is_a_helper.
4065 (dyn_cast): Likewise.
4067 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4068 pointer from the is-a.h API.
4070 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4071 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4073 (is_a_helper <varpool_node>::test): Likewise, convert to...
4074 (is_a_helper <varpool_node *>::test): ...this.
4076 (varpool_first_variable): Update for removal of implicit pointer
4077 from the is-a.h API.
4078 (varpool_next_variable): Likewise.
4079 (varpool_first_static_initializer): Likewise.
4080 (varpool_next_static_initializer): Likewise.
4081 (varpool_first_defined_variable): Likewise.
4082 (varpool_next_defined_variable): Likewise.
4083 (cgraph_first_defined_function): Likewise.
4084 (cgraph_next_defined_function): Likewise.
4085 (cgraph_first_function): Likewise.
4086 (cgraph_next_function): Likewise.
4087 (cgraph_first_function_with_gimple_body): Likewise.
4088 (cgraph_next_function_with_gimple_body): Likewise.
4089 (cgraph_alias_target): Likewise.
4090 (varpool_alias_target): Likewise.
4091 (cgraph_function_or_thunk_node): Likewise.
4092 (varpool_variable_node): Likewise.
4093 (symtab_real_symbol_p): Likewise.
4094 * cgraphunit.c (referred_to_p): Likewise.
4095 (analyze_functions): Likewise.
4096 (handle_alias_pairs): Likewise.
4097 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4098 * gimple-ssa.h (gimple_vuse_op): Likewise.
4099 (gimple_vdef_op): Likewise.
4100 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4101 * gimple.c (gimple_build_asm_1): Likewise.
4102 (gimple_build_try): Likewise.
4103 (gimple_build_resx): Likewise.
4104 (gimple_build_eh_dispatch): Likewise.
4105 (gimple_build_omp_for): Likewise.
4106 (gimple_omp_for_set_clauses): Likewise.
4108 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4109 (is_a_helper <gimple_statement_asm *>::test): ...this.
4110 (is_a_helper <gimple_statement_bind>::test): Convert to...
4111 (is_a_helper <gimple_statement_bind *>::test): ...this.
4112 (is_a_helper <gimple_statement_call>::test): Convert to...
4113 (is_a_helper <gimple_statement_call *>::test): ...this.
4114 (is_a_helper <gimple_statement_catch>::test): Convert to...
4115 (is_a_helper <gimple_statement_catch *>::test): ...this.
4116 (is_a_helper <gimple_statement_resx>::test): Convert to...
4117 (is_a_helper <gimple_statement_resx *>::test): ...this.
4118 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4119 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4120 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4121 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4122 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4123 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4124 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4125 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4126 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4127 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4128 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4129 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4130 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4131 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4132 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4133 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4134 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4135 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4136 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4137 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4138 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4139 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4140 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4141 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4142 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4143 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4144 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4145 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4146 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4147 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4148 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4149 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4150 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4151 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4152 (is_a_helper <gimple_statement_phi>::test): Convert to...
4153 (is_a_helper <gimple_statement_phi *>::test): ...this.
4154 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4155 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4156 (is_a_helper <gimple_statement_try>::test): Convert to...
4157 (is_a_helper <gimple_statement_try *>::test): ...this.
4158 (is_a_helper <gimple_statement_wce>::test): Convert to...
4159 (is_a_helper <gimple_statement_wce *>::test): ...this.
4160 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4161 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4162 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4163 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4164 (is_a_helper <const gimple_statement_call>::test): Convert to...
4165 (is_a_helper <const gimple_statement_call *>::test): ...this.
4166 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4167 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4168 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4169 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4170 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4171 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4172 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4173 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4174 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4176 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4178 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4180 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4182 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4183 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4184 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4186 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4187 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4189 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4190 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4191 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4192 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4193 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4194 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4196 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4197 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4198 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4199 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4201 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4202 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4203 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4204 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4205 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4206 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4207 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4208 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4209 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4210 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4211 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4212 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4213 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4214 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4215 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4216 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4218 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4220 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4221 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4223 (gimple_use_ops): Update for removal of implicit pointer from the
4225 (gimple_set_use_ops): Likewise.
4226 (gimple_vuse): Likewise.
4227 (gimple_vdef): Likewise.
4228 (gimple_vuse_ptr): Likewise.
4229 (gimple_vdef_ptr): Likewise.
4230 (gimple_set_vuse): Likewise.
4231 (gimple_set_vdef): Likewise.
4232 (gimple_omp_return_set_lhs): Likewise.
4233 (gimple_omp_return_lhs): Likewise.
4234 (gimple_omp_return_lhs_ptr): Likewise.
4235 (gimple_call_fntype): Likewise.
4236 (gimple_call_set_fntype): Likewise.
4237 (gimple_call_set_internal_fn): Likewise.
4238 (gimple_call_use_set): Likewise.
4239 (gimple_call_clobber_set): Likewise.
4240 (gimple_bind_vars): Likewise.
4241 (gimple_bind_set_vars): Likewise.
4242 (gimple_bind_body_ptr): Likewise.
4243 (gimple_bind_set_body): Likewise.
4244 (gimple_bind_add_stmt): Likewise.
4245 (gimple_bind_block): Likewise.
4246 (gimple_bind_set_block): Likewise.
4247 (gimple_asm_ninputs): Likewise.
4248 (gimple_asm_noutputs): Likewise.
4249 (gimple_asm_nclobbers): Likewise.
4250 (gimple_asm_nlabels): Likewise.
4251 (gimple_asm_input_op): Likewise.
4252 (gimple_asm_input_op_ptr): Likewise.
4253 (gimple_asm_output_op): Likewise.
4254 (gimple_asm_output_op_ptr): Likewise.
4255 (gimple_asm_set_output_op): Likewise.
4256 (gimple_asm_clobber_op): Likewise.
4257 (gimple_asm_set_clobber_op): Likewise.
4258 (gimple_asm_label_op): Likewise.
4259 (gimple_asm_set_label_op): Likewise.
4260 (gimple_asm_string): Likewise.
4261 (gimple_catch_types): Likewise.
4262 (gimple_catch_types_ptr): Likewise.
4263 (gimple_catch_handler_ptr): Likewise.
4264 (gimple_catch_set_types): Likewise.
4265 (gimple_catch_set_handler): Likewise.
4266 (gimple_eh_filter_types): Likewise.
4267 (gimple_eh_filter_types_ptr): Likewise.
4268 (gimple_eh_filter_failure_ptr): Likewise.
4269 (gimple_eh_filter_set_types): Likewise.
4270 (gimple_eh_filter_set_failure): Likewise.
4271 (gimple_eh_must_not_throw_fndecl): Likewise.
4272 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4273 (gimple_eh_else_n_body_ptr): Likewise.
4274 (gimple_eh_else_e_body_ptr): Likewise.
4275 (gimple_eh_else_set_n_body): Likewise.
4276 (gimple_eh_else_set_e_body): Likewise.
4277 (gimple_try_eval_ptr): Likewise.
4278 (gimple_try_cleanup_ptr): Likewise.
4279 (gimple_try_set_eval): Likewise.
4280 (gimple_try_set_cleanup): Likewise.
4281 (gimple_wce_cleanup_ptr): Likewise.
4282 (gimple_wce_set_cleanup): Likewise.
4283 (gimple_phi_capacity): Likewise.
4284 (gimple_phi_num_args): Likewise.
4285 (gimple_phi_result): Likewise.
4286 (gimple_phi_result_ptr): Likewise.
4287 (gimple_phi_set_result): Likewise.
4288 (gimple_phi_arg): Likewise.
4289 (gimple_phi_set_arg): Likewise.
4290 (gimple_resx_region): Likewise.
4291 (gimple_resx_set_region): Likewise.
4292 (gimple_eh_dispatch_region): Likewise.
4293 (gimple_eh_dispatch_set_region): Likewise.
4294 (gimple_omp_critical_name): Likewise.
4295 (gimple_omp_critical_name_ptr): Likewise.
4296 (gimple_omp_critical_set_name): Likewise.
4297 (gimple_omp_for_clauses): Likewise.
4298 (gimple_omp_for_clauses_ptr): Likewise.
4299 (gimple_omp_for_set_clauses): Likewise.
4300 (gimple_omp_for_collapse): Likewise.
4301 (gimple_omp_for_index): Likewise.
4302 (gimple_omp_for_index_ptr): Likewise.
4303 (gimple_omp_for_set_index): Likewise.
4304 (gimple_omp_for_initial): Likewise.
4305 (gimple_omp_for_initial_ptr): Likewise.
4306 (gimple_omp_for_set_initial): Likewise.
4307 (gimple_omp_for_final): Likewise.
4308 (gimple_omp_for_final_ptr): Likewise.
4309 (gimple_omp_for_set_final): Likewise.
4310 (gimple_omp_for_incr): Likewise.
4311 (gimple_omp_for_incr_ptr): Likewise.
4312 (gimple_omp_for_set_incr): Likewise.
4313 (gimple_omp_for_pre_body_ptr): Likewise.
4314 (gimple_omp_for_set_pre_body): Likewise.
4315 (gimple_omp_parallel_clauses): Likewise.
4316 (gimple_omp_parallel_clauses_ptr): Likewise.
4317 (gimple_omp_parallel_set_clauses): Likewise.
4318 (gimple_omp_parallel_child_fn): Likewise.
4319 (gimple_omp_parallel_child_fn_ptr): Likewise.
4320 (gimple_omp_parallel_set_child_fn): Likewise.
4321 (gimple_omp_parallel_data_arg): Likewise.
4322 (gimple_omp_parallel_data_arg_ptr): Likewise.
4323 (gimple_omp_parallel_set_data_arg): Likewise.
4324 (gimple_omp_task_clauses): Likewise.
4325 (gimple_omp_task_clauses_ptr): Likewise.
4326 (gimple_omp_task_set_clauses): Likewise.
4327 (gimple_omp_task_child_fn): Likewise.
4328 (gimple_omp_task_child_fn_ptr): Likewise.
4329 (gimple_omp_task_set_child_fn): Likewise.
4330 (gimple_omp_task_data_arg): Likewise.
4331 (gimple_omp_task_data_arg_ptr): Likewise.
4332 (gimple_omp_task_set_data_arg): Likewise.
4333 (gimple_omp_taskreg_clauses): Likewise.
4334 (gimple_omp_taskreg_clauses_ptr): Likewise.
4335 (gimple_omp_taskreg_set_clauses): Likewise.
4336 (gimple_omp_taskreg_child_fn): Likewise.
4337 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4338 (gimple_omp_taskreg_set_child_fn): Likewise.
4339 (gimple_omp_taskreg_data_arg): Likewise.
4340 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4341 (gimple_omp_taskreg_set_data_arg): Likewise.
4342 (gimple_omp_task_copy_fn): Likewise.
4343 (gimple_omp_task_copy_fn_ptr): Likewise.
4344 (gimple_omp_task_set_copy_fn): Likewise.
4345 (gimple_omp_task_arg_size): Likewise.
4346 (gimple_omp_task_arg_size_ptr): Likewise.
4347 (gimple_omp_task_set_arg_size): Likewise.
4348 (gimple_omp_task_arg_align): Likewise.
4349 (gimple_omp_task_arg_align_ptr): Likewise.
4350 (gimple_omp_task_set_arg_align): Likewise.
4351 (gimple_omp_single_clauses): Likewise.
4352 (gimple_omp_single_clauses_ptr): Likewise.
4353 (gimple_omp_single_set_clauses): Likewise.
4354 (gimple_omp_target_clauses): Likewise.
4355 (gimple_omp_target_clauses_ptr): Likewise.
4356 (gimple_omp_target_set_clauses): Likewise.
4357 (gimple_omp_target_child_fn): Likewise.
4358 (gimple_omp_target_child_fn_ptr): Likewise.
4359 (gimple_omp_target_set_child_fn): Likewise.
4360 (gimple_omp_target_data_arg): Likewise.
4361 (gimple_omp_target_data_arg_ptr): Likewise.
4362 (gimple_omp_target_set_data_arg): Likewise.
4363 (gimple_omp_teams_clauses): Likewise.
4364 (gimple_omp_teams_clauses_ptr): Likewise.
4365 (gimple_omp_teams_set_clauses): Likewise.
4366 (gimple_omp_sections_clauses): Likewise.
4367 (gimple_omp_sections_clauses_ptr): Likewise.
4368 (gimple_omp_sections_set_clauses): Likewise.
4369 (gimple_omp_sections_control): Likewise.
4370 (gimple_omp_sections_control_ptr): Likewise.
4371 (gimple_omp_sections_set_control): Likewise.
4372 (gimple_omp_for_set_cond): Likewise.
4373 (gimple_omp_for_cond): Likewise.
4374 (gimple_omp_atomic_store_set_val): Likewise.
4375 (gimple_omp_atomic_store_val): Likewise.
4376 (gimple_omp_atomic_store_val_ptr): Likewise.
4377 (gimple_omp_atomic_load_set_lhs): Likewise.
4378 (gimple_omp_atomic_load_lhs): Likewise.
4379 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4380 (gimple_omp_atomic_load_set_rhs): Likewise.
4381 (gimple_omp_atomic_load_rhs): Likewise.
4382 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4383 (gimple_omp_continue_control_def): Likewise.
4384 (gimple_omp_continue_control_def_ptr): Likewise.
4385 (gimple_omp_continue_set_control_def): Likewise.
4386 (gimple_omp_continue_control_use): Likewise.
4387 (gimple_omp_continue_control_use_ptr): Likewise.
4388 (gimple_omp_continue_set_control_use): Likewise.
4389 (gimple_transaction_body_ptr): Likewise.
4390 (gimple_transaction_label): Likewise.
4391 (gimple_transaction_label_ptr): Likewise.
4392 (gimple_transaction_set_body): Likewise.
4393 (gimple_transaction_set_label): Likewise.
4395 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4396 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4397 * ipa-ref.c (ipa_record_reference): Likewise.
4398 * ipa-reference.c (analyze_function): Likewise.
4399 (ipa_reference_write_optimization_summary): Likewise.
4400 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4401 (address_taken_from_non_vtable_p): Likewise.
4402 (comdat_can_be_unshared_p_1): Likewise.
4403 * lto-cgraph.c (lto_output_ref): Likewise.
4404 (add_references): Likewise.
4405 (compute_ltrans_boundary): Likewise.
4406 (output_symtab): Likewise.
4407 (input_ref): Likewise.
4408 (input_cgraph_1): Likewise.
4409 (output_cgraph_opt_summary): Likewise.
4410 * lto-streamer-out.c (lto_output): Likewise.
4411 (output_symbol_p): Likewise.
4412 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4413 (lsei_start_function_in_partition): Likewise.
4414 (lsei_next_variable_in_partition): Likewise.
4415 (lsei_start_variable_in_partition): Likewise.
4416 * symtab.c (insert_to_assembler_name_hash): Likewise.
4417 (unlink_from_assembler_name_hash): Likewise.
4418 (symtab_unregister_node): Likewise.
4419 (symtab_remove_node): Likewise.
4420 (dump_symtab_node): Likewise.
4421 (verify_symtab_base): Likewise.
4422 (verify_symtab_node): Likewise.
4423 (symtab_make_decl_local): Likewise.
4424 (symtab_alias_ultimate_target): Likewise.
4425 (symtab_resolve_alias): Likewise.
4426 (symtab_get_symbol_partitioning_class): Likewise.
4427 * tree-phinodes.c (allocate_phi_node): Likewise.
4428 (reserve_phi_args_for_new_edge): Likewise.
4429 (remove_phi_args): Likewise.
4430 * varpool.c (varpool_node_for_asm): Likewise.
4431 (varpool_remove_unreferenced_decls): Likewise.
4433 2014-04-23 Jeff Law <law@redhat.com>
4435 PR tree-optimization/60902
4436 * tree-ssa-threadedge.c
4437 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4438 invalidate outputs from statements that do not produce useful
4439 outputs for threading.
4441 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4443 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4444 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4445 machine descriptions for Stack Smashing Protector.
4447 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4449 * aarch64.md (<optab>_rol<mode>3): New pattern.
4450 (<optab>_rolsi3_uxtw): Likewise.
4451 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4453 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4455 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4456 (arm_cortex_a12_tune): Likewise.
4458 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4460 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4462 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4464 * config/arm/arm.md (arm_rev16si2): New pattern.
4465 (arm_rev16si2_alt): Likewise.
4466 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4468 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4470 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4471 (rev16<mode>2_alt): Likewise.
4472 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4473 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4474 (aarch_rev16_shleft_mask_imm_p): Likewise.
4475 (aarch_rev16_p_1): Likewise.
4476 (aarch_rev16_p): Likewise.
4477 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4478 (aarch_rev16_shright_mask_imm_p): Likewise.
4479 (aarch_rev16_shleft_mask_imm_p): Likewise.
4481 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4483 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4484 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4486 (cortex_a53_extra_costs): Likewise.
4487 (cortex_a57_extra_costs): Likewise.
4488 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4489 (cortexa7_extra_costs): Likewise.
4490 (cortexa8_extra_costs): Likewise.
4491 (cortexa12_extra_costs): Likewise.
4492 (cortexa15_extra_costs): Likewise.
4493 (v7m_extra_costs): Likewise.
4494 (arm_new_rtx_costs): Handle BSWAP.
4496 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4498 * config/arm/arm.c (cortexa8_extra_costs): New table.
4499 (arm_cortex_a8_tune): New tuning struct.
4500 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4502 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4504 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4506 2014-04-23 Richard Biener <rguenther@suse.de>
4508 * Makefile.in (OBJS): Remove loop-unswitch.o.
4509 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4510 * passes.def (pass_rtl_unswitch): Likewise.
4511 * loop-init.c (gate_rtl_unswitch): Likewise.
4512 (rtl_unswitch): Likewise.
4513 (pass_data_rtl_unswitch): Likewise.
4514 (pass_rtl_unswitch): Likewise.
4515 (make_pass_rtl_unswitch): Likewise.
4516 * rtl.h (reversed_condition): Likewise.
4517 (compare_and_jump_seq): Likewise.
4518 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4520 * loop-unroll.c (compare_and_jump_seq): Likewise.
4522 2014-04-23 Richard Biener <rguenther@suse.de>
4524 PR tree-optimization/60903
4525 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4526 commented code block.
4527 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4528 loop flags to newly created BBs and edges.
4530 2014-04-23 Nick Clifton <nickc@redhat.com>
4532 * config/msp430/msp430.c (msp430_handle_option): Move function
4534 (msp430_option_override): Simplify mcu and mcpu option handling.
4535 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4536 support for -mhwmult command line option.
4537 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4538 -mhwmult command line option.
4539 (msp430_hwmult_enabled): Delete.
4540 (msp43o_output_labelref): Add support for -mhwmult command line option.
4541 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4542 (umulsidi3): Likewise.
4543 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4544 (mcpu, mlarge, msmall): Likewise.
4545 (mhwmult): New option.
4546 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4548 (msp430_is_f5_mcu): Remove prototype.
4549 (msp430_use_f5_series_hwmult): Add prototype.
4550 * config/msp430/msp430-opts.h: New file.
4551 * common/config/msp430: New directory.
4552 * common/config/msp430/msp430-common.c: New file.
4553 * config.gcc (msp430): Remove target_has_targetm_common.
4554 * doc/invoke.texi: Document -mhwmult command line option.
4556 2014-04-23 Nick Clifton <nickc@redhat.com>
4558 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4559 default-manifest.o if it can be found in the search path.
4560 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4562 2014-04-23 Terry Guo <terry.guo@arm.com>
4564 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4566 2014-04-23 Richard Biener <rguenther@suse.de>
4569 * tree-inline.c (declare_return_variable): Use mark_addressable.
4571 2014-04-23 Richard Biener <rguenther@suse.de>
4574 * loop-init.c (loop_optimizer_init): Make sure to apply
4575 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4577 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4580 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4582 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4583 if flag_sanitize_undefined_trap_on_error.
4584 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4585 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4586 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4587 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4588 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4589 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4590 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4591 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4592 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4593 * ubsan.c (ubsan_instrument_unreachable): Return
4594 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4595 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4596 if flag_sanitize_undefined_trap_on_error and
4597 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4598 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4599 instrument_bool_enum_load): Emit __builtin_trap () if
4600 flag_sanitize_undefined_trap_on_error and
4601 __builtin_handle_*_abort () if !flag_sanitize_recover.
4602 * doc/invoke.texi (-fsanitize-recover,
4603 -fsanitize-undefined-trap-on-error): Document.
4605 2014-04-22 Christian Bruel <christian.bruel@st.com>
4607 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4608 Force immediates to SImode.
4610 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4612 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4614 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4615 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4616 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4617 (nios2_fpu_insn): Add entry for round.
4618 (N2FPU_NO_ERRNO_P): Define.
4619 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4621 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4623 2014-04-22 Richard Henderson <rth@redhat.com>
4625 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4626 (add<GPI>3_compare0): Remove leading * from name.
4627 (add<GPI>3_carryin): Likewise.
4628 (sub<GPI>3_compare0): Likewise.
4629 (sub<GPI>3_carryin): Likewise.
4630 (<su_optab>mulditi3): New expander.
4631 (multi3): New expander.
4632 (madd<GPI>): Remove leading * from name.
4634 2014-04-22 Martin Jambor <mjambor@suse.cz>
4636 * cgraphclones.c (cgraph_function_versioning): Copy
4637 ipa_transforms_to_apply instead of asserting it is empty.
4639 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4642 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4643 on count_exp to get mode.
4645 2014-04-22 Andrew Pinski <apinski@cavium.com>
4647 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4648 Handle TLS for ILP32.
4649 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4650 (tlsie_small_<mode>): this and handle PTR.
4651 (tlsie_small_sidi): New pattern.
4652 (tlsle_small): Change to an expand to handle ILP32.
4653 (tlsle_small_<mode>): New pattern.
4654 (tlsdesc_small): Rename to ...
4655 (tlsdesc_small_<mode>): this and handle PTR.
4657 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4659 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4661 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4663 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4664 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4665 (aarch64_types_signed_poly_qualifiers): Likewise.
4666 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4667 (aarch64_types_poly_signed_qualifiers): Likewise.
4668 (TYPES_REINTERP_SS): Type macro added.
4669 (TYPES_REINTERP_SU): Likewise.
4670 (TYPES_REINTERP_SP): Likewise.
4671 (TYPES_REINTERP_US): Likewise.
4672 (TYPES_REINTERP_PS): Likewise.
4673 (aarch64_fold_builtin): New expression folding added.
4674 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4675 Declarations removed.
4676 (REINTERP_SS): Declarations added.
4677 (REINTERP_US): Likewise.
4678 (REINTERP_PS): Likewise.
4679 (REINTERP_SU): Likewise.
4680 (REINTERP_SP): Likewise.
4681 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4682 (vreinterpretq_p8_f64): Likewise.
4683 (vreinterpret_p16_f64): Likewise.
4684 (vreinterpretq_p16_f64): Likewise.
4685 (vreinterpret_f32_f64): Likewise.
4686 (vreinterpretq_f32_f64): Likewise.
4687 (vreinterpret_f64_f32): Likewise.
4688 (vreinterpret_f64_p8): Likewise.
4689 (vreinterpret_f64_p16): Likewise.
4690 (vreinterpret_f64_s8): Likewise.
4691 (vreinterpret_f64_s16): Likewise.
4692 (vreinterpret_f64_s32): Likewise.
4693 (vreinterpret_f64_s64): Likewise.
4694 (vreinterpret_f64_u8): Likewise.
4695 (vreinterpret_f64_u16): Likewise.
4696 (vreinterpret_f64_u32): Likewise.
4697 (vreinterpret_f64_u64): Likewise.
4698 (vreinterpretq_f64_f32): Likewise.
4699 (vreinterpretq_f64_p8): Likewise.
4700 (vreinterpretq_f64_p16): Likewise.
4701 (vreinterpretq_f64_s8): Likewise.
4702 (vreinterpretq_f64_s16): Likewise.
4703 (vreinterpretq_f64_s32): Likewise.
4704 (vreinterpretq_f64_s64): Likewise.
4705 (vreinterpretq_f64_u8): Likewise.
4706 (vreinterpretq_f64_u16): Likewise.
4707 (vreinterpretq_f64_u32): Likewise.
4708 (vreinterpretq_f64_u64): Likewise.
4709 (vreinterpret_s64_f64): Likewise.
4710 (vreinterpretq_s64_f64): Likewise.
4711 (vreinterpret_u64_f64): Likewise.
4712 (vreinterpretq_u64_f64): Likewise.
4713 (vreinterpret_s8_f64): Likewise.
4714 (vreinterpretq_s8_f64): Likewise.
4715 (vreinterpret_s16_f64): Likewise.
4716 (vreinterpretq_s16_f64): Likewise.
4717 (vreinterpret_s32_f64): Likewise.
4718 (vreinterpretq_s32_f64): Likewise.
4719 (vreinterpret_u8_f64): Likewise.
4720 (vreinterpretq_u8_f64): Likewise.
4721 (vreinterpret_u16_f64): Likewise.
4722 (vreinterpretq_u16_f64): Likewise.
4723 (vreinterpret_u32_f64): Likewise.
4724 (vreinterpretq_u32_f64): Likewise.
4726 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4728 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4729 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4730 (vreinterpret_p8_s8): Likewise.
4731 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4732 (vreinterpret_p8_s16): Likewise.
4733 (vreinterpret_p8_s32): Likewise.
4734 (vreinterpret_p8_s64): Likewise.
4735 (vreinterpret_p8_f32): Likewise.
4736 (vreinterpret_p8_u8): Likewise.
4737 (vreinterpret_p8_u16): Likewise.
4738 (vreinterpret_p8_u32): Likewise.
4739 (vreinterpret_p8_u64): Likewise.
4740 (vreinterpret_p8_p16): Likewise.
4741 (vreinterpretq_p8_s8): Likewise.
4742 (vreinterpretq_p8_s16): Likewise.
4743 (vreinterpretq_p8_s32): Likewise.
4744 (vreinterpretq_p8_s64): Likewise.
4745 (vreinterpretq_p8_f32): Likewise.
4746 (vreinterpretq_p8_u8): Likewise.
4747 (vreinterpretq_p8_u16): Likewise.
4748 (vreinterpretq_p8_u32): Likewise.
4749 (vreinterpretq_p8_u64): Likewise.
4750 (vreinterpretq_p8_p16): Likewise.
4751 (vreinterpret_p16_s8): Likewise.
4752 (vreinterpret_p16_s16): Likewise.
4753 (vreinterpret_p16_s32): Likewise.
4754 (vreinterpret_p16_s64): Likewise.
4755 (vreinterpret_p16_f32): Likewise.
4756 (vreinterpret_p16_u8): Likewise.
4757 (vreinterpret_p16_u16): Likewise.
4758 (vreinterpret_p16_u32): Likewise.
4759 (vreinterpret_p16_u64): Likewise.
4760 (vreinterpret_p16_p8): Likewise.
4761 (vreinterpretq_p16_s8): Likewise.
4762 (vreinterpretq_p16_s16): Likewise.
4763 (vreinterpretq_p16_s32): Likewise.
4764 (vreinterpretq_p16_s64): Likewise.
4765 (vreinterpretq_p16_f32): Likewise.
4766 (vreinterpretq_p16_u8): Likewise.
4767 (vreinterpretq_p16_u16): Likewise.
4768 (vreinterpretq_p16_u32): Likewise.
4769 (vreinterpretq_p16_u64): Likewise.
4770 (vreinterpretq_p16_p8): Likewise.
4771 (vreinterpret_f32_s8): Likewise.
4772 (vreinterpret_f32_s16): Likewise.
4773 (vreinterpret_f32_s32): Likewise.
4774 (vreinterpret_f32_s64): Likewise.
4775 (vreinterpret_f32_u8): Likewise.
4776 (vreinterpret_f32_u16): Likewise.
4777 (vreinterpret_f32_u32): Likewise.
4778 (vreinterpret_f32_u64): Likewise.
4779 (vreinterpret_f32_p8): Likewise.
4780 (vreinterpret_f32_p16): Likewise.
4781 (vreinterpretq_f32_s8): Likewise.
4782 (vreinterpretq_f32_s16): Likewise.
4783 (vreinterpretq_f32_s32): Likewise.
4784 (vreinterpretq_f32_s64): Likewise.
4785 (vreinterpretq_f32_u8): Likewise.
4786 (vreinterpretq_f32_u16): Likewise.
4787 (vreinterpretq_f32_u32): Likewise.
4788 (vreinterpretq_f32_u64): Likewise.
4789 (vreinterpretq_f32_p8): Likewise.
4790 (vreinterpretq_f32_p16): Likewise.
4791 (vreinterpret_s64_s8): Likewise.
4792 (vreinterpret_s64_s16): Likewise.
4793 (vreinterpret_s64_s32): Likewise.
4794 (vreinterpret_s64_f32): Likewise.
4795 (vreinterpret_s64_u8): Likewise.
4796 (vreinterpret_s64_u16): Likewise.
4797 (vreinterpret_s64_u32): Likewise.
4798 (vreinterpret_s64_u64): Likewise.
4799 (vreinterpret_s64_p8): Likewise.
4800 (vreinterpret_s64_p16): Likewise.
4801 (vreinterpretq_s64_s8): Likewise.
4802 (vreinterpretq_s64_s16): Likewise.
4803 (vreinterpretq_s64_s32): Likewise.
4804 (vreinterpretq_s64_f32): Likewise.
4805 (vreinterpretq_s64_u8): Likewise.
4806 (vreinterpretq_s64_u16): Likewise.
4807 (vreinterpretq_s64_u32): Likewise.
4808 (vreinterpretq_s64_u64): Likewise.
4809 (vreinterpretq_s64_p8): Likewise.
4810 (vreinterpretq_s64_p16): Likewise.
4811 (vreinterpret_u64_s8): Likewise.
4812 (vreinterpret_u64_s16): Likewise.
4813 (vreinterpret_u64_s32): Likewise.
4814 (vreinterpret_u64_s64): Likewise.
4815 (vreinterpret_u64_f32): Likewise.
4816 (vreinterpret_u64_u8): Likewise.
4817 (vreinterpret_u64_u16): Likewise.
4818 (vreinterpret_u64_u32): Likewise.
4819 (vreinterpret_u64_p8): Likewise.
4820 (vreinterpret_u64_p16): Likewise.
4821 (vreinterpretq_u64_s8): Likewise.
4822 (vreinterpretq_u64_s16): Likewise.
4823 (vreinterpretq_u64_s32): Likewise.
4824 (vreinterpretq_u64_s64): Likewise.
4825 (vreinterpretq_u64_f32): Likewise.
4826 (vreinterpretq_u64_u8): Likewise.
4827 (vreinterpretq_u64_u16): Likewise.
4828 (vreinterpretq_u64_u32): Likewise.
4829 (vreinterpretq_u64_p8): Likewise.
4830 (vreinterpretq_u64_p16): Likewise.
4831 (vreinterpret_s8_s16): Likewise.
4832 (vreinterpret_s8_s32): Likewise.
4833 (vreinterpret_s8_s64): Likewise.
4834 (vreinterpret_s8_f32): Likewise.
4835 (vreinterpret_s8_u8): Likewise.
4836 (vreinterpret_s8_u16): Likewise.
4837 (vreinterpret_s8_u32): Likewise.
4838 (vreinterpret_s8_u64): Likewise.
4839 (vreinterpret_s8_p8): Likewise.
4840 (vreinterpret_s8_p16): Likewise.
4841 (vreinterpretq_s8_s16): Likewise.
4842 (vreinterpretq_s8_s32): Likewise.
4843 (vreinterpretq_s8_s64): Likewise.
4844 (vreinterpretq_s8_f32): Likewise.
4845 (vreinterpretq_s8_u8): Likewise.
4846 (vreinterpretq_s8_u16): Likewise.
4847 (vreinterpretq_s8_u32): Likewise.
4848 (vreinterpretq_s8_u64): Likewise.
4849 (vreinterpretq_s8_p8): Likewise.
4850 (vreinterpretq_s8_p16): Likewise.
4851 (vreinterpret_s16_s8): Likewise.
4852 (vreinterpret_s16_s32): Likewise.
4853 (vreinterpret_s16_s64): Likewise.
4854 (vreinterpret_s16_f32): Likewise.
4855 (vreinterpret_s16_u8): Likewise.
4856 (vreinterpret_s16_u16): Likewise.
4857 (vreinterpret_s16_u32): Likewise.
4858 (vreinterpret_s16_u64): Likewise.
4859 (vreinterpret_s16_p8): Likewise.
4860 (vreinterpret_s16_p16): Likewise.
4861 (vreinterpretq_s16_s8): Likewise.
4862 (vreinterpretq_s16_s32): Likewise.
4863 (vreinterpretq_s16_s64): Likewise.
4864 (vreinterpretq_s16_f32): Likewise.
4865 (vreinterpretq_s16_u8): Likewise.
4866 (vreinterpretq_s16_u16): Likewise.
4867 (vreinterpretq_s16_u32): Likewise.
4868 (vreinterpretq_s16_u64): Likewise.
4869 (vreinterpretq_s16_p8): Likewise.
4870 (vreinterpretq_s16_p16): Likewise.
4871 (vreinterpret_s32_s8): Likewise.
4872 (vreinterpret_s32_s16): Likewise.
4873 (vreinterpret_s32_s64): Likewise.
4874 (vreinterpret_s32_f32): Likewise.
4875 (vreinterpret_s32_u8): Likewise.
4876 (vreinterpret_s32_u16): Likewise.
4877 (vreinterpret_s32_u32): Likewise.
4878 (vreinterpret_s32_u64): Likewise.
4879 (vreinterpret_s32_p8): Likewise.
4880 (vreinterpret_s32_p16): Likewise.
4881 (vreinterpretq_s32_s8): Likewise.
4882 (vreinterpretq_s32_s16): Likewise.
4883 (vreinterpretq_s32_s64): Likewise.
4884 (vreinterpretq_s32_f32): Likewise.
4885 (vreinterpretq_s32_u8): Likewise.
4886 (vreinterpretq_s32_u16): Likewise.
4887 (vreinterpretq_s32_u32): Likewise.
4888 (vreinterpretq_s32_u64): Likewise.
4889 (vreinterpretq_s32_p8): Likewise.
4890 (vreinterpretq_s32_p16): Likewise.
4891 (vreinterpret_u8_s8): Likewise.
4892 (vreinterpret_u8_s16): Likewise.
4893 (vreinterpret_u8_s32): Likewise.
4894 (vreinterpret_u8_s64): Likewise.
4895 (vreinterpret_u8_f32): Likewise.
4896 (vreinterpret_u8_u16): Likewise.
4897 (vreinterpret_u8_u32): Likewise.
4898 (vreinterpret_u8_u64): Likewise.
4899 (vreinterpret_u8_p8): Likewise.
4900 (vreinterpret_u8_p16): Likewise.
4901 (vreinterpretq_u8_s8): Likewise.
4902 (vreinterpretq_u8_s16): Likewise.
4903 (vreinterpretq_u8_s32): Likewise.
4904 (vreinterpretq_u8_s64): Likewise.
4905 (vreinterpretq_u8_f32): Likewise.
4906 (vreinterpretq_u8_u16): Likewise.
4907 (vreinterpretq_u8_u32): Likewise.
4908 (vreinterpretq_u8_u64): Likewise.
4909 (vreinterpretq_u8_p8): Likewise.
4910 (vreinterpretq_u8_p16): Likewise.
4911 (vreinterpret_u16_s8): Likewise.
4912 (vreinterpret_u16_s16): Likewise.
4913 (vreinterpret_u16_s32): Likewise.
4914 (vreinterpret_u16_s64): Likewise.
4915 (vreinterpret_u16_f32): Likewise.
4916 (vreinterpret_u16_u8): Likewise.
4917 (vreinterpret_u16_u32): Likewise.
4918 (vreinterpret_u16_u64): Likewise.
4919 (vreinterpret_u16_p8): Likewise.
4920 (vreinterpret_u16_p16): Likewise.
4921 (vreinterpretq_u16_s8): Likewise.
4922 (vreinterpretq_u16_s16): Likewise.
4923 (vreinterpretq_u16_s32): Likewise.
4924 (vreinterpretq_u16_s64): Likewise.
4925 (vreinterpretq_u16_f32): Likewise.
4926 (vreinterpretq_u16_u8): Likewise.
4927 (vreinterpretq_u16_u32): Likewise.
4928 (vreinterpretq_u16_u64): Likewise.
4929 (vreinterpretq_u16_p8): Likewise.
4930 (vreinterpretq_u16_p16): Likewise.
4931 (vreinterpret_u32_s8): Likewise.
4932 (vreinterpret_u32_s16): Likewise.
4933 (vreinterpret_u32_s32): Likewise.
4934 (vreinterpret_u32_s64): Likewise.
4935 (vreinterpret_u32_f32): Likewise.
4936 (vreinterpret_u32_u8): Likewise.
4937 (vreinterpret_u32_u16): Likewise.
4938 (vreinterpret_u32_u64): Likewise.
4939 (vreinterpret_u32_p8): Likewise.
4940 (vreinterpret_u32_p16): Likewise.
4941 (vreinterpretq_u32_s8): Likewise.
4942 (vreinterpretq_u32_s16): Likewise.
4943 (vreinterpretq_u32_s32): Likewise.
4944 (vreinterpretq_u32_s64): Likewise.
4945 (vreinterpretq_u32_f32): Likewise.
4946 (vreinterpretq_u32_u8): Likewise.
4947 (vreinterpretq_u32_u16): Likewise.
4948 (vreinterpretq_u32_u64): Likewise.
4949 (vreinterpretq_u32_p8): Likewise.
4950 (vreinterpretq_u32_p16): Likewise.
4952 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4954 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4956 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4958 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4959 (vqnegd_s64): Likewise.
4960 (vqabs_s64): Likewise.
4961 (vqabsd_s64): Likewise.
4963 2014-04-22 Richard Henderson <rth@redhat.com>
4965 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4966 computation to the top of the loop.
4968 2014-04-22 Renlin <renlin.li@arm.com>
4969 Jiong Wang <jiong.wang@arm.com>
4971 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4972 * config/aarch64/aarch64.c (aarch64_layout_frame)
4973 (aarch64_initial_elimination_offset): Likewise.
4975 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4977 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4980 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4982 * machmode.h (bitwise_mode_for_mode): Declare.
4983 * stor-layout.h (bitwise_type_for_mode): Likewise.
4984 * stor-layout.c (bitwise_mode_for_mode): New function.
4985 (bitwise_type_for_mode): Likewise.
4986 * builtins.c (fold_builtin_memory_op): Use it instead of
4987 int_mode_for_mode and build_nonstandard_integer_type.
4989 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4991 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4992 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4993 (*-*-solaris2*): Simplify.
4994 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4995 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4996 *-*-solaris2.9* handling.
4998 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5000 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5001 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5003 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5004 * configure: Regenerate.
5006 * config/i386/sol2-9.h: Remove.
5008 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5009 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5010 Remove Solaris 9 references.
5012 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5014 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5015 (floatuns<GPI:mode><GPF:mode>2): Remove.
5016 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5017 and floatuns conversions.
5018 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5019 and floatuns conversions.
5020 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5021 (w1,w2): New mode attributes for inequal width conversions.
5023 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5025 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5026 the output asm format.
5028 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5030 * config/aarch64/aarch64-simd.md
5031 (aarch64_cm<optab>di): Always split.
5032 (*aarch64_cm<optab>di): New.
5033 (aarch64_cmtstdi): Always split.
5034 (*aarch64_cmtstdi): New.
5036 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5038 PR tree-optimization/60823
5039 * omp-low.c (ipa_simd_modify_function_body): Go through
5040 all SSA_NAMEs and for those refering to vector arguments
5041 which are going to be replaced adjust SSA_NAME_VAR and,
5042 if it is a default definition, change it into a non-default
5043 definition assigned at the beginning of function from new_decl.
5044 (ipa_simd_modify_stmt_ops): Rewritten.
5045 * tree-dfa.c (set_ssa_default_def): When removing default def,
5046 check for NULL loc instead of NULL *loc.
5048 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5050 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5051 restrictions on core registers for DImode values in Thumb2.
5053 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5055 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5056 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5058 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5060 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5061 (*iordi_notzesidi_di): Likewise.
5062 (*iordi_notsesidi_di): Likewise.
5064 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5066 * config/arm/arm-protos.h (tune_params): New struct members.
5067 * config/arm/arm.c: Initialise tune_params per processor.
5068 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5069 for speed, based on new tune_params.
5071 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5073 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5074 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5075 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5076 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5077 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5078 (vrnda_f64): Likewise.
5079 (vrndi_f64): Likewise.
5080 (vrndm_f64): Likewise.
5081 (vrndn_f64): Likewise.
5082 (vrndp_f64): Likewise.
5083 (vrndx_f64): Likewise.
5085 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5087 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5088 GET_MODE_SIZE argument is enum machine_mode.
5090 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5093 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5094 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5096 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5099 * asan.c (asan_emit_stack_protection): Force the base to align to
5100 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5101 appropriate bits if STRICT_ALIGNMENT.
5102 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5104 (expand_used_vars): Leave a space in the stack frame for alignment
5105 if STRICT_ALIGNMENT.
5107 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5109 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5111 (gimple_store_p): Likewise.
5112 (gimple_assign_load_p): Likewise.
5113 (gimple_assign_cast_p): Likewise.
5114 (gimple_clobber_p): Likewise.
5116 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5117 rather than a gimple.
5118 (gimple_assign_cast_p): Likewise.
5120 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5123 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5124 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5126 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5127 more debug information for E500 if -mdebug=reg.
5129 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5132 * config/i386/i386.c (ix86_expand_builtin)
5133 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5134 register for target RTX.
5135 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5137 2014-04-18 Cong Hou <congh@google.com>
5139 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5140 the widen-mult pattern by handling two operands with different sizes,
5141 and operands whose size is smaller than half of the result type.
5143 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5145 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5146 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5147 (do_estimate_edge_time): Compute it.
5148 * ipa-inline.c (want_inline_small_function_p): Bypass
5149 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5151 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5153 * ipa-inline.c (spec_rem): New static variable.
5154 (dump_overall_stats): New function.
5155 (dump_inline_stats): New function.
5157 2014-04-18 Richard Henderson <rth@redhat.com>
5159 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5160 to GET_MODE_SIZE, not a reg_class_t.
5162 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5164 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5165 (vsx_xxmrglw_<mode>): Likewise.
5167 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5170 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5171 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5172 (rs6000_init_hard_regno_mode_ok): Likewise.
5174 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5176 * ipa-inline.c (inline_small_functions): Account only non-cold
5178 * doc/invoke.texi (inline-unit-growth): Update documentation.
5180 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5182 * config/rs6000/rs6000.md (addti3, subti3): New.
5184 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5187 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5188 comment. Check optimize_insn_for_size_p instead of
5189 optimize_insn_for_speed_p.
5191 2014-04-17 Martin Jambor <mjambor@suse.cz>
5193 * gimple-iterator.c (gsi_start_edge): New function.
5194 * gimple-iterator.h (gsi_start_edge): Declare.
5195 * tree-sra.c (single_non_eh_succ): New function.
5196 (disqualify_ops_if_throwing_stmt): Renamed to
5197 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5198 having one non-EH successor BB.
5199 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5200 generate loads into replacements.
5201 (sra_modify_assign): Likewise and and also use the simple path for
5203 (sra_modify_function_body): Commit statements on edges.
5205 2014-04-17 Richard Biener <rguenther@suse.de>
5208 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5209 comparison results and add clarifying comment.
5211 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5213 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5214 (blank_mode): Initialize it.
5215 (emit_mode_size_inline, emit_mode_nunits_inline,
5216 emit_mode_inner_inline): New functions.
5217 (emit_insn_modes_h): Call them and surround their output with
5218 #if GCC_VERSION >= 4001 ... #endif.
5219 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5220 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5221 mode_* arrays if the argument is __builtin_constant_p.
5222 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5223 is enum machine_mode.
5225 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5227 * passes.c (opt_pass::execute): Adjust.
5228 (pass_manager::execute_pass_mode_switching): Likewise.
5229 (early_local_passes::execute): Likewise.
5230 (execute_one_pass): Pass cfun to the pass's execute method.
5231 * tree-pass.h (opt_pass::execute): Add function * argument.
5232 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5233 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5234 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5235 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5236 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5237 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5238 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5239 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5240 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5241 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5242 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5243 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5244 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5245 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5246 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5247 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5248 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5249 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5250 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5251 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5252 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5253 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5254 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5255 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5256 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5257 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5258 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5259 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5260 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5263 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5265 * passes.c (opt_pass::gate): Take function * argument.
5266 (gate_all_early_local_passes): Merge into
5267 (early_local_passes::gate): this.
5268 (gate_all_early_optimizations): Merge into
5269 (all_early_optimizations::gate): this.
5270 (gate_all_optimizations): Mege into
5271 (all_optimizations::gate): this.
5272 (gate_all_optimizations_g): Merge into
5273 (all_optimizations_g::gate): this.
5274 (gate_rest_of_compilation): Mege into
5275 (rest_of_compilation::gate): this.
5276 (gate_postreload): Merge into
5277 (postreload::gate): this.
5278 (dump_one_pass): Pass cfun to the pass's gate method.
5279 (execute_ipa_summary_passes): Likewise.
5280 (execute_one_pass): Likewise.
5281 (ipa_write_summaries_2): Likewise.
5282 (ipa_write_optimization_summaries_1): Likewise.
5283 (ipa_read_summaries_1): Likewise.
5284 (ipa_read_optimization_summaries_1): Likewise.
5285 (execute_ipa_stmt_fixups): Likewise.
5286 * tree-pass.h (opt_pass::gate): Add function * argument.
5287 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5288 combine-stack-adj.c, combine.c, compare-elim.c,
5289 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5290 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5291 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5292 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5293 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5294 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5295 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5296 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5297 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5298 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5299 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5300 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5301 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5302 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5303 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5304 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5305 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5306 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5307 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5308 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5309 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5310 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5311 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5312 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5313 var-tracking.c, vtable-verify.c, web.c: Adjust.
5315 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5317 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5318 * configure: Regenerate.
5320 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5322 * passes.c (dump_one_pass): don't check pass->has_gate.
5323 (execute_ipa_summary_passes): Likewise.
5324 (execute_one_pass): Likewise.
5325 (ipa_write_summaries_2): Likewise.
5326 (ipa_write_optimization_summaries_1): Likewise.
5327 (ipa_read_optimization_summaries_1): Likewise.
5328 (execute_ipa_stmt_fixups): Likewise.
5329 * tree-pass.h (pass_data::has_gate): Remove.
5330 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5331 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5332 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5333 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5334 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5335 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5336 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5337 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5338 gimple-low.c, gimple-ssa-isolate-paths.c,
5339 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5340 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5341 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5342 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5343 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5344 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5345 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5346 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5347 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5348 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5349 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5350 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5351 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5352 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5353 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5354 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5355 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5356 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5357 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5358 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5359 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5360 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5363 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5365 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5367 * passes.c (pass_manager::register_dump_files_1): Merge into
5368 (pass_manager::register_dump_files): this, and remove its handling of
5369 properties since the pass always has the properties anyway.
5370 (pass_manager::pass_manager): Adjust.
5372 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5374 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5375 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5376 dealing with properties.
5377 (pass_manager::register_dump_files): Adjust.
5379 2014-03-20 Mark Wielaard <mjw@redhat.com>
5381 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5382 then represent the bound as normal constant value.
5384 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5387 Forward port from 4.8 branch
5388 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5390 * config/i386/bmiintrin.h (_blsi_u32): New.
5394 (_blsmsk_u32): Ditto.
5395 (_blsmsk_u64): Ditto.
5396 (_tzcnt_u32): Ditto.
5397 (_tzcnt_u64): Ditto.
5399 2014-04-17 Kito Cheng <kito@0xlab.org>
5401 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5403 2014-04-17 Richard Biener <rguenther@suse.de>
5406 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5407 boolean results for comparisons.
5409 2014-04-17 Richard Biener <rguenther@suse.de>
5411 PR tree-optimization/60836
5412 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5413 initial PHI args to be gimple values.
5415 2014-04-17 Richard Biener <rguenther@suse.de>
5417 PR tree-optimization/60841
5418 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5419 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5420 of stmts to SLP build.
5421 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5422 (vect_analyze_slp): Likewise.
5423 (vect_analyze_slp_instance): Likewise.
5424 (vect_build_slp_tree): Limit overall SLP tree growth.
5425 * tree-vectorizer.h (vect_analyze_data_refs,
5426 vect_analyze_slp): Adjust prototypes.
5428 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5430 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5433 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5435 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5436 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5437 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5438 for TARGET_SLOW_PSHUFB
5440 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5442 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5443 * config/i386/i386.c (intel_cost): Ditto.
5445 2014-04-17 Joey Ye <joey.ye@arm.com>
5447 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5449 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5451 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5454 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5456 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5457 (type_all_derivations_known_p): New predicate.
5458 (type_all_ctors_visible_p): New predicate.
5459 (type_possibly_instantiated_p): New predicate.
5460 (get_odr_type): Compute all_derivations_known.
5461 (dump_odr_type): Dump the flag.
5462 (maybe_record_type): Cleanup.
5463 (record_target_from_binfo): Add bases_to_consider array;
5464 record bases for types w/o instances and skip CXX destructor.
5465 (possible_polymorphic_call_targets_1): Add bases_to_consider
5466 and consider_construction parameters; check if type may have instance.
5467 (get_polymorphic_call_info): Set maybe_in_construction to true
5468 when we know nothing.
5469 (record_targets_from_bases): Skip CXX destructors; they are
5470 never called for types in construction.
5471 (possible_polymorphic_call_targets): Do not record target when
5472 type may not have instance.
5474 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5477 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5478 external aliases alive, too.
5480 2014-04-16 Andrew Pinski <apinski@cavium.com>
5482 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5485 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5487 * final.c (compute_alignments): Do not apply loop alignment to a block
5488 falling through to the exit.
5490 2014-04-16 Catherine Moore <clm@codesourcery.com>
5492 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5493 Adjust constraints for microMIPS store patterns.
5495 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5497 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5499 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5501 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5502 (append_use): Run at -O0.
5503 (append_vdef): Likewise.
5504 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5505 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5507 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5509 PR tree-optimization/60844
5510 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5511 (propagate_op_to_single_use, remove_visited_stmt_chain,
5512 linearize_expr, repropagate_negates, reassociate_bb): Use it
5513 instead of gsi_remove.
5515 2014-04-16 Martin Jambor <mjambor@suse.cz>
5517 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5518 ipa_transforms_to_apply.
5519 (cgraph_function_versioning): Assert that old_node has empty
5520 ipa_transforms_to_apply.
5521 * trans-mem.c (ipa_tm_create_version): Likewise.
5522 * tree-inline.c (tree_function_versioning): Do not duplicate
5523 ipa_transforms_to_apply.
5525 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5528 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5530 Pass necessary as flags on 64-bit Solaris/x86.
5531 Use lowercase relocs for x86_64-*-*.
5532 * configure: Regenerate.
5534 2014-04-15 Jan Hubicka <jh@suse.cz>
5536 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5537 (maybe_record_node, likely_target_p): Use it.
5539 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5542 Revert following patch
5544 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5547 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5548 software floating point or no floating point registers, do not
5549 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5550 in GPRs that occurs after we tested for GPRs that would never be
5553 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5554 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5555 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5556 specifically allow DDmode, since that does not use the SPE SIMD
5559 2014-03-21 Mark Wielaard <mjw@redhat.com>
5561 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5562 as unsigned or int depending on type and value used.
5564 2014-04-15 Richard Biener <rguenther@suse.de>
5566 PR rtl-optimization/56965
5567 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5568 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5570 * alias.c (true_dependence_1): Do not call
5571 nonoverlapping_component_refs_p.
5572 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5573 nonoverlapping_component_refs_p.
5574 (indirect_refs_may_alias_p): Likewise.
5576 2014-04-15 Teresa Johnson <tejohnson@google.com>
5578 * cfg.c (dump_bb_info): Fix flags check.
5579 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5581 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5583 PR rtl-optimization/60663
5584 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5587 2014-04-15 Richard Biener <rguenther@suse.de>
5589 * lto-streamer.h (LTO_major_version): Bump to 4.
5591 2014-04-15 Richard Biener <rguenther@suse.de>
5593 * common.opt (lto_partition_model): New enum.
5594 (flto-partition=): Merge separate options with a single with argument,
5595 add -flto-partition=one support.
5596 * flag-types.h (enum lto_partition_model): Declare.
5597 * opts.c (finish_options): Remove duplicate -flto-partition=
5599 * lto-wrapper.c (run_gcc): Adjust.
5601 2014-04-15 Richard Biener <rguenther@suse.de>
5603 * alias.c (ncr_compar): New function.
5604 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5606 2014-04-15 Richard Biener <rguenther@suse.de>
5608 * alias.c (record_component_aliases): Do not walk BINFOs.
5610 2014-04-15 Richard Biener <rguenther@suse.de>
5612 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5613 Add struct function argument and adjust.
5614 (find_func_aliases_for_call): Likewise.
5615 (find_func_aliases): Likewise.
5616 (find_func_clobbers): Likewise.
5617 (intra_create_variable_infos): Likewise.
5618 (compute_points_to_sets): Likewise.
5619 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5621 2014-04-15 Richard Biener <rguenther@suse.de>
5623 * tree.c (iterative_hash_expr): Use enum tree_code_class
5624 to store TREE_CODE_CLASS.
5625 (tree_block): Likewise.
5626 (tree_set_block): Likewise.
5627 * tree.h (fold_build_pointer_plus_loc): Use
5628 convert_to_ptrofftype_loc.
5630 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5633 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5636 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5638 * cfgloop.h (struct loop): Move force_vectorize down.
5639 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5640 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5641 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5642 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5643 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5644 * tree-core.h (enum annot_expr_kind): Add new kind values.
5645 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5646 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5648 * tree.def (ANNOTATE_EXPR): Tweak comment.
5650 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5652 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5655 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5657 * tree.h (TYPE_IDENTIFIER): Declare.
5658 * tree.c (subrange_type_for_debug_p): Use it.
5659 * godump.c (go_format_type): Likewise.
5660 * dwarf2out.c (is_cxx_auto, modified_type_die,
5661 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5662 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5664 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5667 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5669 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5671 * config/i386/i386.c (examine_argument): Return bool. Return true if
5672 parameter should be passed in memory.
5673 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5674 (construct_container): Update calls to examine_argument.
5675 (function_arg_advance_64): Ditto.
5676 (return_in_memory_32): Merge with ix86_return_in_memory.
5677 (return_in_memory_64): Ditto.
5678 (return_in_memory_ms_64): Ditto.
5680 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5682 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5683 * coverage.c (coverage_compute_profile_id): Handle externally visible
5686 2014-04-14 Martin Jambor <mjambor@suse.cz>
5688 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5689 DECL_DISREGARD_INLINE_LIMITS functions.
5691 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5694 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5696 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5699 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5700 optimize_insn_for_speed_p instead of
5701 optimize_function_for_speed_p.
5703 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5705 * doc/invoke.texi (free): Document AArch64.
5707 2014-04-14 Richard Biener <rguenther@suse.de>
5709 PR tree-optimization/60042
5710 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5711 (insert_into_preds_of_block): Do not prevent PHI insertion
5712 for REFERENCE exprs here ...
5713 (eliminate_dom_walker::before_dom_children): ... but prevent
5714 their use here under similar conditions when applied to the
5715 IL after PRE optimizations.
5717 2014-04-14 Richard Biener <rguenther@suse.de>
5719 * passes.def: Move early points-to after early SRA.
5721 2014-04-14 Richard Biener <rguenther@suse.de>
5723 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5724 check for which sign-changes we allow when forwarding
5725 a converted value into a switch.
5727 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5729 * stor-layout.c (place_field): Finalize non-constant offset for the
5732 2014-04-14 Richard Biener <rguenther@suse.de>
5734 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5736 (expand_switch_using_bit_tests_p): Likewise.
5737 (process_switch): Compute and pass on speed_p based on the
5739 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5740 optimize_bb_for_speed_p.
5742 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5744 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5745 * function.h (struct function): Rename has_force_vect_loops into
5746 has_force_vectorize_loops.
5747 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5748 (input_struct_function_base): Likewise.
5749 * lto-streamer-out.c (output_cfg): Likewise.
5750 (output_struct_function_base): Likewise.
5751 * omp-low.c (expand_omp_simd): Likewise.
5752 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5753 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5754 (version_loop_for_if_conversion): Likewise.
5755 (tree_if_conversion): Likewise.
5756 (main_tree_if_conversion): Likewise.
5757 (gate_tree_if_conversion): Likewise.
5758 * tree-inline.c (copy_loops): Likewise.
5759 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5760 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5761 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5762 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5763 * tree-vectorizer.c (vectorize_loops): Likewise.
5764 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5766 2014-04-14 Richard Biener <rguenther@suse.de>
5769 * lto-streamer-out.c (wrap_refs): New function.
5770 (lto_output): Wrap symbol references in global initializes in
5771 type-preserving MEM_REFs.
5773 2014-04-14 Christian Bruel <christian.bruel@st.com>
5775 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5777 2014-04-14 Christian Bruel <christian.bruel@st.com>
5779 * config/sh/sh.md (setmemqi): New expand pattern.
5780 * config/sh/sh.h (CLEAR_RATIO): Define.
5781 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5782 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5784 2014-04-14 Richard Biener <rguenther@suse.de>
5787 * fold-const.c (negate_expr_p): Don't negate directional rounding
5789 (fold_negate_expr): Likewise.
5791 2014-04-14 Richard Biener <rguenther@suse.de>
5793 PR tree-optimization/59817
5794 PR tree-optimization/60453
5795 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5796 recursion to catch all CHRECs in the scalar evolution and restrict
5797 the predicate for the remains appropriately.
5799 2014-04-12 Catherine Moore <clm@codesourcery.com>
5801 * config/mips/constraints.md: Add new register constraint "kb".
5802 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5803 (*movhi_internal): Likewise.
5804 (*movqi_internal): Likewise.
5805 * config/mips/mips.h (M16_STORE_REGS): New register class.
5806 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5807 (REG_CLASS_CONTENTS): Likewise.
5808 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5810 2014-04-11 Tobias Burnus <burnus@net-b.de>
5813 * doc/invoke.texi (-Wformat-signedness): Document it.
5814 (Wformat=2): Mention that this enables -Wformat-signedness.
5816 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5818 * common/config/epiphany/epiphany-common.c
5819 (epiphany_option_optimization_table): Enable section anchors by
5820 default at -O1 or higher.
5821 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5822 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5823 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5824 carries no extra cost.
5825 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5826 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5827 * config/epiphany/predicates.md (memclob_operand): New predicate.
5828 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5829 Use memclob_operand predicate and X constraint for operand 3.
5831 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5833 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5834 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5837 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5839 PR rtl-optimization/60651
5840 * mode-switching.c (optimize_mode_switching): Make sure to emit
5841 sets of a lower numbered entity before sets of a higher numbered
5842 entity to a mode of the same or lower priority.
5843 When creating a seginfo for a basic block that starts with a code
5844 label, move the insertion point past the code label.
5845 (new_seginfo): Document and enforce requirement that
5846 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5847 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5848 * doc/tm.texi: Regenerate.
5850 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5853 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5855 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5857 * BASE-VER: Set to 4.10.0.
5859 2014-04-11 Tobias Burnus <burnus@net-b.de>
5862 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5863 * doc/gcc.texi (Service): Update description in the @menu
5864 * doc/invoke.texi (Option Summary): Remove misplaced and
5867 2014-04-11 Steve Ellcey <sellcey@mips.com>
5868 Jakub Jelinek <jakub@redhat.com>
5871 * expr.c (convert_move): Use emit_store_flag_force instead of
5872 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5875 2014-04-11 Richard Biener <rguenther@suse.de>
5878 * varasm.c (assemble_alias): Avoid endless error reporting
5879 recursion by setting TREE_ASM_WRITTEN.
5881 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5883 * config/s390/s390.md: Add a splitter for NOT rtx.
5885 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5887 PR rtl-optimization/60663
5888 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5890 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5891 Jakub Jelinek <jakub@redhat.com>
5894 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5895 flag from decl_node to node.
5897 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5900 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5901 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5902 ameliorating the cases where it can be.
5904 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5907 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5909 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5910 (loadsync_<mode>): Change mode.
5911 (load_quadpti, store_quadpti): New.
5912 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5913 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5914 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5916 2014-04-09 Cong Hou <congh@google.com>
5919 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5922 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5924 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5925 instead of vnor to exploit possible fusion opportunity in the
5927 (altivec_expand_vec_perm_const_le): Likewise.
5929 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5931 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5932 (loadsync_<mode>): Change mode.
5933 (load_quadpti, store_quadpti): New.
5934 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5935 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5937 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5940 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5941 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5942 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5944 2014-04-08 Richard Biener <rguenther@suse.de>
5947 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5948 a 64bit widest int print double-int similar to on HWI64 hosts.
5950 2014-04-08 Richard Biener <rguenther@suse.de>
5952 PR tree-optimization/60785
5953 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5954 default defs properly.
5956 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5958 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5959 (Weffc++): Likewise.
5961 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5963 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5964 set completep to false rather than true.
5966 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5969 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5970 ARM_TARGET2_DWARF_FORMAT.
5972 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5975 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5976 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5979 2014-04-07 Richard Biener <rguenther@suse.de>
5981 PR tree-optimization/60766
5982 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5983 (may_eliminate_iv): Convert cand_value_at result to desired type.
5985 2014-04-07 Jason Merrill <jason@redhat.com>
5988 * common.opt (-fno-gnu-unique): Add.
5989 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5991 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5993 * haifa-sched.c: Fix outdated function reference and minor
5994 grammar errors in introductory comment.
5996 2014-04-07 Richard Biener <rguenther@suse.de>
5999 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6001 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6003 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6006 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6007 size accounting for thunks.
6008 (pa_asm_output_mi_thunk): Use final_start_function() and
6009 final_end_function() to output function start and end directives.
6011 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6013 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6014 device specific ISA/ feature information. Remove short_sp and
6015 errata_skip ds. Add avr_device_specific_features enum to have device
6017 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6018 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6019 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6020 updated device specific info.
6021 * config/avr/avr-mcus.def: Merge device specific details to
6022 dev_attribute field.
6023 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6025 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6026 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6027 assembler if RMW isa supported by current device.
6028 * config/avr/genmultilib.awk: Update as device info structure changed.
6029 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6031 2014-04-04 Cong Hou <congh@google.com>
6033 PR tree-optimization/60656
6034 * tree-vect-stmts.c (supportable_widening_operation):
6035 Fix a bug that elements in a vector with vect_used_by_reduction
6036 property are incorrectly reordered when the operation on it is not
6037 consistant with the one in reduction operation.
6039 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6041 PR rtl-optimization/60155
6042 * gcse.c (record_set_data): New function.
6043 (single_set_gcse): New function.
6044 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6045 (hoist_code): Likewise.
6046 (get_pressure_class_and_nregs): Likewise.
6048 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6050 * explow.c (probe_stack_range): Emit a final optimization blockage.
6052 2014-04-04 Anthony Green <green@moxielogic.com>
6054 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6057 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6060 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6061 flags are set only during streaming.
6062 * ipa.c (process_references, walk_polymorphic_call_targets,
6063 symtab_remove_unreachable_nodes): Drop bodies of always inline
6064 after early inlining.
6065 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6067 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6068 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6071 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6074 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6077 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6079 (cortex_a53_fdivd): Likewise.
6081 2014-04-04 Martin Jambor <mjambor@suse.cz>
6084 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6086 * cgraph.c (clone_of_p): Also return true if thunks match.
6087 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6088 cgraph_function_or_thunk_node and an obsolete comment.
6089 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6091 (build_function_decl_skip_args): Likewise.
6092 (set_new_clone_decl_and_node_flags): New function.
6093 (duplicate_thunk_for_node): Likewise.
6094 (redirect_edge_duplicating_thunks): Likewise.
6095 (cgraph_clone_node): New parameter args_to_skip, pass it to
6096 redirect_edge_duplicating_thunks which is called instead of
6097 cgraph_redirect_edge_callee.
6098 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6099 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6101 2014-04-04 Jeff Law <law@redhat.com>
6104 * config/arm/predicates.md (const_int_I_operand): New predicate.
6105 (const_int_M_operand): Similarly.
6106 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6108 (insv_t2, extv_reg, extzv_t2): Likewise.
6109 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6110 (pop_multiple_with_writeback_and_return): Likewise.
6111 (vfp_pop_multiple_with_writeback): Likewise
6113 2014-04-04 Richard Biener <rguenther@suse.de>
6116 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6117 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6119 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6120 * gimplify.c (gimple_add_tmp_var_fn): New function.
6121 * gimple-expr.h (create_tmp_reg_fn): Declare.
6122 * gimple-expr.c (create_tmp_reg_fn): New function.
6123 * gimple-low.c (record_vars_into): Don't change cfun.
6124 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6125 code generation without cfun.
6127 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6130 * Makefile.in (install-driver): Fix shell scripting.
6132 2014-04-03 Cong Hou <congh@google.com>
6134 PR tree-optimization/60505
6135 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6136 threshold of number of iterations below which no vectorization
6138 * tree-vect-loop.c (new_loop_vec_info):
6139 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6140 * tree-vect-loop.c (vect_analyze_loop_operations):
6141 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6142 * tree-vect-loop.c (vect_transform_loop):
6143 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6144 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6145 of iterations of the loop and see if we should build the epilogue.
6147 2014-04-03 Richard Biener <rguenther@suse.de>
6149 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6150 (streamer_tree_cache_create): Adjust.
6151 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6152 to allow optional nodes array.
6153 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6154 (streamer_tree_cache_append): Likewise.
6155 (streamer_tree_cache_create): Create nodes array optionally
6156 as specified by parameter.
6157 * lto-streamer-out.c (create_output_block): Avoid maintaining
6158 the node array in the writer cache.
6159 (DFS_write_tree): Remove assertion.
6160 (produce_asm_for_decls): Free the out decl state hash table early.
6161 * lto-streamer-in.c (lto_data_in_create): Adjust for
6162 streamer_tree_cache_create prototype change.
6164 2014-04-03 Richard Biener <rguenther@suse.de>
6166 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6167 set TREE_CHAIN to NULL_TREE.
6169 2014-04-03 Richard Biener <rguenther@suse.de>
6171 PR tree-optimization/60740
6172 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6173 over all GIMPLE_COND operands.
6175 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6177 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6178 (Weffc++): Remove Scott's numbering, merge lists and reference
6181 2014-04-03 Nick Clifton <nickc@redhat.com>
6183 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6186 2014-04-03 Martin Jambor <mjambor@suse.cz>
6188 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6189 mention gcc_unreachable before failing.
6190 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6193 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6196 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6197 inconsistent code and instead mark the context inconsistent.
6198 (possible_polymorphic_call_targets): For inconsistent contexts
6199 return empty complete list.
6201 2014-04-02 Anthony Green <green@moxielogic.com>
6203 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6204 (extendqisi2, extendhisi2): Define.
6205 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6206 (WCHAR_TYPE): Change to unsigned int.
6208 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6210 PR tree-optimization/60733
6211 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6212 insertion point for PHI candidates to be the end of the feeding
6213 block for the PHI argument.
6215 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6217 PR rtl-optimization/60650
6218 * lra-constraints.c (process_alt_operands): Decrease reject for
6219 earlyclobber matching.
6221 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6223 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6225 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6227 * config/spu/spu.c (pad_bb): Do not crash when the last
6228 insn is CODE_FOR_blockage.
6230 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6232 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6233 lies outside the target mode.
6235 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6238 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6239 software floating point or no floating point registers, do not
6240 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6241 in GPRs that occurs after we tested for GPRs that would never be
6244 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6245 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6246 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6247 specifically allow DDmode, since that does not use the SPE SIMD
6250 2014-04-02 Richard Biener <rguenther@suse.de>
6253 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6254 MODE_INTs. Properly use negv_optab.
6255 (expand_abs): Likewise.
6257 2014-04-02 Richard Biener <rguenther@suse.de>
6260 * Makefile.in (install-driver): Guard extra installs with special
6263 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6265 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6268 2014-04-01 Richard Henderson <rth@redhat.com>
6271 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6272 alternative enabled before register allocation.
6274 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6276 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6277 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6279 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6280 (nios2_got_address): Update nios2_large_got_address call site.
6281 (nios2_delegitimize_address): New function.
6282 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6283 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6284 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6286 2014-04-01 Martin Husemann <martin@duskware.de>
6288 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6291 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6293 PR rtl-optimization/60604
6294 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6295 check from register_operand.
6296 (register_operand): Redefine in terms of general_operand.
6297 (nonmemory_operand): Use register_operand for the non-constant cases.
6299 2014-04-01 Richard Biener <rguenther@suse.de>
6301 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6303 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6305 * doc/invoke.texi (mapp-regs): Clarify.
6307 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6309 * config/i386/avx512fintrin.h (__v32hi): Define type.
6310 (__v64qi): Likewise.
6311 (_mm512_set1_epi8): Define.
6312 (_mm512_set1_epi16): Define.
6313 (_mm512_set4_epi32): Define.
6314 (_mm512_set4_epi64): Define.
6315 (_mm512_set4_pd): Define.
6316 (_mm512_set4_ps): Define.
6317 (_mm512_setr4_epi64): Define.
6318 (_mm512_setr4_epi32): Define.
6319 (_mm512_setr4_pd): Define.
6320 (_mm512_setr4_ps): Define.
6321 (_mm512_setzero_epi32): Define.
6323 2014-03-31 Martin Jambor <mjambor@suse.cz>
6326 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6327 callsite_arguments_match_p. Updated all callers. Also check types of
6328 corresponding formal parameters and actual arguments.
6329 (not_all_callers_have_enough_arguments_p) Renamed to
6330 some_callers_have_mismatched_arguments_p.
6332 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6334 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6336 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6339 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6342 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6344 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6345 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6347 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6348 Use FMAMODE_NOVF512 mode iterator.
6349 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6350 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6351 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6352 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6354 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6355 Use VF_128_256 mode iterator.
6356 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6359 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6361 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6362 static chain if needed.
6364 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6367 * lra-constraints.c (index_part_to_reg): New.
6368 (process_address): Use it.
6370 2014-03-27 Jeff Law <law@redhat.com>
6371 Jakub Jelinek <jakub@redhat.com>
6374 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6375 gen_rtx_{PLUS,MULT} to build up the address expression.
6377 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6378 creating non-canonical RTL.
6380 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6383 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6384 functions; reorganize to make cheap checks first.
6385 (inline_small_functions): Do not estimate growth when dumping;
6387 * ipa-inline.h (inline_summary): Add min_size.
6388 (growth_likely_positive): New function.
6389 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6390 (set_cond_stmt_execution_predicate): Cleanup.
6391 (estimate_edge_size_and_time): Compute min_size.
6392 (estimate_calls_size_and_time): Likewise.
6393 (estimate_node_size_and_time): Likewise.
6394 (inline_update_overall_summary): Update min_size.
6395 (do_estimate_edge_time): Likewise.
6396 (do_estimate_edge_size): Update.
6397 (do_estimate_edge_hints): Update.
6398 (growth_likely_positive): New function.
6400 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6403 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6404 also if addr has VOIDmode.
6406 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6408 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6409 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6411 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6412 instructions as well as AdvancedSIMD loads.
6414 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6416 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6417 Use crypto_aese type.
6418 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6419 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6420 crypto_aese, crypto_aesmc. Move to types.md.
6421 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6423 * config/arm/iterators.md (crypto_type): Likewise.
6425 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6427 * cgraph.c: Include expr.h and tree-dfa.h.
6428 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6431 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6434 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6435 regs from checking multi-reg pseudos.
6437 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6439 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6441 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6443 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6444 if it would clobber the stack pointer, even temporarily.
6446 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6448 * mode-switching.c: Make small adjustments to the top comment.
6450 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6452 * config/rs6000/constraints.md (wD constraint): New constraint to
6453 match the constant integer to get the top DImode/DFmode out of a
6454 vector in a VSX register.
6456 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6457 match the constant integer to get the top DImode/DFmode out of a
6458 vector in a VSX register.
6460 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6463 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6466 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6467 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6469 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6470 Optimize vec_extract of 64-bit values, where the value being
6471 extracted is in the top word, where we can use scalar
6472 instructions. Add direct move and store support. Combine the big
6473 endian/little endian vector select load support into a single insn.
6474 (vsx_extract_<mode>_internal1): Likewise.
6475 (vsx_extract_<mode>_internal2): Likewise.
6476 (vsx_extract_<mode>_load): Likewise.
6477 (vsx_extract_<mode>_store): Likewise.
6478 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6479 combined into vsx_extract_<mode>_load.
6480 (vsx_extract_<mode>_one_le): Likewise.
6482 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6483 define the top 64-bit vector element.
6485 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6488 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6489 Document vec_vbpermq builtin.
6492 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6493 enable use of xxsldwi and xxpermdi builtin functions.
6494 (vec_xxpermdi): Likewise.
6496 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6497 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6499 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6501 PR rtl-optimization/60650
6502 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6504 (find_spills_for): New.
6505 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6506 Spill all pseudos on the second iteration.
6508 2014-03-27 Marek Polacek <polacek@redhat.com>
6511 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6514 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6516 * config/s390/s390.c (s390_can_use_return_insn): Check for
6517 call-saved FPRs on 31 bit.
6519 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6522 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6523 if they need regimplification, just drop them instead of
6524 calling gimple_regimplify_operands on them.
6526 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6529 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6530 (aarch64_frame_pointer_required): Adjust logic.
6531 (aarch64_can_eliminate): Adjust logic.
6532 (aarch64_override_options_after_change): Adjust logic.
6534 2014-03-27 Dehao Chen <dehao@google.com>
6536 * ipa-inline.c (early_inliner): Update node's inline info.
6538 2014-03-26 Dehao Chen <dehao@google.com>
6540 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6541 compiler inserted conditional jumps for NAN float check.
6543 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6545 * ubsan.h (ubsan_create_data): Change second argument's type
6546 to const location_t *.
6547 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6549 (ubsan_create_data): Change second argument to const location_t *PLOC.
6550 Create Loc field whenever PLOC is non-NULL.
6551 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6552 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6556 * real.c (real_to_integer2): Change type of low to UHWI.
6558 2014-03-26 Tobias Burnus <burnus@net-b.de>
6560 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6561 (CILK_SELF_SPECS): New define.
6562 (driver_self_specs): Use it.
6564 2014-03-26 Richard Biener <rguenther@suse.de>
6566 * tree-pretty-print.c (percent_K_format): Implement special
6567 case for LTO and its stripped down BLOCK tree.
6569 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6572 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6574 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6575 one range is range_int_cst_p, but not both, at least optimize
6576 addition/subtraction of 0 and multiplication by 0 or 1.
6577 * gimple-fold.c (gimple_fold_call): Fold
6578 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6579 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6580 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6582 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6584 PR rtl-optimization/60452
6585 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6586 <case REG>: Return 1 for invalid offsets from the frame pointer.
6588 2014-03-26 Marek Polacek <polacek@redhat.com>
6591 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6594 2014-03-26 Marek Polacek <polacek@redhat.com>
6597 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6600 2014-03-26 Marek Polacek <polacek@redhat.com>
6603 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6604 multiplication in unsigned type.
6606 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6608 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6610 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6612 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6614 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6617 * cif-code.def (UNREACHABLE) New code.
6618 * ipa-inline.c (inline_small_functions): Skip edges to
6619 __builtlin_unreachable.
6620 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6621 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6622 predicate to __bulitin_unreachable.
6623 (set_cond_stmt_execution_predicate): Fix issue when
6624 invert_tree_comparison returns ERROR_MARK.
6625 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6626 propagate to inline clones.
6627 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6629 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6630 * cgraphclones.c (cgraph_clone_node): If call destination is already
6631 ureachable, do not redirect it back.
6632 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6635 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6637 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6638 Do not modify inline clones.
6640 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6642 * config/i386/i386.md (general_sext_operand): New mode attr.
6643 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6644 don't generate (sign_extend (const_int)).
6645 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6646 operands[2]. Use We constraint instead of <i> and
6647 <general_sext_operand> predicate instead of <general_operand>.
6648 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6649 * config/i386/constraints.md (We): New constraint.
6650 * config/i386/predicates.md (x86_64_sext_operand,
6651 sext_operand): New predicates.
6653 2014-03-25 Martin Jambor <mjambor@suse.cz>
6656 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6657 inconsistent devirtualizations to __builtin_unreachable.
6659 2014-03-25 Marek Polacek <polacek@redhat.com>
6662 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6664 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6666 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6667 order of elements for big-endian.
6669 2014-03-25 Richard Biener <rguenther@suse.de>
6672 * gimplify-me.c (gimple_regimplify_operands): Update the
6675 2014-03-25 Martin Jambor <mjambor@suse.cz>
6678 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6679 (lto_output_varpool_node): Likewise.
6680 (input_overwrite_node): Likewise.
6681 (input_varpool_node): Likewise.
6683 2014-03-25 Richard Biener <rguenther@suse.de>
6685 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6686 (run_gcc): Likewise.
6688 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6690 * combine.c (simplify_compare_const): Add MODE argument.
6691 Handle mode_width 0 as very large mode_width.
6692 (try_combine, simplify_comparison): Adjust callers.
6694 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6695 type to avoid signed integer overflow.
6696 * explow.c (plus_constant): Likewise.
6698 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6700 * doc/generic.texi: Correct typos.
6702 2014-03-24 Tobias Burnus <burnus@net-b.de>
6704 * doc/invoke.texi (-flto): Expand section about
6705 using static libraries with LTO.
6707 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6709 PR rtl-optimization/60501
6710 * optabs.def (addptr3_optab): New optab.
6711 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6712 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6713 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6715 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6717 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6719 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6721 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6724 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6725 (_mm256_undefined_ps): Define.
6726 (_mm256_undefined_pd): Define.
6727 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6728 (_mm_undefined_pd): Define.
6729 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6730 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6731 (_mm512_undefined_ps): Define.
6732 (_mm512_undefined_pd): Define.
6733 Use _mm*_undefined_*.
6734 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6736 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6738 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6739 (lshr_simd): DI mode added.
6740 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6741 (aarch64_ushr_simddi): Likewise.
6742 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6743 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6744 (vshrd_n_u64): Likewise.
6746 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6748 * Makefile.in (s-macro_list): Depend on cc1.
6750 2014-03-23 Teresa Johnson <tejohnson@google.com>
6752 * ipa-utils.c (ipa_print_order): Use specified dump file.
6754 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6756 PR rtl-optimization/60601
6757 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6759 * gcc.c (eval_spec_function): Initialize save_growing_value.
6761 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6764 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6765 code == MINUS_EXPR, never swap op0 with op1.
6767 * toplev.c (init_local_tick): Avoid signed integer multiplication
6769 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6770 shift by first operand's bitsize.
6772 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6775 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6777 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6778 TARGET_ISA_64BIT_P(x).
6780 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6782 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6783 pattern for vector nor instead of subtract from splat(-1).
6784 (altivec_expand_vec_perm_const_le): Likewise.
6786 2014-03-21 Richard Henderson <rth@twiddle.net>
6789 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6790 related insns after epilogue_completed.
6792 2014-03-21 Martin Jambor <mjambor@suse.cz>
6795 * cgraph.h (symtab_node): New flag body_removed.
6796 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6797 when removing bodies.
6798 * symtab.c (dump_symtab_base): Dump body_removed flag.
6799 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6800 had their bodies removed.
6802 2014-03-21 Martin Jambor <mjambor@suse.cz>
6805 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6808 2014-03-21 Richard Biener <rguenther@suse.de>
6810 PR tree-optimization/60577
6811 * tree-core.h (struct tree_base): Document nothrow_flag use
6812 in VAR_DECL_NONALIASED.
6813 * tree.h (VAR_DECL_NONALIASED): New.
6814 (may_be_aliased): Adjust.
6815 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6817 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6819 * expr.c (expand_expr_real_1): Remove outdated comment.
6821 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6824 * ira.c (adjust_cleared_regs): Call copy_rtx on
6825 *reg_equiv[REGNO (loc)].src_p before passing it to
6826 simplify_replace_fn_rtx.
6829 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6830 into CONST, put pic register as first operand of PLUS. Use
6831 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6833 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6835 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6837 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6839 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6840 around for store forwarding issue in the FPU on the UT699.
6841 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6842 loads and operations if -mfix-ut699 is specified.
6843 (divtf3_hq): Tweak attribute.
6844 (sqrttf2_hq): Likewise.
6846 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6848 * calls.c (store_one_arg): Remove incorrect const qualification on the
6849 type of the temporary.
6850 * cfgexpand.c (expand_return): Likewise.
6851 * expr.c (expand_constructor): Likewise.
6852 (expand_expr_real_1): Likewise.
6854 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6856 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6859 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6862 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6864 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6866 * config/arm/aarch-common-protos.h
6867 (alu_cost_table): Fix spelling of "extend".
6868 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6870 2014-03-19 Richard Biener <rguenther@suse.de>
6873 * tree-core.h (tree_type_common): Re-order pointer members
6874 to reduce recursion depth during GC walks.
6876 2014-03-19 Marek Polacek <polacek@redhat.com>
6879 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6880 before accessing it.
6882 2014-03-19 Richard Biener <rguenther@suse.de>
6885 * lto-streamer-in.c (input_function): In WPA stage do not drop
6888 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6890 PR tree-optimization/60559
6891 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6892 with build_zero_cst assignment.
6894 2014-03-18 Kai Tietz <ktietz@redhat.com>
6896 PR rtl-optimization/56356
6897 * sdbout.c (sdbout_parms): Verify that parms'
6898 incoming argument is valid.
6899 (sdbout_reg_parms): Likewise.
6901 2014-03-18 Richard Henderson <rth@redhat.com>
6904 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6905 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6906 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6908 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6910 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6911 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6912 Italicize plugin event names in description. Explain that
6913 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6914 Remind that no GCC functions should be called after PLUGIN_FINISH.
6915 Explain what pragmas with expansion are.
6917 2014-03-18 Martin Liska <mliska@suse.cz>
6919 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6920 gimple call statement is update.
6921 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6922 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6924 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6927 * ubsan.c (ubsan_instrument_unreachable): Call
6928 initialize_sanitizer_builtins.
6929 (ubsan_pass): Likewise.
6932 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6933 varpool_finalize_decl instead of rest_of_decl_compilation.
6935 2014-03-18 Richard Biener <rguenther@suse.de>
6937 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6938 by using bitmap_and_compl instead of bitmap_and_compl_into.
6939 (df_rd_transfer_function): Likewise.
6941 2014-03-18 Richard Biener <rguenther@suse.de>
6943 * doc/lto.texi (fresolution): Fix typo.
6945 2014-03-18 Richard Biener <rguenther@suse.de>
6947 * doc/invoke.texi (flto): Update for changes in 4.9.
6949 2014-03-18 Richard Biener <rguenther@suse.de>
6951 * doc/loop.texi: Remove section on the removed lambda framework.
6952 Update loop docs with recent changes in preserving loop structure.
6954 2014-03-18 Richard Biener <rguenther@suse.de>
6956 * doc/lto.texi (-fresolution): Document.
6958 2014-03-18 Richard Biener <rguenther@suse.de>
6960 * doc/contrib.texi: Adjust my name.
6962 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6965 * internal-fn.c: Include diagnostic-core.h.
6966 (expand_BUILTIN_EXPECT): New function.
6967 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6968 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6969 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6970 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6972 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6973 Revert 3 argument __builtin_expect code.
6974 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6975 * gimple-fold.c (gimple_fold_call): Likewise.
6976 * tree.h (fold_builtin_expect): New prototype.
6977 * builtins.c (build_builtin_expect_predicate): Add predictor
6978 argument, if non-NULL, create 3 argument __builtin_expect.
6979 (fold_builtin_expect): No longer static. Add ARG2 argument,
6980 pass it through to build_builtin_expect_predicate.
6981 (fold_builtin_2): Adjust caller.
6982 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6983 * internal-fn.def (BUILTIN_EXPECT): New.
6985 2014-03-18 Tobias Burnus <burnus@net-b.de>
6988 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6989 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6990 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6992 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6995 * predict.c (combine_predictions_for_bb): Fix up formatting.
6996 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6997 fill what it points to if non-NULL.
6998 (tree_predict_by_opcode): Adjust caller, use the predictor.
6999 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7001 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7003 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7004 proper constant for the store mode.
7006 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7008 * symtab.c (change_decl_assembler_name): Fix transparent alias
7011 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7013 * config/aarch64/aarch64.c: Correct the comments about the
7014 aarch64 stack layout.
7016 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7018 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7019 check for GF_OMP_FOR_KIND_FOR.
7021 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7023 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7024 ymm and zmm register names.
7026 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7029 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7030 note creation for the 2010-08-31 changes.
7032 2014-03-17 Marek Polacek <polacek@redhat.com>
7035 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7036 as -fno-tree-loop-vectorize.
7037 (expand_omp_simd): Likewise.
7039 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7041 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7042 (eligible_for_call_delay): New prototype.
7043 * config/sparc/sparc.c (tls_call_delay): Rename into...
7044 (eligible_for_call_delay): ...this. Return false if the instruction
7045 cannot be put in the delay slot of a branch.
7046 (eligible_for_restore_insn): Simplify.
7047 (eligible_for_return_delay): Return false if the instruction cannot be
7048 put in the delay slot of a branch and simplify.
7049 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7050 put in the delay slot of a branch.
7051 * config/sparc/sparc.md (fix_ut699): New attribute.
7052 (tls_call_delay): Delete.
7053 (in_call_delay): Reimplement.
7054 (eligible_for_sibcall_delay): Rename into...
7055 (in_sibcall_delay): ...this.
7056 (eligible_for_return_delay): Rename into...
7057 (in_return_delay): ...this.
7058 (in_branch_delay): Reimplement.
7059 (in_uncond_branch_delay): Delete.
7060 (in_annul_branch_delay): Delete.
7062 2014-03-14 Richard Henderson <rth@redhat.com>
7065 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7066 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7067 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7068 (floathi splitters): Remove.
7069 (float<SWI48x>xf2): New pattern.
7070 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7071 code that tried to handle DImode for 32-bit, but which was excluded
7072 by the pattern's condition. Drop allocation of stack temporary.
7073 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7074 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7075 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7076 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7077 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7078 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7079 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7080 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7081 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7082 (*float<SWI48x><X87MODEF>2_i387): Remove.
7083 (all float _with_temp splitters): Remove.
7084 (*float<SWI48x><MODEF>2_i387): New pattern.
7085 (*float<SWI48><MODEF>2_sse): New pattern.
7086 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7087 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7089 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7090 Marek Polacek <polacek@redhat.com>
7093 * common.opt (dump_base_name_prefixed): New Variable.
7094 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7095 if x_dump_base_name_prefixed is already set, set it at the end.
7097 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7099 PR rtl-optimization/60508
7100 * lra-constraints.c (get_reload_reg): Add new parameter
7102 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7103 Pass the new parameter values.
7105 2014-03-14 Richard Biener <rguenther@suse.de>
7107 * common.opt: Revert unintented changes from r205065.
7110 2014-03-14 Richard Biener <rguenther@suse.de>
7113 * cfghooks.c (split_block): Properly adjust all loops the
7114 block was a latch of.
7116 2014-03-14 Martin Jambor <mjambor@suse.cz>
7119 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7122 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7125 * config/avr/avr.c (avr_set_current_function): Pass function name
7126 through default_strip_name_encoding before sanity checking instead
7127 of skipping the first char of the assembler name.
7129 2014-03-13 Richard Henderson <rth@redhat.com>
7132 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7133 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7134 * config/i386/i386-protos.h: Likewise.
7135 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7136 in the expander instead of a splitter.
7137 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7138 any possibility of requiring a memory.
7139 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7140 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7141 (fp branch splitters): Update for ix86_split_fp_branch.
7142 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7143 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7144 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7145 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7146 (*fop_<MODEF>_3_i387): Likewise.
7147 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7148 (splitters for the fop_* register patterns): Remove.
7149 (fscalexf4_i387): Rename from *fscalexf4_i387.
7150 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7152 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7154 PR tree-optimization/59779
7155 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7156 type for bitsize and maxsize instead of HOST_WIDE_INT.
7158 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7160 PR rtl-optimization/57320
7161 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7162 the CFG after thread_prologue_and_epilogue_insns.
7164 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7166 PR rtl-optimization/57189
7167 * lra-constraints.c (process_alt_operands): Disfavor spilling
7170 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7172 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7174 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7176 PR tree-optimization/59025
7178 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7179 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7181 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7184 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7185 calls of avr_out_plus_1.
7187 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7189 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7190 BB's single pred and update the father loop's latch info later.
7192 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7194 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7198 (VEC_base): Likewise.
7199 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7200 registers, we need to swap double words in little endian mode.
7202 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7203 to be a container mode for 128-bit integer operations added in ISA
7204 2.07. Unlike TImode and PTImode, the preferred register set is
7205 the Altivec/VMX registers for the 128-bit operations.
7207 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7209 (rs6000_split_128bit_ok_p): Likewise.
7211 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7212 macros for creating ISA 2.07 normal and overloaded builtin
7213 functions with 3 arguments.
7214 (BU_P8V_OVERLOAD_3): Likewise.
7215 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7216 for use as overloaded functions.
7217 (VPERM_1TI_UNS): Likewise.
7218 (VSEL_1TI): Likewise.
7219 (VSEL_1TI_UNS): Likewise.
7220 (ST_INTERNAL_1ti): Likewise.
7221 (LD_INTERNAL_1ti): Likewise.
7222 (XXSEL_1TI): Likewise.
7223 (XXSEL_1TI_UNS): Likewise.
7224 (VPERM_1TI): Likewise.
7225 (VPERM_1TI_UNS): Likewise.
7226 (XXPERMDI_1TI): Likewise.
7227 (SET_1TI): Likewise.
7228 (LXVD2X_V1TI): Likewise.
7229 (STXVD2X_V1TI): Likewise.
7230 (VEC_INIT_V1TI): Likewise.
7231 (VEC_SET_V1TI): Likewise.
7232 (VEC_EXT_V1TI): Likewise.
7233 (EQV_V1TI): Likewise.
7234 (NAND_V1TI): Likewise.
7235 (ORC_V1TI): Likewise.
7236 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7237 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7239 (VADDUQM): Likewise.
7240 (VSUBCUQ): Likewise.
7241 (VADDEUQM): Likewise.
7242 (VADDECUQ): Likewise.
7243 (VSUBEUQM): Likewise.
7244 (VSUBECUQ): Likewise.
7246 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7247 __int128_t and __uint128_t types.
7248 (__uint128_type): Likewise.
7249 (altivec_categorize_keyword): Add support for vector __int128_t,
7250 vector __uint128_t, vector __int128, and vector unsigned __int128
7251 as a container type for TImode operations that need to be done in
7252 VSX/Altivec registers.
7253 (rs6000_macro_to_expand): Likewise.
7254 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7255 to support 128-bit integer instructions vaddcuq, vadduqm,
7256 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7257 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7259 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7260 for V1TImode, and set up preferences to use VSX/Altivec registers.
7261 Setup VSX reload handlers.
7262 (rs6000_debug_reg_global): Likewise.
7263 (rs6000_init_hard_regno_mode_ok): Likewise.
7264 (rs6000_preferred_simd_mode): Likewise.
7265 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7266 (easy_altivec_constant): Likewise.
7267 (output_vec_const_move): Likewise.
7268 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7270 (rs6000_expand_vector_extract): Likewise.
7271 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7273 (rs6000_const_vec): Add support for V1TImode.
7274 (rs6000_emit_le_vsx_load): Swap double words when loading or
7275 storing TImode/V1TImode.
7276 (rs6000_emit_le_vsx_store): Likewise.
7277 (rs6000_emit_le_vsx_move): Likewise.
7278 (rs6000_emit_move): Add support for V1TImode.
7279 (altivec_expand_ld_builtin): Likewise.
7280 (altivec_expand_st_builtin): Likewise.
7281 (altivec_expand_vec_init_builtin): Likewise.
7282 (altivec_expand_builtin): Likewise.
7283 (rs6000_init_builtins): Add support for V1TImode type. Add
7284 support for ISA 2.07 128-bit integer builtins. Define type names
7285 for the VSX/Altivec vector types.
7286 (altivec_init_builtins): Add support for overloaded vector
7287 functions with V1TImode type.
7288 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7289 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7291 (rs6000_split_128bit_ok_p): Likewise.
7292 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7293 __int128_t and vector __uint128_t.
7295 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7296 and mode attributes.
7303 (VS_scalar): Likewise.
7304 (VS_double): Likewise.
7305 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7307 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7308 we support the ISA 2.07 128-bit integer arithmetic instructions.
7309 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7310 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7311 and TImode types for use with the builtin functions.
7312 (V1TI_type_node): Likewise.
7313 (unsigned_V1TI_type_node): Likewise.
7314 (intTI_type_internal_node): Likewise.
7315 (uintTI_type_internal_node): Likewise.
7317 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7318 128-bit builtin functions.
7319 (UNSPEC_VADDEUQM): Likewise.
7320 (UNSPEC_VADDECUQ): Likewise.
7321 (UNSPEC_VSUBCUQ): Likewise.
7322 (UNSPEC_VSUBEUQM): Likewise.
7323 (UNSPEC_VSUBECUQ): Likewise.
7324 (VM): Add V1TImode to vector mode iterators.
7326 (VI_unit): Likewise.
7327 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7328 (altivec_vaddcuq): Likewise.
7329 (altivec_vsubuqm): Likewise.
7330 (altivec_vsubcuq): Likewise.
7331 (altivec_vaddeuqm): Likewise.
7332 (altivec_vaddecuq): Likewise.
7333 (altivec_vsubeuqm): Likewise.
7334 (altivec_vsubecuq): Likewise.
7336 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7338 (BOOL_128): Likewise.
7339 (BOOL_REGS_OUTPUT): Likewise.
7340 (BOOL_REGS_OP1): Likewise.
7341 (BOOL_REGS_OP2): Likewise.
7342 (BOOL_REGS_UNARY): Likewise.
7343 (BOOL_REGS_AND_CR0): Likewise.
7345 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7346 128-bit integer builtin support.
7347 (vec_vadduqm): Likewise.
7348 (vec_vaddecuq): Likewise.
7349 (vec_vaddeuqm): Likewise.
7350 (vec_vsubecuq): Likewise.
7351 (vec_vsubeuqm): Likewise.
7352 (vec_vsubcuq): Likewise.
7353 (vec_vsubuqm): Likewise.
7355 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7356 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7357 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7358 128-bit integer add/subtract to ISA 2.07.
7360 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7362 * config/arc/arc.c (arc_predicate_delay_insns):
7363 Fix third argument passed to conditionalize_nonjump.
7365 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7367 * config/aarch64/aarch64-builtins.c
7368 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7369 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7370 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7371 instead of __builtin_lfloor.
7372 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7374 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7376 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7377 (tree_ssa_ifcombine_bb_1): New function.
7378 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7379 is an empty forwarder block to then_bb or vice versa and then_bb
7380 and else_bb are effectively swapped.
7382 2014-03-12 Christian Bruel <christian.bruel@st.com>
7385 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7386 REG_CFA_DEF_CFA note.
7387 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7388 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7390 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7392 PR tree-optimization/60454
7393 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7395 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7397 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7398 Do not define target_cpu_default2 to generic.
7399 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7400 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7401 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7403 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7404 Marc Glisse <marc.glisse@inria.fr>
7406 PR tree-optimization/60502
7407 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7408 instead of build_low_bits_mask.
7410 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7413 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7414 if there are multiple uses, but op doesn't live on E edge.
7415 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7416 clobber stmts before __builtin_unreachable.
7418 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7420 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7421 hard_frame_pointer_rtx.
7422 * cse.c (cse_insn): Remove volatile check.
7423 * cselib.c (cselib_process_insn): Likewise.
7424 * dse.c (scan_insn): Likewise.
7426 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7428 * config/arc/arc.c (conditionalize_nonjump): New function,
7430 (arc_ifcvt): ... this.
7431 (arc_predicate_delay_insns): Use it.
7433 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7435 * config/arc/predicates.md (extend_operand): During/after reload,
7436 allow const_int_operand.
7437 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7438 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7439 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7441 (umulsi3_highpart_i): Likewise.
7443 2014-03-11 Richard Biener <rguenther@suse.de>
7445 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7446 Add asserts to guard possible wrong-code bugs.
7448 2014-03-11 Richard Biener <rguenther@suse.de>
7450 PR tree-optimization/60429
7451 PR tree-optimization/60485
7452 * tree-ssa-structalias.c (set_union_with_increment): Properly
7453 take into account all fields that overlap the shifted vars.
7454 (do_sd_constraint): Likewise.
7455 (do_ds_constraint): Likewise.
7456 (get_constraint_for_ptr_offset): Likewise.
7458 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7460 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7461 (nios2_compute_frame_layout):
7462 Add calculation of cfun->machine->fp_save_offset.
7463 (nios2_expand_prologue): Correct setting of frame pointer register
7465 (nios2_expand_epilogue): Update recovery of stack pointer from
7466 frame pointer accordingly.
7467 (nios2_initial_elimination_offset): Update calculation of offset
7468 for eliminating to HARD_FRAME_POINTER_REGNUM.
7470 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7473 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7474 cgraph_get_create_node on VAR_DECLs.
7476 2014-03-10 Richard Biener <rguenther@suse.de>
7479 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7481 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7483 * config/vms/vms.opt (vms_float_format): New variable.
7485 2014-03-08 Tobias Burnus <burnus@net-b.de>
7487 * doc/invoke.texi (-fcilkplus): Update implementation status.
7489 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7490 Richard Biener <rguenther@suse.de>
7492 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7493 consistently accross all TUs.
7494 (run_gcc): Enable -fshort-double automatically at link at link-time
7495 and disallow override.
7497 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7500 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7501 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7502 if they can't be used.
7504 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7506 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7507 for Solaris 11/x86 ld.
7508 * configure: Regenerate.
7510 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7512 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7513 (LIB_TLS_SPEC): Save as ld_tls_libs.
7514 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7515 (HAVE_AS_IX86_TLSLDM): New test.
7516 * configure, config.in: Regenerate.
7517 * config/i386/i386.c (legitimize_tls_address): Fall back to
7518 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7519 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7520 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7521 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7523 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7525 * common.opt (fira-loop-pressure): Mark as optimization.
7527 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7529 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7530 an OpenMP mappable type.
7532 2014-03-06 Matthias Klose <doko@ubuntu.com>
7534 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7535 MULTILIB_OSDIRNAMES is not defined.
7537 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7538 Meador Inge <meadori@codesourcery.com>
7541 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7542 (arm_legitimize_address): Call legitimize_tls_address for any
7543 arm_tls_referenced_p expression, handle constant addend. Call it
7544 before testing for !TARGET_ARM.
7545 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7547 2014-03-06 Richard Biener <rguenther@suse.de>
7553 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7555 * tree-streamer.c (record_common_node): Assert we don't record
7556 nodes with type double.
7557 (preload_common_node): Skip type double, complex double and double
7558 pointer since it is now frontend dependent due to fshort-double option.
7560 2014-03-06 Richard Biener <rguenther@suse.de>
7562 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7563 or -fno-lto is specified and the linker has full plugin support.
7564 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7565 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7566 * lto-wrapper.c (merge_and_complain): Merge compile-time
7567 optimization levels.
7568 (run_gcc): And pass it through to the link options.
7570 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7574 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7576 * cselib.c (remove_useless_values): Skip to avoid quadratic
7577 behavior if the condition moved from...
7578 (cselib_process_insn): ... here holds.
7580 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7583 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7584 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7587 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7588 (TM_H): Add x86-tune.def.
7590 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7592 * config/aarch64/aarch64.c (generic_tunings):
7593 Use cortexa57_extra_costs.
7595 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7598 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7599 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7600 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7603 2014-03-04 Heiher <r@hev.cc>
7605 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7606 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7608 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7610 * config/i386/predicates.md (const2356_operand): Change to ...
7611 (const2367_operand): ... this.
7612 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7614 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7615 (*avx512pf_scatterpf<mode>sf): Ditto.
7616 (avx512pf_scatterpf<mode>df): Ditto.
7617 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7618 (*avx512pf_scatterpf<mode>df): Ditto.
7619 * config/i386/i386.c (ix86_expand_builtin): Update
7620 incorrect hint operand error message.
7622 2014-03-04 Richard Biener <rguenther@suse.de>
7624 * lto-section-in.c (lto_get_section_data): Fix const cast.
7626 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7628 * tree-streamer.c (record_common_node): Assert we don't record
7629 nodes with type double.
7630 (preload_common_node): Skip type double, complex double and double
7631 pointer since it is now frontend dependent due to fshort-double option.
7633 2014-03-04 Richard Biener <rguenther@suse.de>
7636 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7637 (lto_input_toplevel_asms): Likewise.
7638 * lto-section-in.c (lto_get_section_data): Instead do it here
7641 2014-03-04 Richard Biener <rguenther@suse.de>
7643 PR tree-optimization/60382
7644 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7645 dead PHIs a reduction.
7647 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7649 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7651 (_mm_prefetch): Move out of GCC target("sse") pragma.
7652 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7653 GCC target("prfchw") pragma.
7654 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7656 * config/i386/i386.c (ix86_option_override_internal): Enable
7657 -mprfchw with -mprefetchwt1.
7659 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7661 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7664 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7666 * opts.h (CL_PCH_IGNORE): Define.
7667 * targhooks.c (option_affects_pch_p):
7668 Return false for options that have CL_PCH_IGNORE set.
7669 * opt-functions.awk: Process PchIgnore.
7670 * doc/options.texi: Document PchIgnore.
7672 * config/arc/arc.opt (misize): Add PchIgnore property.
7674 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7676 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7677 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7678 constraint on constants to permit them being loaded into
7679 GENERAL_REGS or BASE_REGS.
7681 2014-03-03 Nick Clifton <nickc@redhat.com>
7683 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7684 anti-cacnonical alternatives.
7685 (negandhi3_real): New pattern.
7686 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7688 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7690 * config/avr/avr-mcus.def: Remove atxmega16x1.
7691 * config/avr/avr-tables.opt: Regenerate.
7692 * config/avr/t-multilib: Regenerate.
7693 * doc/avr-mmcu.texi: Regenerate.
7695 2014-03-03 Tobias Grosser <tobias@grosser.es>
7696 Mircea Namolaru <mircea.namolaru@inria.fr>
7698 PR tree-optimization/58028
7699 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7702 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7704 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7705 not handled by recognizers.
7707 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7710 * function.c (expand_function_end): Don't emit
7711 clobber_return_register sequence if clobber_after is a BARRIER.
7712 * cfgexpand.c (construct_exit_block): Append instructions before
7713 return_label to prev_bb.
7715 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7717 * config/rs6000/constraints.md: Document reserved use of "wc".
7719 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7722 * ipa.c (function_and_variable_visibility): When dissolving comdat
7723 group, also set all symbols to local.
7725 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7730 2013-12-14 Jan Hubicka <jh@suse.cz>
7732 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7734 2014-03-02 Jon Beniston <jon@beniston.com>
7740 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7741 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7742 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7743 (simple_return, *simple_return): New patterns
7744 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7745 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7747 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7749 * dwarf2out.c (gen_subprogram_die): Tidy.
7751 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7754 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7755 (*mov_t_msb_neg_negc): ... this new insn.
7757 2014-02-28 Jason Merrill <jason@redhat.com>
7760 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7763 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7766 * dwarf2out.c (decltype_auto_die): New static.
7767 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7768 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7769 (is_cxx_auto): Likewise.
7771 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7773 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7774 we are not using general regs only.
7776 2014-02-28 Richard Biener <rguenther@suse.de>
7779 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7780 previous fix and only allow to remove trivial pre-headers
7781 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7782 (remove_forwarder_block): Properly update the latch of a loop.
7784 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7787 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7788 (cselib_preserved_hash_table): New.
7789 (preserve_constants_and_equivs): Move preserved vals to it.
7790 (cselib_find_slot): Look it up first.
7791 (cselib_init): Initialize it.
7792 (cselib_finish): Release it.
7793 (dump_cselib_table): Dump it.
7795 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7798 * cselib.c (remove_useless_values): Skip to avoid quadratic
7799 behavior if the condition moved from...
7800 (cselib_process_insn): ... here holds.
7802 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7805 * var-tracking.c (vt_initialize): Apply the same condition to
7806 preserve the CFA base value.
7808 2014-02-28 Joey Ye <joey.ye@arm.com>
7811 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7812 if reload in progress or completed.
7814 2014-02-28 Tobias Burnus <burnus@net-b.de>
7817 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7820 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7822 * doc/tm.texi.in (Condition Code Status): Update documention for
7823 relative locations of cc0-setter and cc0-user.
7825 2014-02-27 Jeff Law <law@redhat.com>
7827 PR rtl-optimization/52714
7828 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7829 into two independent simple sets, if I3 is a jump, ensure the
7830 pattern we place into I3 is a (set (pc) ...).
7832 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7833 Jeff Law <law@redhat.com>
7835 PR rtl-optimization/49847
7836 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7837 are in different blocks.
7838 * doc/tm.texi (Condition Code Status): Update documention for
7839 relative locations of cc0-setter and cc0-user.
7841 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7844 * lra.c (lra_emit_add): Check SUBREG too.
7846 2014-02-27 Andreas Schwab <schwab@suse.de>
7848 * config/m68k/m68k.c (m68k_option_override): Disable
7849 -flive-range-shrinkage for classic m68k.
7850 (m68k_override_options_after_change): Likewise.
7852 2014-02-27 Marek Polacek <polacek@redhat.com>
7855 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7856 -Wmaybe-uninitialized.
7858 2014-02-27 Alan Modra <amodra@gmail.com>
7861 * reload1.c (emit_input_reload_insns): When reload_override_in,
7862 set old to rl->in_reg when rl->in_reg is a subreg.
7864 2014-02-26 Richard Biener <rguenther@suse.de>
7867 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7869 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7871 * common/config/i386/predicates.md (const1256_operand): Remove.
7872 (const2356_operand): New.
7873 (const_1_to_2_operand): Remove.
7874 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7875 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7876 (*avx512pf_gatherpf<mode>sf): Ditto.
7877 (avx512pf_gatherpf<mode>df): Ditto.
7878 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7879 (*avx512pf_gatherpf<mode>df): Ditto.
7880 (avx512pf_scatterpf<mode>sf): Ditto.
7881 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7882 (*avx512pf_scatterpf<mode>sf): Ditto.
7883 (avx512pf_scatterpf<mode>df): Ditto.
7884 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7885 (*avx512pf_scatterpf<mode>df): Ditto.
7886 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7888 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7890 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7891 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7892 (_mm512_mask_testn_epi64_mask): Move to ...
7893 * config/i386/avx512cdintrin.h: Here.
7894 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7895 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7896 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7897 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7898 TARGET_AVX512F from TARGET_AVX512CD.
7900 2014-02-26 Richard Biener <rguenther@suse.de>
7903 * ipa.c (walk_polymorphic_call_targets): Properly guard
7904 call to inline_update_overall_summary.
7906 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7909 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7910 and latches only if requested. Fix latch if it is removed.
7911 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7912 LOOPS_HAVE_PREHEADERS.
7914 2014-02-25 Andrew Pinski <apinski@cavium.com>
7916 * builtins.c (expand_builtin_thread_pointer): Create a new target
7917 when the target is NULL.
7919 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7921 PR rtl-optimization/60317
7922 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7923 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7924 * lra-assigns.c: Include params.h.
7925 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7926 other reload pseudos considerations.
7928 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7930 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7931 to use canonical form for nor<mode>3.
7933 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7936 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7939 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7941 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7942 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7943 (ix86_handle_option): Handle OPT_mprefetchwt1.
7944 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7945 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7947 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7948 OPTION_MASK_ISA_PREFETCHWT1.
7949 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7950 (PTA_PREFETCHWT1): New.
7951 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7952 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7953 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7954 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7955 (*prefetch_avx512pf_<mode>_: Change into ...
7956 (*prefetch_prefetchwt1_<mode>: This.
7957 * config/i386/i386.opt (mprefetchwt1): New.
7958 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7959 (_mm_prefetch): Handle intent to write.
7960 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7962 2014-02-25 Richard Biener <rguenther@suse.de>
7965 * emit-rtl.c (mem_attrs_htab): Remove.
7966 (mem_attrs_htab_hash): Likewise.
7967 (mem_attrs_htab_eq): Likewise.
7968 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7969 (init_emit_once): Do not allocate mem_attrs_htab.
7971 2014-02-25 Richard Biener <rguenther@suse.de>
7974 * lto-opts.c (lto_write_options): Output non-explicit conservative
7975 -fwrapv, -fno-trapv and -fno-strict-overflow.
7976 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7977 (run_gcc): And pass them through.
7979 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7981 * sel-sched.c (calculate_new_fences): New parameter ptime.
7982 Calculate it as a maximum over all fence cycles.
7983 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7984 Print the final schedule timing when sched_verbose.
7986 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7988 PR rtl-optimization/60292
7989 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7990 bit fot the fence instruction.
7992 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7994 * calls.h: Fix typo in comment.
7996 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7998 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7999 adjusting offsetable addresses.
8001 2014-02-24 Guozhi Wei <carrot@google.com>
8003 * sparseset.h (sparseset_pop): Fix the wrong index.
8005 2014-02-24 Walter Lee <walt@tilera.com>
8007 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8008 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8010 * common/config/tilegx/tilegx-common.c
8011 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8012 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8014 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8015 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8016 (tilegx_gimplify_va_arg_expr): Handle big endian.
8017 (tilegx_expand_unaligned_load): Ditto.
8018 (tilegx_expand_unaligned_store): Ditto.
8019 (TARGET_RETURN_IN_MSB): New.
8020 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8021 (TARGET_ENDIAN_DEFAULT): New.
8022 (TARGET_BIG_ENDIAN): Handle big endian.
8023 (BYTES_BIG_ENDIAN): Ditto.
8024 (WORDS_BIG_ENDIAN): Ditto.
8025 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8028 * config/tilegx/tilegx.md (extv): Handle big endian.
8030 (insn_st<n>): Ditto.
8031 (insn_st<n>_add<bitsuffix>): Ditto.
8032 (insn_stnt<n>): Ditto.
8033 (insn_stnt<n>_add<bitsuffix>):Ditto.
8034 (vec_interleave_highv8qi): Handle big endian.
8035 (vec_interleave_highv8qi_be): New.
8036 (vec_interleave_highv8qi_le): New.
8037 (insn_v1int_h): Handle big endian.
8038 (vec_interleave_lowv8qi): Handle big endian.
8039 (vec_interleave_lowv8qi_be): New.
8040 (vec_interleave_lowv8qi_le): New.
8041 (insn_v1int_l): Handle big endian.
8042 (vec_interleave_highv4hi): Handle big endian.
8043 (vec_interleave_highv4hi_be): New.
8044 (vec_interleave_highv4hi_le): New.
8045 (insn_v2int_h): Handle big endian.
8046 (vec_interleave_lowv4hi): Handle big endian.
8047 (vec_interleave_lowv4hi_be): New.
8048 (vec_interleave_lowv4hi_le): New.
8049 (insn_v2int_l): Handle big endian.
8050 (vec_interleave_highv2si): Handle big endian.
8051 (vec_interleave_highv2si_be): New.
8052 (vec_interleave_highv2si_le): New.
8053 (insn_v4int_h): Handle big endian.
8054 (vec_interleave_lowv2si): Handle big endian.
8055 (vec_interleave_lowv2si_be): New.
8056 (vec_interleave_lowv2si_le): New.
8057 (insn_v4int_l): Handle big endian.
8058 * config/tilegx/tilegx.opt (mbig-endian): New option.
8059 (mlittle-endian): New option.
8060 * doc/install.texi: Document tilegxbe-linux.
8061 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8063 2014-02-24 Martin Jambor <mjambor@suse.cz>
8066 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8067 there are no parameter descriptors.
8069 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8071 PR rtl-optimization/60268
8072 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8073 initialization to ...
8074 (sched_rgn_init): ... here.
8075 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8077 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8079 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8082 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8084 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8087 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8089 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8090 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8092 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8094 * config/microblaze/predicates.md: Add cmp_op predicate.
8095 * config/microblaze/microblaze.md: Add branch_compare instruction
8096 which uses cmp_op predicate and emits cmp insn before branch.
8097 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8098 to microblaze_expand_conditional_branch and consolidate logic.
8099 (microblaze_expand_conditional_branch): emit branch_compare
8100 insn instead of handling cmp op separate from branch insn.
8102 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8104 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8107 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8109 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8110 define_insn with define_expand and new define_insn
8111 *altivec_lve<VI_char>x_internal.
8112 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8113 and new define_insn *altivec_stve<VI_char>x_internal.
8114 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8116 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8118 (altivec_expand_stvex_be): New function.
8120 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8122 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8123 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8124 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8125 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8127 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8130 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8131 instead of emit_move_insn.
8133 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8135 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8137 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8138 gen_altivec_vsumsws.
8140 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8142 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8143 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8144 (altivec_lvxl_<mode>): New define_expand incorporating
8145 -maltivec=be semantics where needed.
8146 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8147 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8148 semantics where needed.
8149 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8150 (altivec_stvx_<mode>): New define_expand incorporating
8151 -maltivec=be semantics where needed.
8152 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8153 VM2 iterator instead of V4SI.
8154 (altivec_stvxl_<mode>): New define_expand incorporating
8155 -maltivec=be semantics where needed.
8156 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8157 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8158 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8159 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8160 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8161 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8162 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8163 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8164 ALTIVEC_BUILTIN_STVXL.
8165 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8166 (altivec_expand_stvx_be): Likewise.
8167 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8168 (altivec_expand_lvx_be): Likewise.
8169 (altivec_expand_stvx_be): Likewise.
8170 (altivec_expand_builtin): Add cases for
8171 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8172 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8173 (altivec_init_builtins): Add definitions for
8174 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8175 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8177 2014-02-21 Catherine Moore <clm@codesourcery.com>
8179 * doc/invoke.texi (mvirt, mno-virt): Document.
8180 * config/mips/mips.opt (mvirt): New option.
8181 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8183 2014-02-21 Richard Biener <rguenther@suse.de>
8185 PR tree-optimization/60276
8186 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8187 (STMT_VINFO_MIN_NEG_DIST): New macro.
8188 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8189 STMT_VINFO_MIN_NEG_DIST.
8190 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8191 made for negative dependence distances still hold.
8193 2014-02-21 Richard Biener <rguenther@suse.de>
8196 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8197 DECL_INITIAL for globals not in the current function context.
8199 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8201 PR tree-optimization/56490
8202 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8203 * tree-ssa-uninit.c: Include params.h.
8204 (compute_control_dep_chain): Add num_calls argument, return false
8205 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8206 num_calls to recursive call.
8207 (find_predicates): Change dep_chain into normal array,
8208 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8209 variable and adjust compute_control_dep_chain caller.
8210 (find_def_preds): Likewise.
8212 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8214 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8215 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8217 2014-02-21 Nick Clifton <nickc@redhat.com>
8219 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8220 (pushhi1): Likewise.
8221 (popqi1): Add mode to pre_dec.
8224 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8226 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8227 mode for mask of V8SFmode permutation.
8229 2014-02-20 Richard Henderson <rth@redhat.com>
8232 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8233 a new pseudo for OLDVAL.
8235 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8238 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8239 gen_reg_rtx if d->testing_p.
8240 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8241 if d->testing_p and we will certainly return true.
8242 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8245 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8247 * emit-rtl.c (gen_reg_rtx): Assert that
8248 crtl->emit.regno_pointer_align_length is non-zero.
8250 2014-02-20 Richard Henderson <rth@redhat.com>
8253 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8254 on failure the store back into EXPECT.
8256 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8257 Sandra Loosemore <sandra@codesourcery.com>
8259 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8260 * config/nios2/nios2.c (nios2_function_profiler): Add
8261 -fPIC (flag_pic == 2) support.
8262 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8263 (nios2_large_offset_p): New function.
8264 (nios2_unspec_reloc_p): Move up position, update to use
8265 nios2_large_offset_p.
8266 (nios2_unspec_address): Remove function.
8267 (nios2_unspec_offset): New function.
8268 (nios2_large_got_address): New function.
8269 (nios2_got_address): Add large offset support.
8270 (nios2_legitimize_tls_address): Update usage of removed and new
8272 (nios2_symbol_binds_local_p): New function.
8273 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8274 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8275 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8276 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8277 processing for (const (unspec ...)).
8278 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8280 2014-02-20 Richard Biener <rguenther@suse.de>
8282 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8283 doing the substitution.
8284 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8286 2014-02-20 Martin Jambor <mjambor@suse.cz>
8289 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8290 info when checking whether lattices are bottom.
8292 2014-02-20 Richard Biener <rguenther@suse.de>
8295 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8298 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8301 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8302 parameter specifying the scaling.
8303 (inline_call): Update.
8304 (want_inline_recursively): Guard division by zero.
8305 (recursive_inlining): Update.
8306 * ipa-inline.h (clone_inlined_nodes): Update.
8308 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8311 * config/i386/i386.c (classify_argument): Pass structures of size
8312 64 bytes or less in register.
8314 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8315 Kirill Yukhin <kirill.yukhin@intel.com>
8317 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8318 (_mm_rcp28_round_ss): Ditto.
8319 (_mm_rsqrt28_round_sd): Ditto.
8320 (_mm_rsqrt28_round_ss): Ditto.
8321 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8322 (_mm_rcp14_round_ss): Ditto.
8323 (_mm_rsqrt14_round_sd): Ditto.
8324 (_mm_rsqrt14_round_ss): Ditto.
8325 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8326 the first input operand, get rid of match_dup.
8327 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8329 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8331 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8332 operand as the first input operand, set type attribute.
8333 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8335 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8336 operand as the first input operand, set type attribute.
8338 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8340 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8343 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8346 * config/i386/i386.c (construct_container): Remove TFmode check
8347 for X86_64_INTEGER_CLASS.
8349 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8352 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8353 only when -Wpsabi is enabled.
8355 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8358 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8359 passing arrays in registers are the same as for structs, so remove the
8360 special case for them.
8362 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8364 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8365 destination type, extract only the valid bits if the source type is not
8366 integral and has a different mode.
8368 2014-02-19 Richard Biener <rguenther@suse.de>
8371 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8374 2014-02-19 Richard Biener <rguenther@suse.de>
8377 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8378 (ipa_modify_call_arguments): Emit an argument load explicitely and
8379 preserve virtual SSA form there and for the replacement call.
8380 Do not update SSA form nor free dominance info.
8382 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8384 * ipa.c (function_and_variable_visibility): Also clear WEAK
8385 flag when disolving COMDAT_GROUP.
8387 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8389 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8390 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8392 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8393 do more sanity checks.
8394 (detect_type_change): Return true when giving up early.
8395 (compute_complex_assign_jump_func): Fix type parameter of
8396 ipa_set_ancestor_jf.
8397 (compute_complex_ancestor_jump_func): Likewise.
8398 (update_jump_functions_after_inlining): Fix updating of
8400 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8402 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8404 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8405 inline clones when edge disappears.
8407 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8410 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8411 Split 64-bit moves into 2 patterns. Do not allow the use of
8412 direct move for TDmode in little endian, since the decimal value
8413 has little endian bytes within a word, but the 64-bit pieces are
8414 ordered in a big endian fashion, and normal subreg's of TDmode are
8416 (mov<mode>_64bit_dm): Likewise.
8417 (movtd_64bit_nodm): Likewise.
8419 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8421 PR tree-optimization/60174
8422 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8423 statement of an SSA_NAME that occurs in an abnormal PHI node.
8425 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8428 * final.c (SEEN_BB): Remove.
8429 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8430 (final_scan_insn): Don't force_source_line on second
8431 NOTE_INSN_BASIC_BLOCK.
8433 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8436 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8437 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8438 (type_natural_mode): Warn ABI change when %zmm register is not
8439 available for AVX512F vector value passing.
8441 2014-02-18 Kai Tietz <ktietz@redhat.com>
8444 * config/i386/i386.c (ix86_expand_prologue): Use value in
8445 rax register as displacement when restoring %r10 or %rax.
8446 Fix wrong offset when restoring both registers.
8448 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8450 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8451 assertion with conditional return.
8453 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8454 Uros Bizjak <ubizjak@gmail.com>
8457 * config/i386/driver-i386.c (host_detect_local_cpu): If
8458 YMM state is not saved by the OS, also clear has_f16c. Move
8459 CPUID 0x80000001 handling before YMM state saving checking.
8461 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8463 PR rtl-optimization/58960
8464 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8465 factored out from ...
8466 (sched_init): ... here.
8467 (free_global_sched_pressure_data): New, factored out from ...
8468 (sched_finish): ... here.
8469 * sched-int.h (free_global_sched_pressure_data): Declare.
8470 * sched-rgn.c (nr_regions_initial): New static global.
8471 (haifa_find_rgns): Initialize it.
8472 (schedule_region): Disable sched-pressure for the newly
8475 2014-02-17 Richard Biener <rguenther@suse.de>
8477 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8478 release SSA defs of pattern stmts.
8480 2014-02-17 Richard Biener <rguenther@suse.de>
8482 * tree-inline.c (expand_call_inline): Release the virtual
8483 operand defined by the call we are about to inline.
8485 2014-02-17 Richard Biener <rguenther@suse.de>
8487 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8489 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8490 Ilya Tocar <ilya.tocar@intel.com>
8492 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8493 arguments order in builtin.
8494 (_mm512_permutexvar_epi64): Ditto.
8495 (_mm512_mask_permutexvar_epi64): Ditto
8496 (_mm512_maskz_permutexvar_epi32): Ditto
8497 (_mm512_permutexvar_epi32): Ditto
8498 (_mm512_mask_permutexvar_epi32): Ditto
8500 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8502 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8503 (p8_vmrgow): Likewise.
8505 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8507 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8510 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8513 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8514 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8515 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8516 using direct move instructions on ISA 2.07. Also adjust
8517 instruction length for 64-bit.
8518 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8519 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8521 2014-02-15 Alan Modra <amodra@gmail.com>
8525 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8526 find_replacement on parts of insn rtl that might be reloaded.
8528 2014-02-15 Richard Biener <rguenther@suse.de>
8530 PR tree-optimization/60183
8531 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8532 (tree_ssa_phiprop): Calculate and free post-dominators.
8534 2014-02-14 Jeff Law <law@redhat.com>
8536 PR rtl-optimization/60131
8537 * ree.c (get_extended_src_reg): New function.
8538 (combine_reaching_defs): Use it rather than assuming location of REG.
8539 (find_and_remove_re): Verify first operand of extension is
8540 a REG before adding the insns to the copy list.
8542 2014-02-14 Roland McGrath <mcgrathr@google.com>
8544 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8545 * configure: Regenerated.
8546 * config.in: Regenerated.
8547 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8548 instead of ASM_SHORT.
8550 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8551 Richard Earnshaw <rearnsha@arm.com>
8553 PR rtl-optimization/59535
8554 * lra-constraints.c (process_alt_operands): Encourage alternative
8555 when unassigned pseudo class is superset of the alternative class.
8556 (inherit_reload_reg): Don't inherit when optimizing for code size.
8557 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8558 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8559 modes not less than 4 for Thumb1.
8561 2014-02-14 Kyle McMartin <kyle@redhat.com>
8564 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8566 2014-02-14 Richard Biener <rguenther@suse.de>
8568 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8569 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8570 Do not strip INDIRECT_REFs.
8572 2014-02-14 Richard Biener <rguenther@suse.de>
8575 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8576 DECL_FUNCTION_SPECIFIC_TARGET.
8577 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8578 * tree-streamer-out.c (pack_ts_target_option): Remove.
8579 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8580 (write_ts_function_decl_tree_pointers): Do not stream
8581 DECL_FUNCTION_SPECIFIC_TARGET.
8582 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8583 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8584 (lto_input_ts_function_decl_tree_pointers): Do not stream
8585 DECL_FUNCTION_SPECIFIC_TARGET.
8587 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8589 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8590 (get_initial_def_for_induction, vectorizable_induction): Ignore
8591 debug stmts when looking for exit_phi.
8592 (vectorizable_live_operation): Fix up condition.
8594 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8596 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8597 nreverse() because it changes the content of original tree list.
8599 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8601 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8602 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8604 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8606 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8607 GNU coding standards.
8609 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8612 * dwarf2out.c (gen_subprogram_die): Don't call
8613 add_calling_convention_attribute if subr_die is old_die.
8615 2014-02-13 Sharad Singhai <singhai@google.com>
8617 * doc/optinfo.texi: Fix order of nodes.
8619 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8621 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8622 operands[2], not operands[3].
8624 2014-02-13 Richard Biener <rguenther@suse.de>
8627 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8628 mention the possibility of an in-tree build.
8629 (CLooG): Update recommended version to 0.18.1, mention the
8630 possibility of an in-tree build and clarify that the ISL
8631 bundled with CLooG does not work.
8633 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8636 * expr.c (compress_float_constant): If x is a hard register,
8637 extend into a pseudo and then move to x.
8639 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8641 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8642 caused by bad second argument to warning_at() with -mhotpatch and
8643 nested functions (e.g. with gfortran).
8645 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8647 * opts.c (option_name): Remove "enabled by default" rider.
8649 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8651 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8653 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8654 Uros Bizjak <ubizjak@gmail.com>
8657 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8658 * configure: Regenerated.
8660 2014-02-12 Richard Biener <rguenther@suse.de>
8662 * vec.c (vec_prefix::calculate_allocation): Move as
8663 inline variant to vec.h.
8664 (vec_prefix::calculate_allocation_1): New out-of-line version.
8665 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8666 (vec_prefix::m_has_auto_buf): Rename to ...
8667 (vec_prefix::m_using_auto_storage): ... this.
8668 (vec_prefix::calculate_allocation): Inline the easy cases
8669 and dispatch to calculate_allocation_1 which doesn't need the
8671 (va_heap::reserve): Use gcc_checking_assert.
8672 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8673 m_using_auto_storage.
8674 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8676 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8677 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8678 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8680 2014-02-12 Richard Biener <rguenther@suse.de>
8682 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8683 when we found a dependence.
8685 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8687 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8689 (maybe_fold_stmt): ... into this new function.
8690 * omp-low.c (lower_omp): Update comment.
8692 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8695 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8698 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8700 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8701 identifiers in comments.
8702 (cortexa53_extra_costs): Likewise.
8703 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8704 (cortexa7_extra_costs): Likewise.
8705 (cortexa12_extra_costs): Likewise.
8706 (cortexa15_extra_costs): Likewise.
8707 (v7m_extra_costs): Likewise.
8709 2014-02-12 Richard Biener <rguenther@suse.de>
8712 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8713 of posix_memalign being successful.
8714 (lower_stmt): Restrict lowering of posix_memalign to when
8715 -ftree-bit-ccp is enabled.
8717 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8719 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8721 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8723 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8725 PR rtl-optimization/60116
8726 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8727 other_insn once the combination has been validated.
8729 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8732 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8734 * ipa-devirt.c: Include demangle.h
8735 (odr_violation_reported): New static variable.
8736 (add_type_duplicate): Update odr_violations.
8737 (maybe_record_node): Add completep parameter; update it.
8738 (record_target_from_binfo): Add COMPLETEP parameter;
8739 update it as needed.
8740 (possible_polymorphic_call_targets_1): Likewise.
8741 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8742 rename FINAL to COMPLETE.
8743 (record_targets_from_bases): Sanity check we found the binfo;
8744 fix COMPLETEP updating.
8745 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8746 parameter, fix computing of COMPLETEP.
8747 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8748 at LTO time do demangling.
8749 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8750 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8752 (gimple_get_virt_method_for_binfo): Likewise.
8753 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8754 gimple_get_virt_method_for_vtable): Update prototypes.
8756 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8759 * genautomata.c (add_presence_absence): Fix typo with
8760 {final_}presence_list.
8762 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8765 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8766 for VSX/Altivec vectors that land in GPR registers.
8768 2014-02-11 Richard Henderson <rth@redhat.com>
8769 Jakub Jelinek <jakub@redhat.com>
8772 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8773 around drhs if type conversion to lacc->type is not useless.
8775 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8777 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8779 (cortex-a57.cortex-a53): Likewise.
8780 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8782 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8784 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8787 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8789 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8790 add_options_for_arm_vfp3.
8792 2014-02-11 Jeff Law <law@redhat.com>
8795 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8796 object with an undesirable mode.
8798 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8801 * config/i386/sol2-9.h: New file.
8802 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8803 *-*-solaris2.9*): Use it.
8805 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8807 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8808 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8810 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8812 * config/microblaze/microblaze.c: Extend mcpu version format
8814 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8816 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8818 2014-02-10 Richard Henderson <rth@redhat.com>
8821 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8822 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8823 ms-abi vs -mno-accumulate-outgoing-args.
8824 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8825 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8828 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8831 * cfgexpand.c (expand_asm_operands): Attach source location to
8832 ASM_INPUT rtx objects.
8833 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8835 2014-02-10 Nick Clifton <nickc@redhat.com>
8837 * config/mn10300/mn10300.c (popcount): New function.
8838 (mn10300_expand_prologue): Include saved registers in stack usage
8841 2014-02-10 Jeff Law <law@redhat.com>
8844 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8845 when changing the SET_DEST of a prior insn to avoid an input reload.
8847 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8849 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8850 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8851 -mcall-openbsd, or -mcall-linux.
8852 (CC1_ENDIAN_BIG_SPEC): Remove.
8853 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8854 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8855 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8856 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8857 and %cc1_endian_default.
8858 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8860 2014-02-10 Richard Biener <rguenther@suse.de>
8862 PR tree-optimization/60115
8863 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8864 MEM_REF handling. Properly verify that the accesses are not
8865 out of the objects bound.
8867 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8869 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8872 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8874 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8875 proper constants and fix formatting.
8876 (possible_polymorphic_call_targets): Fix formatting.
8878 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8879 Ilya Tocar <ilya.tocar@intel.com>
8881 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8882 (_mm512_loadu_epi32): Renamed into...
8883 (_mm512_loadu_si512): This.
8884 (_mm512_storeu_epi32): Renamed into...
8885 (_mm512_storeu_si512): This.
8886 (_mm512_maskz_ceil_ps): Removed.
8887 (_mm512_maskz_ceil_pd): Ditto.
8888 (_mm512_maskz_floor_ps): Ditto.
8889 (_mm512_maskz_floor_pd): Ditto.
8890 (_mm512_floor_round_ps): Ditto.
8891 (_mm512_floor_round_pd): Ditto.
8892 (_mm512_ceil_round_ps): Ditto.
8893 (_mm512_ceil_round_pd): Ditto.
8894 (_mm512_mask_floor_round_ps): Ditto.
8895 (_mm512_mask_floor_round_pd): Ditto.
8896 (_mm512_mask_ceil_round_ps): Ditto.
8897 (_mm512_mask_ceil_round_pd): Ditto.
8898 (_mm512_maskz_floor_round_ps): Ditto.
8899 (_mm512_maskz_floor_round_pd): Ditto.
8900 (_mm512_maskz_ceil_round_ps): Ditto.
8901 (_mm512_maskz_ceil_round_pd): Ditto.
8902 (_mm512_expand_pd): Ditto.
8903 (_mm512_expand_ps): Ditto.
8904 * config/i386/i386.c (ix86_builtins): Remove
8905 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8906 (bdesc_args): Ditto.
8907 * config/i386/predicates.md (const1256_operand): New.
8908 (const_1_to_2_operand): Ditto.
8909 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8910 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8911 (*avx512pf_gatherpf<mode>sf): Ditto.
8912 (avx512pf_gatherpf<mode>df): Ditto.
8913 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8914 (*avx512pf_gatherpf<mode>df): Ditto.
8915 (avx512pf_scatterpf<mode>sf): Ditto.
8916 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8917 (*avx512pf_scatterpf<mode>sf): Ditto.
8918 (avx512pf_scatterpf<mode>df): Ditto.
8919 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8920 (*avx512pf_scatterpf<mode>df): Ditto.
8921 (avx512f_expand<mode>): Removed.
8922 (<shift_insn><mode>3<mask_name>): Change predicate type.
8924 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8926 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8927 not at the end of datarefs vector use ordered_remove to avoid
8928 reordering datarefs vector.
8931 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8932 mark local addressable non-static vars as GOVD_PRIVATE
8933 instead of GOVD_LOCAL.
8934 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8935 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8939 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8940 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8941 assume_aligned or alloc_align attributes.
8942 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8943 arguments. Handle also assume_aligned and alloc_align attributes.
8944 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8945 calls to functions with assume_aligned or alloc_align attributes.
8946 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8948 2014-02-08 Terry Guo <terry.guo@arm.com>
8950 * doc/invoke.texi: Document ARM -march=armv7e-m.
8952 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8954 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8955 flag on __cilkrts_rethrow builtin.
8958 * ipa-cp.c (determine_versionability): Fail at -O0
8959 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8960 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8963 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8966 * tree-inline.c (copy_forbidden): Fail for
8967 __attribute__((optimize (0))) functions.
8969 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8971 * varpool.c: Include pointer-set.h.
8972 (varpool_remove_unreferenced_decls): Variables in other partitions
8973 will not be output; be however careful to not lose information
8976 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8978 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8979 lookup in the vtable constructor.
8981 2014-02-07 Jeff Law <law@redhat.com>
8984 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8985 define_insn_and_split.
8987 * ipa-inline.c (inline_small_functions): Fix typos.
8989 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8991 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8992 (s390_can_use_return_insn): Declare.
8993 * config/s390/s390.h (EPILOGUE_USES): Define.
8994 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8996 (s390_chunkify_start): Handle return JUMP_LABELs.
8997 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8998 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8999 (s390_can_use_return_insn): New functions.
9000 (s390_fix_long_loop_prediction): Handle conditional returns.
9001 (TARGET_SET_UP_BY_PROLOGUE): Define.
9002 * config/s390/s390.md (ANY_RETURN): New code iterator.
9003 (*creturn, *csimple_return, return, simple_return): New patterns.
9005 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9007 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9008 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9009 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9010 REG_CFA_RESTORE list when deciding not to restore a register.
9012 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9014 * config/s390/s390.c: Include tree-pass.h and context.h.
9015 (s390_early_mach): New function, split out from...
9016 (s390_emit_prologue): ...here.
9017 (pass_data_s390_early_mach): New pass structure.
9018 (pass_s390_early_mach): New class.
9019 (s390_option_override): Create and register early_mach pass.
9020 Move to end of file.
9022 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9024 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9025 to match for the exit block.
9027 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9029 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9030 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9031 Reject misaligned operands.
9033 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9035 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9037 2014-02-07 Richard Biener <rguenther@suse.de>
9040 * gimple-low.c (lower_builtin_posix_memalign): New function.
9041 (lower_stmt): Call it to lower posix_memalign in a way
9042 to make alignment info accessible.
9044 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9047 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9048 __builtin_setjmp_receiver.
9050 2014-02-07 Richard Biener <rguenther@suse.de>
9053 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9054 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9055 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9056 Handle BUILT_IN_POSIX_MEMALIGN.
9057 (find_func_clobbers): Likewise.
9058 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9059 (call_may_clobber_ref_p_1): Likewise.
9061 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9064 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9067 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9070 * lto-cgraph.c (lto_output_node): Use
9071 symtab_get_symbol_partitioning_class.
9072 (lto_output_varpool_node): likewise.
9073 (symtab_get_symbol_partitioning_class): Move here from
9075 * cgraph.h (symbol_partitioning_class): Likewise.
9076 (symtab_get_symbol_partitioning_class): Declare.
9078 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9080 * ggc.h (ggc_internal_cleared_alloc): New macro.
9081 * vec.h (vec_safe_copy): Handle memory stats.
9082 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9083 * target-globals.c (save_target_globals): Likewise.
9085 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9088 * expr.c (emit_move_resolve_push): Export; be bit more selective
9089 on when to clear alias set.
9090 * expr.h (emit_move_resolve_push): Declare.
9091 * function.h (struct function): Add tail_call_marked.
9092 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9093 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9094 * config/i386/i386.md (TImode move expander): De not call
9096 (FP push expanders): Preserve memory attributes.
9097 * config/i386/sse.md (push<mode>1): Remove.
9098 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9099 (ix86_expand_push): Remove.
9100 * config/i386/mmx.md (push<mode>1): Remove.
9102 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9104 PR rtl-optimization/60030
9105 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9106 lopart with paradoxical subreg before shifting it up by hprec.
9108 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9110 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9111 Remove extra newline at end of file.
9112 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9113 (arm_issue_rate): Handle cortexa57.
9114 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9115 (cortex-a57.cortex-a53): Likewise.
9117 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9120 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9121 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9123 (arm_expand_prologue): Adjust all callers.
9124 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9125 info, registers also at the lowest numbered registers side. Use
9126 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9130 * var-tracking.c (adjust_mems): Before adding a SET to
9131 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9133 2014-02-06 Alan Modra <amodra@gmail.com>
9136 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9137 change SDmode to DDmode when lra_in_progress.
9139 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9142 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9143 free_data_ref on the dr first, and before goto again also set dr
9144 to the next dr. For simd_lane_access, free old datarefs[i] before
9145 overwriting it. For get_vectype_for_scalar_type failure, don't
9146 free_data_ref if simd_lane_access.
9148 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9151 * tree.h (opts_for_fn): New inline function.
9152 (opt_for_fn): Define.
9153 * config/i386/i386.c (ix86_function_regparm): Use
9154 opt_for_fn (decl, optimize) instead of optimize.
9156 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9158 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9159 for SYMBOL_REF in large memory model.
9161 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9163 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9165 (cortex-a57): Likewise.
9166 (cortex-a57.cortex-a53): Likewise.
9168 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9169 Kugan Vivekanandarajah <kuganv@linaro.org>
9171 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9173 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9175 2014-02-06 Richard Biener <rguenther@suse.de>
9177 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9178 set_loop_copy and initialize_original_copy_tables.
9180 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9182 * config/aarch64/aarch64-simd.md
9183 (aarch64_ashr_simddi): Change QI to SI.
9185 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9186 Jakub Jelinek <jakub@redhat.com>
9189 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9192 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9194 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9195 CODE_FOR_altivec_vpku[hw]um to
9196 CODE_FOR_altivec_vpku[hw]um_direct.
9197 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9198 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9199 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9200 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9202 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9204 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9205 generation for -maltivec=be.
9206 (altivec_vsumsws): Simplify redundant test.
9208 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9210 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9211 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9212 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9213 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9214 gen_altivec_vpkuwum.
9215 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9217 (altivec_vpks<VI_char>ss): Likewise.
9218 (altivec_vpks<VI_char>us): Likewise.
9219 (altivec_vpku<VI_char>us): Likewise.
9220 (altivec_vpku<VI_char>um): Likewise.
9221 (altivec_vpku<VI_char>um_direct): New (copy of
9222 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9224 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9225 target is little endian and -maltivec=be is not specified.
9226 (*altivec_vupkhs<VU_char>_direct): New (copy of
9227 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9228 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9229 target is little endian and -maltivec=be is not specified.
9230 (*altivec_vupkls<VU_char>_direct): New (copy of
9231 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9232 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9233 little endian and -maltivec=be is not specified.
9234 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9235 little endian and -maltivec=be is not specified.
9237 2014-02-05 Richard Henderson <rth@redhat.com>
9240 * combine-stack-adj.c: Revert r206943.
9241 * sched-int.h (struct deps_desc): Add last_args_size.
9242 * sched-deps.c (init_deps): Initialize it.
9243 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9244 contain REG_ARGS_SIZE notes.
9246 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9248 * lto-cgraph.c (asm_nodes_output): Make global.
9249 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9250 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9251 (driver_handle_option): Handle OPT_fwpa.
9253 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9256 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9257 a comment typo and formatting issue. If odr_hash hasn't been
9258 created, return vNULL and set *completep to false.
9261 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9262 bb with no successors.
9264 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9267 * doc/invoke.texi (-march): Clarify documentation for ARM.
9271 2014-02-05 Richard Biener <rguenther@suse.de>
9273 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9274 when not vectorizing because of too many alias checks.
9275 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9276 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9278 2014-02-05 Nick Clifton <nickc@redhat.com>
9280 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9281 accept extended registers in any mode when compiling for the MN10300.
9283 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9285 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9286 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9287 sanitization attributes.
9288 (can_inline_edge_p): Likewise.
9289 (sanitize_attrs_match_for_inline_p): New function.
9291 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9293 * ipa-prop.c (detect_type_change): Shor circuit testing of
9294 type changes on THIS pointer.
9296 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9299 * config/pa/pa.c (legitimize_tls_address): Return original address
9300 if not passed a SYMBOL_REF rtx.
9301 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9303 (pa_emit_move_sequence): Simplify TLS source operands.
9304 (pa_legitimate_constant_p): Reject all TLS constants.
9305 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9306 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9308 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9310 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9311 groups when we know they are controlled by LTO.
9312 * varasm.c (default_binds_local_p_1): If object is in other partition,
9313 it will be resolved locally.
9315 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9317 * config/host-linux.c (linux_gt_pch_use_address): Don't
9318 use SSIZE_MAX because it is not always defined.
9320 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9323 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9324 threshold for pseudo splitting.
9325 (update_ebb_live_info): Process call argument hard registers and
9326 hard registers from insn definition too.
9327 (max_small_class_regs_num): New constant.
9328 (inherit_in_ebb): Update live hard regs through EBBs. Update
9329 reloads_num only for small register classes. Don't split for
9332 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9335 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9338 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9340 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9343 2014-02-04 Marek Polacek <polacek@redhat.com>
9345 * gdbinit.in (pel): Define.
9347 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9349 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9352 2014-02-04 Richard Biener <rguenther@suse.de>
9355 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9356 in function context local.
9357 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9358 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9359 similar to LTO_imported_decl_ref.
9361 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9363 PR tree-optimization/60002
9364 * cgraphclones.c (build_function_decl_skip_args): Clear
9367 PR tree-optimization/60023
9368 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9369 false to gsi_replace.
9370 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9371 has been in some EH region and vec_stmt could throw, add
9372 vec_stmt into the same EH region.
9373 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9374 has no lhs, ignore it.
9375 * internal-fn.c (expand_MASK_LOAD): Likewise.
9378 * tree-inline.c (copy_forbidden): Fail for
9379 __attribute__((optimize (0))) functions.
9382 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9383 is set, copy one less argument.
9384 (expand_simd_clones): Don't subtract clone_info->inbranch
9385 from simd_clone_struct_alloc argument.
9387 PR rtl-optimization/57915
9388 * recog.c (simplify_while_replacing): If all unary/binary/relational
9389 operation arguments are constant, attempt to simplify those.
9392 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9393 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9395 2014-02-04 Richard Biener <rguenther@suse.de>
9397 PR tree-optimization/60012
9398 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9399 TBAA disambiguation to all DDRs.
9401 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9404 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9405 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9407 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9410 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9412 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9414 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9415 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9417 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9420 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9421 to figure out targets of polymorphic calls with known decl.
9422 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9423 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9424 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9425 (get_polymorphic_call_info): ... here.
9426 (get_polymorphic_call_info_from_invariant): New function.
9428 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9430 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9431 lookup via vtable pointer; check for type consistency
9432 and turn inconsitent facts into UNREACHABLE.
9433 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9434 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9435 type inconsistent querries; return UNREACHABLE instead.
9437 2014-02-03 Richard Henderson <rth@twiddle.net>
9440 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9441 already processed this node.
9442 (normalize_one_pred_1): Pass along mark_set.
9443 (normalize_one_pred): Create and destroy a pointer_set_t.
9444 (normalize_one_pred_chain): Likewise.
9446 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9448 PR gcov-profile/58602
9449 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9451 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9454 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9455 -fno-devirtualize; try to devirtualize by the knowledge of
9456 virtual table pointer given by aggregate propagation.
9457 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9458 (ipa_print_node_jump_functions): Dump also offset that
9459 is relevant for polymorphic calls.
9460 (determine_known_aggregate_parts): Add arg_type parameter; use it
9461 instead of determining the type from pointer type.
9462 (ipa_compute_jump_functions_for_edge): Update call of
9463 determine_known_aggregate_parts.
9464 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9465 (gimple_get_virt_method_for_binfo): ... here; simplify using
9466 vtable_pointer_value_to_vtable.
9467 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9468 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9469 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9470 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9472 (vtable_pointer_value_to_binfo): ... here.
9473 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9475 2014-02-03 Teresa Johnson <tejohnson@google.com>
9477 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9478 redef of outer loop index variable.
9480 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9484 * doc/extend.texi (Function Attributes): Typo.
9486 2014-02-03 Cong Hou <congh@google.com>
9488 PR tree-optimization/60000
9489 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9490 if the vectorized statement is a store. A store statement can only
9491 appear at the end of pattern statements.
9493 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9495 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9496 (ix86_option_override_internal): Default long double to 64-bit for
9497 32-bit Bionic and to 128-bit for 64-bit Bionic.
9499 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9500 TARGET_LONG_DOUBLE_128 is true.
9501 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9503 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9504 (mlong-double-64): Negate -mlong-double-128.
9505 (mlong-double-128): New option.
9507 * config/i386/i386-c.c (ix86_target_macros): Define
9508 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9510 * doc/invoke.texi: Document -mlong-double-128.
9512 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9514 PR rtl-optimization/60024
9515 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9517 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9519 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9521 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9523 PR rtl-optimization/57662
9524 * sel-sched.c (code_motion_path_driver): Do not mark already not
9525 existing blocks in the visiting bitmap.
9527 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9529 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9530 on the insn being emitted.
9532 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9533 Will Deacon <will.deacon@arm.com>
9535 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9537 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9539 * config/arm/arm-tables.opt: Regenerate.
9541 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9543 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9544 for vector types other than V16QImode.
9545 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9546 define_expand, and call altivec_expand_vec_perm_le when producing
9547 code with little endian element order.
9548 (*altivec_vperm_<mode>_internal): New insn having previous
9549 behavior of altivec_vperm_<mode>.
9550 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9551 altivec_expand_vec_perm_le when producing code with little endian
9553 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9554 behavior of altivec_vperm_<mode>_uns.
9556 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9558 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9559 (altivec_vsumsws): Add handling for -maltivec=be with a little
9561 (altivec_vsumsws_direct): New.
9562 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9563 gen_altivec_vsumsws.
9565 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9567 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9568 vtable_pointer_value_to_binfo): New functions.
9569 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9570 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9572 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9574 * config/nios2/nios2.md (load_got_register): Initialize GOT
9575 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9576 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9578 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9580 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9581 preserverd by passthrough, do not propagate the type.
9583 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9585 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9586 (mips_atomic_assign_expand_fenv): New function.
9587 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9589 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9591 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9592 (__builtin_mips_set_fcsr): Likewise.
9593 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9594 MIPS_USI_FTYPE_VOID.
9595 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9596 (mips16_expand_set_fcsr): Likewise.
9597 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9598 (mips16_set_fcsr_stub): Likewise.
9599 (mips16_get_fcsr_one_only_stub): New class.
9600 (mips16_set_fcsr_one_only_stub): Likewise.
9601 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9602 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9603 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9604 (hard_float): New availability predicate.
9605 (mips_builtins): Add get_fcsr and set_fcsr.
9606 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9607 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9608 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9609 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9610 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9613 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9615 * config/mips/mips.c (mips_one_only_stub): New class.
9616 (mips_need_mips16_rdhwr_p): Replace with...
9617 (mips16_rdhwr_stub): ...this new variable.
9618 (mips16_stub_call_address): New function.
9619 (mips16_rdhwr_one_only_stub): New class.
9620 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9621 (mips_output_mips16_rdhwr): Delete.
9622 (mips_finish_stub): New function.
9623 (mips_code_end): Use it to handle rdhwr stubs.
9625 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9628 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9629 when calculating size of integer atomic types.
9631 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9633 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9635 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9637 PR tree-optimization/60003
9638 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9639 * profile.c (branch_prob): Use gimple_call_builtin_p
9640 to check for BUILT_IN_SETJMP_RECEIVER.
9641 * tree-inline.c (copy_bb): Call notice_special_calls.
9643 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9646 * lra-constraints.c (process_alt_operands): Update reload_sum only
9649 2014-01-31 Richard Henderson <rth@redhat.com>
9652 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9653 until after else_eh is processed.
9655 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9657 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9658 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9659 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9660 in smmintrin.h, remove them.
9661 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9662 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9663 * config/i386/i386.md (ROUND_SAE): Fix value.
9664 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9665 (const48_operand): New.
9666 * config/i386/subst.md (round), (round_expand): Use
9667 const_4_or_8_to_11_operand.
9668 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9670 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9672 * config/i386/constraints.md (Yk): Swap meaning with k.
9673 * config/i386/i386.md (movhi_internal): Change Yk to k.
9674 (movqi_internal): Ditto.
9675 (*k<logic><mode>): Ditto.
9678 (kandn<mode>): Ditto.
9679 (*<code>hi_1): Ditto.
9680 (*<code>qi_1): Ditto.
9681 (kxnor<mode>): Ditto.
9682 (kortestzhi): Ditto.
9683 (kortestchi): Ditto.
9685 (*one_cmplhi2_1): Ditto.
9686 (*one_cmplqi2_1): Ditto.
9687 * config/i386/sse.md (): Change k to Yk.
9688 (avx512f_load<mode>_mask): Ditto.
9689 (avx512f_blendm<mode>): Ditto.
9690 (avx512f_store<mode>_mask): Ditto.
9691 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9692 (avx512f_storedqu<mode>_mask): Ditto.
9693 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9695 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9696 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9697 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9698 (avx512f_maskcmp<mode>3): Ditto.
9699 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9700 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9701 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9702 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9703 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9704 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9705 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9706 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9707 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9708 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9709 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9710 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9711 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9712 (vec_extract_lo_<mode>_maskm): Ditto.
9713 (vec_extract_hi_<mode>_maskm): Ditto.
9714 (avx512f_vternlog<mode>_mask): Ditto.
9715 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9716 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9717 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9718 (avx512f_<code>v8div16qi2_mask): Ditto.
9719 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9720 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9721 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9722 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9723 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9724 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9725 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9726 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9727 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9728 (avx512cd_maskb_vec_dupv8di): Ditto.
9729 (avx512cd_maskw_vec_dupv16si): Ditto.
9730 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9731 (avx512f_vpermi2var<mode>3_mask): Ditto.
9732 (avx512f_vpermi2var<mode>3_mask): Ditto.
9733 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9734 (*avx512f_gathersi<mode>): Ditto.
9735 (*avx512f_gathersi<mode>_2): Ditto.
9736 (*avx512f_gatherdi<mode>): Ditto.
9737 (*avx512f_gatherdi<mode>_2): Ditto.
9738 (*avx512f_scattersi<mode>): Ditto.
9739 (*avx512f_scatterdi<mode>): Ditto.
9740 (avx512f_compress<mode>_mask): Ditto.
9741 (avx512f_compressstore<mode>_mask): Ditto.
9742 (avx512f_expand<mode>_mask): Ditto.
9743 * config/i386/subst.md (mask): Change k to Yk.
9744 (mask_scalar_merge): Ditto.
9747 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9749 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9751 2014-01-31 Richard Biener <rguenther@suse.de>
9754 * builtins.c (fold_builtin_memory_op): Make sure to not
9755 use a floating-point mode or a boolean or enumeral type for
9758 2014-01-30 DJ Delorie <dj@redhat.com>
9760 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9761 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9762 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9763 whenever main() has an epilogue.
9765 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9767 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9768 unused variable "field".
9769 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9770 (vsx_mergeh_<mode>): Likewise.
9771 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9772 (altivec_vmrghh): Likewise.
9773 (altivec_vmrghw): Likewise.
9774 (altivec_vmrglb): Likewise.
9775 (altivec_vmrglh): Likewise.
9776 (altivec_vmrglw): Likewise.
9777 (altivec_vspltb): Add missing uses.
9778 (altivec_vsplth): Likewise.
9779 (altivec_vspltw): Likewise.
9780 (altivec_vspltsf): Likewise.
9782 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9785 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9786 frame related instructions.
9788 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9790 PR rtl-optimization/59959
9791 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9792 any reload of register whose subreg is invalid.
9794 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9796 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9797 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9798 Add missing return type - void.
9800 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9802 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9803 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9804 remove element index adjustment for endian (now handled in vsx.md
9806 (altivec_expand_vec_perm_const): Use
9807 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9808 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9809 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9810 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9811 define_expand and a new define_insn *altivec_vspltb_internal;
9812 adjust for -maltivec=be on a little endian target.
9813 (altivec_vspltb_direct): New.
9814 (altivec_vsplth): Divide into a define_expand and a new
9815 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9816 little endian target.
9817 (altivec_vsplth_direct): New.
9818 (altivec_vspltw): Divide into a define_expand and a new
9819 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9820 little endian target.
9821 (altivec_vspltw_direct): New.
9822 (altivec_vspltsf): Divide into a define_expand and a new
9823 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9824 a little endian target.
9826 2014-01-30 Richard Biener <rguenther@suse.de>
9828 PR tree-optimization/59993
9829 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9830 can propagate form the earlier stmt and avoid the transform
9831 when the intermediate result is needed.
9833 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9835 * README.Portability: Fix typo.
9837 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9839 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9840 comparison_operator with ordered_comparison_operator.
9842 2014-01-30 Nick Clifton <nickc@redhat.com>
9844 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9845 Rename to mn10300_store_multiple_regs.
9846 * config/mn10300/mn10300.c: Likewise.
9847 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9848 store_multiple_regs.
9849 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9850 Call mn10300_store_multiple_regs.
9852 2014-01-30 Nick Clifton <nickc@redhat.com>
9853 DJ Delorie <dj@redhat.com>
9855 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9856 %fp 2 to keep registers after it properly word-aligned.
9857 (rl78_alloc_physical_registers_umul): Handle the case where both
9858 input operands are the same.
9860 2014-01-30 Richard Biener <rguenther@suse.de>
9862 PR tree-optimization/59903
9863 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9866 2014-01-30 Jason Merrill <jason@redhat.com>
9869 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9872 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9874 2014-01-30 Richard Biener <rguenther@suse.de>
9876 PR tree-optimization/59951
9877 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9879 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9882 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9883 SFmode to DFmode case.
9885 2014-01-29 DJ Delorie <dj@redhat.com>
9887 * config/msp430/msp430.opt (-minrt): New.
9888 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9890 (ENDFILE_SPEC): Likewise.
9892 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9894 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9895 (estimate_function_body_sizes): Use it.
9897 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9900 * dwarf2out.c (is_cxx_auto): New.
9901 (is_base_type): Use it.
9902 (gen_type_die_with_usage): Likewise.
9904 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9906 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9907 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9908 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9909 -maltivec=be with LE targets.
9910 (vsx_mergeh_<mode>): Likewise.
9911 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9912 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9913 (altivec_vmrghb): Replace with define_expand and new
9914 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9915 (altivec_vmrghb_direct): New define_insn.
9916 (altivec_vmrghh): Replace with define_expand and new
9917 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9918 (altivec_vmrghh_direct): New define_insn.
9919 (altivec_vmrghw): Replace with define_expand and new
9920 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9921 (altivec_vmrghw_direct): New define_insn.
9922 (*altivec_vmrghsf): Adjust for endianness.
9923 (altivec_vmrglb): Replace with define_expand and new
9924 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9925 (altivec_vmrglb_direct): New define_insn.
9926 (altivec_vmrglh): Replace with define_expand and new
9927 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9928 (altivec_vmrglh_direct): New define_insn.
9929 (altivec_vmrglw): Replace with define_expand and new
9930 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9931 (altivec_vmrglw_direct): New define_insn.
9932 (*altivec_vmrglsf): Adjust for endianness.
9933 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9934 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9935 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9936 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9937 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9938 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9939 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9940 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9942 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9944 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9945 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9948 2014-01-29 Richard Biener <rguenther@suse.de>
9950 PR tree-optimization/58742
9951 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9952 associate_pointerplus_align.
9953 (associate_pointerplus_diff): New function.
9954 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9955 and associate_pointerplus_diff.
9957 2014-01-29 Richard Biener <rguenther@suse.de>
9959 * lto-streamer.h (LTO_major_version): Bump to 3.
9960 (LTO_minor_version): Reset to 0.
9962 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9964 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9965 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9966 (arm_file_start): Generate correct asm header for armv7ve.
9967 * config/arm/bpabi.h: Add multilib support for armv7ve.
9968 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9969 and cortex-a15 to armv7ve.
9970 * config/arm/t-aprofile: Add multilib support for armv7ve.
9971 * doc/invoke.texi: Document -march=armv7ve.
9973 2014-01-29 Richard Biener <rguenther@suse.de>
9975 PR tree-optimization/58742
9976 * tree-ssa-forwprop.c (associate_plusminus): Return true
9977 if we changed sth, defer EH cleanup to ...
9978 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9979 (simplify_mult): New function.
9981 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9984 PR tree-optimization/59920
9985 * tree.c (build_common_builtin_nodes): Remove
9986 __builtin_setjmp_dispatcher initialization.
9987 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9988 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9989 instead of gsi_after_labels + manually skipping debug stmts.
9990 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9991 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9992 * tree-inline.c (copy_edges_for_bb): Remove
9993 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9994 argument. Ignore computed_goto_p stmts. Don't call
9995 make_abnormal_goto_edges. If a call might need abnormal edges
9996 for non-local gotos, see if it already has an edge to
9997 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9998 with true argument, don't do anything then, otherwise add
9999 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10000 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10002 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10003 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10004 (lower_stmt): Don't set data->calls_builtin_setjmp.
10005 (lower_builtin_setjmp): Adjust comment.
10006 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10007 * tree-cfg.c (found_computed_goto): Remove.
10008 (factor_computed_gotos): Remove.
10009 (make_goto_expr_edges): Return bool, true for computed gotos.
10010 Don't call make_abnormal_goto_edges.
10011 (build_gimple_cfg): Don't set found_computed_goto, don't call
10012 factor_computed_gotos.
10013 (computed_goto_p): No longer static.
10014 (make_blocks): Don't set found_computed_goto.
10015 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10016 (make_edges): If make_goto_expr_edges returns true, push bb
10017 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10018 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10019 vector. Record mapping between bbs and OpenMP regions if there
10020 are any, adjust make_gimple_omp_edges caller. Call
10021 handle_abnormal_edges.
10022 (make_abnormal_goto_edges): Remove.
10023 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10024 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10025 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10026 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10027 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10028 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10029 filling *region also set *region_idx to (*region)->entry->index.
10032 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10033 For REGs set ORIGINAL_REGNO.
10035 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10037 * doc/md.texi: Mention that a target shouldn't implement
10038 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10041 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10043 PR tree-optimization/59594
10044 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10045 a copy of the datarefs vector rather than the vector itself.
10047 2014-01-28 Jason Merrill <jason@redhat.com>
10050 * dwarf2out.c (auto_die): New static.
10051 (gen_type_die_with_usage): Handle C++1y 'auto'.
10052 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10055 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10058 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10059 (SPEC_X32): Likewise.
10060 (SPEC_64): Likewise.
10061 * config/i386/i386.c (ix86_option_override_internal): Turn off
10062 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10064 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10065 * config/i386/i386.h (TARGET_16BIT): New macro.
10066 (TARGET_16BIT_P): Likewise.
10067 * config/i386/i386.opt: Add m16.
10068 * doc/invoke.texi: Document -m16.
10070 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10072 PR preprocessor/59935
10073 * input.c (location_get_source_line): Bail out on when line number
10074 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10076 2014-01-28 Richard Biener <rguenther@suse.de>
10078 PR tree-optimization/58742
10079 * tree-ssa-forwprop.c (associate_plusminus): Handle
10080 pointer subtraction of the form (T)(P + A) - (T)P.
10082 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10084 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10087 2014-01-28 Richard Biener <rguenther@suse.de>
10090 2014-01-28 Richard Biener <rguenther@suse.de>
10092 PR rtl-optimization/45364
10093 PR rtl-optimization/59890
10094 * var-tracking.c (local_get_addr_clear_given_value): Handle
10095 already cleared slot.
10096 (val_reset): Handle not allocated local_get_addr_cache.
10097 (vt_find_locations): Use post-order on the inverted CFG.
10099 2014-01-28 Richard Biener <rguenther@suse.de>
10101 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10103 2014-01-28 Richard Biener <rguenther@suse.de>
10105 PR rtl-optimization/45364
10106 PR rtl-optimization/59890
10107 * var-tracking.c (local_get_addr_clear_given_value): Handle
10108 already cleared slot.
10109 (val_reset): Handle not allocated local_get_addr_cache.
10110 (vt_find_locations): Use post-order on the inverted CFG.
10112 2014-01-28 Alan Modra <amodra@gmail.com>
10114 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10115 * configure.ac <recursive call for build != host>: Define
10116 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10117 and LD_FOR_BUILD too.
10118 * configure: Regenerate.
10120 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10122 * config/i386/i386.c (get_builtin_code_for_version): Separate
10123 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10124 Broadwell from Haswell.
10126 2014-01-27 Steve Ellcey <sellcey@mips.com>
10128 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10129 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10130 * config/mips/mips.c (mips_option_override): Change setting
10132 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10133 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10134 Change from Mask to Var.
10136 2014-01-27 Jeff Law <law@redhat.com>
10138 * ipa-inline.c (inline_small_functions): Fix typo.
10140 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10142 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10143 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10144 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10145 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10146 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10147 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10148 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10149 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10150 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10151 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10152 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10153 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10154 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10155 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10156 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10157 (_mm512_storeu_epi64): Ditto.
10158 (_mm512_cmpge_epi32_mask): Ditto.
10159 (_mm512_cmpge_epu32_mask): Ditto.
10160 (_mm512_cmpge_epi64_mask): Ditto.
10161 (_mm512_cmpge_epu64_mask): Ditto.
10162 (_mm512_cmple_epi32_mask): Ditto.
10163 (_mm512_cmple_epu32_mask): Ditto.
10164 (_mm512_cmple_epi64_mask): Ditto.
10165 (_mm512_cmple_epu64_mask): Ditto.
10166 (_mm512_cmplt_epi32_mask): Ditto.
10167 (_mm512_cmplt_epu32_mask): Ditto.
10168 (_mm512_cmplt_epi64_mask): Ditto.
10169 (_mm512_cmplt_epu64_mask): Ditto.
10170 (_mm512_cmpneq_epi32_mask): Ditto.
10171 (_mm512_cmpneq_epu32_mask): Ditto.
10172 (_mm512_cmpneq_epi64_mask): Ditto.
10173 (_mm512_cmpneq_epu64_mask): Ditto.
10174 (_mm512_expand_pd): Ditto.
10175 (_mm512_expand_ps): Ditto.
10176 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10177 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10178 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10179 * config/i386/i386.c (ix86_builtins): Add
10180 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10181 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10182 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10183 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10184 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10185 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10186 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10187 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10188 IX86_BUILTIN_PMOVUSQW512_MEM.
10189 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10190 __builtin_ia32_pmovsqd512mem_mask,
10191 __builtin_ia32_pmovqd512mem_mask,
10192 __builtin_ia32_pmovusqw512mem_mask,
10193 __builtin_ia32_pmovsqw512mem_mask,
10194 __builtin_ia32_pmovqw512mem_mask,
10195 __builtin_ia32_pmovusdw512mem_mask,
10196 __builtin_ia32_pmovsdw512mem_mask,
10197 __builtin_ia32_pmovdw512mem_mask,
10198 __builtin_ia32_pmovqb512mem_mask,
10199 __builtin_ia32_pmovusqb512mem_mask,
10200 __builtin_ia32_pmovsqb512mem_mask,
10201 __builtin_ia32_pmovusdb512mem_mask,
10202 __builtin_ia32_pmovsdb512mem_mask,
10203 __builtin_ia32_pmovdb512mem_mask.
10204 (bdesc_args): Add __builtin_ia32_expanddf512,
10205 __builtin_ia32_expandsf512.
10206 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10207 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10208 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10209 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10210 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10211 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10212 (avx512f_<code>v8div16qi2_mask_store): This.
10213 (avx512f_expand<mode>): New.
10215 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10217 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10219 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10220 (_mm512_prefetch_i32scatter_pd): Ditto.
10221 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10222 (_mm512_prefetch_i64scatter_pd): Ditto.
10223 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10224 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10225 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10226 (_mm512_prefetch_i32scatter_ps): Ditto.
10227 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10228 (_mm512_prefetch_i64scatter_ps): Ditto.
10229 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10230 * config/i386/i386-builtin-types.def: Define
10231 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10232 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10233 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10234 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10235 IX86_BUILTIN_SCATTERPFQPD.
10236 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10237 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10238 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10239 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10240 __builtin_ia32_scatterpfqps.
10241 (ix86_expand_builtin): Expand new built-ins.
10242 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10243 fix memory access data type.
10244 (*avx512pf_gatherpf<mode>_mask): Ditto.
10245 (*avx512pf_gatherpf<mode>): Ditto.
10246 (avx512pf_scatterpf<mode>): Ditto.
10247 (*avx512pf_scatterpf<mode>_mask): Ditto.
10248 (*avx512pf_scatterpf<mode>): Ditto.
10249 (GATHER_SCATTER_SF_MEM_MODE): New.
10250 (avx512pf_gatherpf<mode>df): Ditto.
10251 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10252 (*avx512pf_scatterpf<mode>df): Ditto.
10254 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10257 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10258 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10261 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10263 * common/config/arm/arm-common.c
10264 (arm_rewrite_mcpu): Handle multiple names.
10266 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10268 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10270 * gimple-builder.h (create_gimple_tmp): Delete.
10272 2014-01-27 Christian Bruel <christian.bruel@st.com>
10274 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10277 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10279 * config/pa/pa.md (call): Generate indirect long calls to non-local
10280 functions when outputing 32-bit code.
10281 (call_value): Likewise except for special call to buggy powf function.
10283 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10284 portable runtime and PIC indirect calls.
10285 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10286 and PIC call sequences. Use ldo instead of blr to set return register
10287 in PIC call sequence.
10289 2014-01-25 Walter Lee <walt@tilera.com>
10291 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10292 avoid clobbering a live register.
10294 2014-01-25 Walter Lee <walt@tilera.com>
10296 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10297 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10298 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10299 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10301 2014-01-25 Walter Lee <walt@tilera.com>
10303 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10304 arguments on even registers.
10305 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10307 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10308 (BIGGEST_ALIGNMENT): Ditto.
10309 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10311 2014-01-25 Walter Lee <walt@tilera.com>
10313 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10314 insns before bundling.
10315 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10317 2014-01-25 Walter Lee <walt@tilera.com>
10319 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10320 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10321 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10323 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10325 * config/mips/constraints.md (kl): Delete.
10326 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10327 define expands, using...
10328 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10329 instructions for MIPS16.
10330 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10331 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10333 2014-01-25 Walter Lee <walt@tilera.com>
10335 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10339 2014-01-25 Walter Lee <walt@tilera.com>
10341 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10342 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10344 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10346 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10349 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10351 * print-rtl.c (in_call_function_usage): New var.
10352 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10353 EXPR_LIST mode as mode and not as reg note name.
10355 PR middle-end/59561
10356 * cfgloopmanip.c (copy_loop_info): If
10357 loop->warned_aggressive_loop_optimizations, make sure
10358 the flag is set in target loop too.
10360 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10362 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10364 * builtins.def: Likewise.
10365 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10366 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10367 * ira.c (ira_setup_eliminable_regset): Likewise.
10368 * omp-low.c (gate_expand_omp): Likewise.
10369 (execute_lower_omp): Likewise.
10370 (diagnose_sb_0): Likewise.
10371 (gate_diagnose_omp_blocks): Likewise.
10372 (simd_clone_clauses_extract): Likewise.
10375 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10377 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10378 correction for little endian...
10379 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10382 2014-01-24 Jeff Law <law@redhat.com>
10384 PR tree-optimization/59919
10385 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10386 for non-returning calls.
10388 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10390 * common/config/aarch64/aarch64-common.c
10391 (aarch64_rewrite_mcpu): Handle multiple names.
10392 * config/aarch64/aarch64.h
10393 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10395 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10397 * input.c (add_file_to_cache_tab): Handle the case where fopen
10400 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10403 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10404 from push operand if code of push isn't PRE_DEC.
10406 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10409 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10410 -mquad-memory-atomic. Update -mquad-memory documentation to say
10411 it is only used for non-atomic loads/stores.
10413 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10414 -mquad-memory or -mquad-memory-atomic switches.
10416 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10417 -mquad-memory-atomic to ISA 2.07 support.
10419 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10420 to separate support of normal quad word memory operations (ldq, stq)
10421 from the atomic quad word memory operations.
10423 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10424 support to separate non-atomic quad word operations from atomic
10425 quad word operations. Disable non-atomic quad word operations in
10426 little endian mode so that we don't have to swap words after the
10427 load and before the store.
10428 (quad_load_store_p): Add comment about atomic quad word support.
10429 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10430 options printed with -mdebug=reg.
10432 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10433 -mquad-memory-atomic as the test for whether we have quad word
10434 atomic instructions.
10435 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10436 or -mp8-vector are used, allow byte/half-word atomic operations.
10438 * config/rs6000/sync.md (load_lockedti): Insure that the address
10439 is a proper indexed or indirect address for the lqarx instruction.
10440 On little endian systems, swap the hi/lo registers after the lqarx
10442 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10443 insure the address is valid for the lqarx instruction.
10444 (store_conditionalti): Insure that the address is a proper indexed
10445 or indirect address for the stqcrx. instruction. On little endian
10446 systems, swap the hi/lo registers before doing the stqcrx.
10448 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10449 insure the address is valid for the stqcrx. instruction.
10451 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10452 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10453 type of quad memory support is available.
10455 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10457 PR regression/59915
10458 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10459 there is a danger of looping.
10461 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10463 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10464 force flag_ira_loop_pressure if set via command line.
10466 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10468 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10469 (ashr_simd): New builtin handling DI mode.
10470 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10471 (aarch64_sshr_simddi): New match pattern.
10472 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10473 (vshrd_n_s64): Likewise.
10474 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10476 2014-01-23 Nick Clifton <nickc@redhat.com>
10478 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10479 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10480 favour of mcu specific scripts.
10481 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10484 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10485 Alex Velenko <Alex.Velenko@arm.com>
10487 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10488 (vaddv_s16): Likewise.
10489 (vaddv_s32): Likewise.
10490 (vaddv_u8): Likewise.
10491 (vaddv_u16): Likewise.
10492 (vaddv_u32): Likewise.
10493 (vaddvq_s8): Likewise.
10494 (vaddvq_s16): Likewise.
10495 (vaddvq_s32): Likewise.
10496 (vaddvq_s64): Likewise.
10497 (vaddvq_u8): Likewise.
10498 (vaddvq_u16): Likewise.
10499 (vaddvq_u32): Likewise.
10500 (vaddvq_u64): Likewise.
10501 (vaddv_f32): Likewise.
10502 (vaddvq_f32): Likewise.
10503 (vaddvq_f64): Likewise.
10504 (vmaxv_f32): Likewise.
10505 (vmaxv_s8): Likewise.
10506 (vmaxv_s16): Likewise.
10507 (vmaxv_s32): Likewise.
10508 (vmaxv_u8): Likewise.
10509 (vmaxv_u16): Likewise.
10510 (vmaxv_u32): Likewise.
10511 (vmaxvq_f32): Likewise.
10512 (vmaxvq_f64): Likewise.
10513 (vmaxvq_s8): Likewise.
10514 (vmaxvq_s16): Likewise.
10515 (vmaxvq_s32): Likewise.
10516 (vmaxvq_u8): Likewise.
10517 (vmaxvq_u16): Likewise.
10518 (vmaxvq_u32): Likewise.
10519 (vmaxnmv_f32): Likewise.
10520 (vmaxnmvq_f32): Likewise.
10521 (vmaxnmvq_f64): Likewise.
10522 (vminv_f32): Likewise.
10523 (vminv_s8): Likewise.
10524 (vminv_s16): Likewise.
10525 (vminv_s32): Likewise.
10526 (vminv_u8): Likewise.
10527 (vminv_u16): Likewise.
10528 (vminv_u32): Likewise.
10529 (vminvq_f32): Likewise.
10530 (vminvq_f64): Likewise.
10531 (vminvq_s8): Likewise.
10532 (vminvq_s16): Likewise.
10533 (vminvq_s32): Likewise.
10534 (vminvq_u8): Likewise.
10535 (vminvq_u16): Likewise.
10536 (vminvq_u32): Likewise.
10537 (vminnmv_f32): Likewise.
10538 (vminnmvq_f32): Likewise.
10539 (vminnmvq_f64): Likewise.
10541 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10543 * config/aarch64/aarch64-simd.md
10544 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10545 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10546 (*aarch64_mul3_elt<mode>): Likewise.
10547 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10548 (*aarch64_mul3_elt_to_64v2df): Likewise.
10549 (*aarch64_mla_elt<mode>): Likewise.
10550 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10551 (*aarch64_mls_elt<mode>): Likewise.
10552 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10553 (*aarch64_fma4_elt<mode>): Likewise.
10554 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10555 (*aarch64_fma4_elt_to_64v2df): Likewise.
10556 (*aarch64_fnma4_elt<mode>): Likewise.
10557 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10558 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10559 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10560 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10561 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10562 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10563 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10564 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10565 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10567 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10569 * config/aarch64/aarch64-simd.md
10570 (aarch64_be_checked_get_lane<mode>): New define_expand.
10571 * config/aarch64/aarch64-simd-builtins.def
10572 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10573 New builtin definition.
10574 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10575 Use new safe be builtin.
10577 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10579 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10581 (aarch64_be_st1<mode>): Likewise.
10582 (aarch_ld1<VALL:mode>): Define_expand modified.
10583 (aarch_st1<VALL:mode>): Likewise.
10584 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10585 (UNSPEC_ST1): Likewise.
10587 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10589 * config/microblaze/microblaze.md: Add trap insn and attribute
10591 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10593 PR preprocessor/58580
10594 * input.h (location_get_source_line): Take an additional line_size
10596 (void diagnostics_file_cache_fini): Declare new function.
10597 * input.c (struct fcache): New type.
10598 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10599 New static constants.
10600 (diagnostic_file_cache_init, total_lines_num)
10601 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10602 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10603 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10604 (get_next_line, read_next_line, goto_next_line, read_line_num):
10605 New static function definitions.
10606 (diagnostic_file_cache_fini): New function.
10607 (location_get_source_line): Take an additional output line_len
10608 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10610 * diagnostic.c (diagnostic_finish): Call
10611 diagnostic_file_cache_fini.
10612 (adjust_line): Take an additional input parameter for the length
10613 of the line, rather than calculating it with strlen.
10614 (diagnostic_show_locus): Adjust the use of
10615 location_get_source_line and adjust_line with respect to their new
10616 signature. While displaying a line now, do not stop at the first
10617 null byte. Rather, display the zero byte as a space and keep
10618 going until we reach the size of the line.
10619 * Makefile.in: Add vec.o to OBJS-libcommon
10621 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10622 Ilya Tocar <ilya.tocar@intel.com>
10624 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10625 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10626 (__builtin_ia32_kmov16): Ditto.
10627 * config/i386/i386.md (UNSPEC_KMOV): New.
10630 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10632 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10633 (_mm512_storeu_si512): Ditto.
10635 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10638 * rtl.h (get_referenced_operands): Declare.
10639 * recog.c (get_referenced_operands): New function.
10640 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10641 operands have been referenced when recording LO_SUM references.
10643 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10645 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10647 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10649 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10650 Enable for generic and recent AMD targets.
10652 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10654 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10655 ARG_SIZE note when adjustment was eliminated.
10657 2014-01-22 Jeff Law <law@redhat.com>
10659 PR tree-optimization/59597
10660 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10661 in file. Accept new argument REGISTERING and use it to modify
10662 dump output appropriately.
10663 (register_jump_thread): Corresponding changes.
10664 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10665 thread paths involving joiner blocks. Add code to dump cancelled
10666 jump threading paths.
10668 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10670 PR rtl-optimization/59477
10671 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10672 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10674 2014-01-22 Tom Tromey <tromey@redhat.com>
10676 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10678 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10680 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10682 PR rtl-optimization/59896
10683 * lra-constraints.c (process_alt_operands): Check unused note for
10684 matched operands of insn with no output reloads.
10686 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10688 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10689 (mips_move_from_gpr_cost): Likewise.
10691 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10693 PR rtl-optimization/59858
10694 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10695 ira_class_hard_regs_num.
10696 (process_alt_operands): Increase reject for dying matched operand.
10698 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10701 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10702 smaller than size, perform several stores or loads and stores
10703 at dst + count - size to store or copy all of size bytes, rather
10704 than just last modesize bytes.
10706 2014-01-20 DJ Delorie <dj@redhat.com>
10708 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10709 that CLOBBERs are REGs before propogating their values.
10711 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10713 PR middle-end/59789
10714 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10715 (cgraph_inline_failed_type): New function.
10716 * cgraph.h (DEFCIFCODE): Add type.
10717 (cgraph_inline_failed_type_t): New enum.
10718 (cgraph_inline_failed_type): New prototype.
10719 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10720 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10721 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10722 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10723 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10724 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10725 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10726 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10727 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10728 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10729 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10730 OPTIMIZATION_MISMATCH.
10731 * tree-inline.c (expand_call_inline): Emit errors during
10732 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10734 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10737 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10738 mode attribute in insn output.
10740 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10742 * output.h (output_constant): Delete.
10743 * varasm.c (output_constant): Make private.
10745 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10747 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10749 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10751 PR middle-end/59860
10752 * tree.h (fold_builtin_strcat): New prototype.
10753 * builtins.c (fold_builtin_strcat): No longer static. Add len
10754 argument, if non-NULL, don't call c_strlen. Optimize
10755 directly into __builtin_memcpy instead of __builtin_strcpy.
10756 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10757 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10759 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10761 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10762 for SImode_address_operand operands, having only a REG argument.
10764 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10766 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10767 loader name using mbig-endian.
10768 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10770 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10772 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10773 (-mtune): Likewise.
10776 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10778 * config/aarch64/aarch64-protos.h
10779 (aarch64_cannot_change_mode_class_ptr): Declare.
10780 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10781 aarch64_cannot_change_mode_class_ptr): New.
10782 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10783 backend hook aarch64_cannot_change_mode_class.
10785 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10787 * common/config/aarch64/aarch64-common.c
10788 (aarch64_handle_option): Don't handle any option order logic here.
10789 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10790 selected_cpu, warn on architecture version mismatch.
10791 (aarch64_override_options): Fix parsing order for option strings.
10793 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10794 Iain Sandoe <iain@codesourcery.com>
10797 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10798 warning. Amend comment to reflect current functionality.
10800 2014-01-20 Richard Biener <rguenther@suse.de>
10802 PR middle-end/59860
10803 * builtins.c (fold_builtin_strcat): Remove case better handled
10804 by tree-ssa-strlen.c.
10806 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10808 * config/aarch64/aarch64.opt
10809 (mcpu, march, mtune): Make case-insensitive.
10811 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10814 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10815 if operands[1] is a REG or ZERO_EXTEND of a REG.
10817 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10819 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10821 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10823 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10824 long non-pic millicode calls.
10826 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10828 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10830 2014-01-19 Kito Cheng <kito@0xlab.org>
10832 * builtins.c (expand_movstr): Check movstr expand done or fail.
10834 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10835 H.J. Lu <hongjiu.lu@intel.com>
10838 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10839 to DImode for zero-extended addresses.
10841 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10843 PR rtl-optimization/57763
10844 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10845 on the new indirect jump_insn and increment LABEL_NUSES (label).
10847 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10851 * config.gcc (x86_archs): New variable.
10852 (x86_64_archs): Likewise.
10853 (x86_cpus): Likewise.
10854 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10855 --with-arch/--with-cpu= options.
10856 Support --with-arch=/--with-cpu={nehalem,westmere,
10857 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10859 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10861 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10862 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10864 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10866 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10868 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10871 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10872 clear cpp_get_options (parse_in)->warn_unused_macros for
10873 ix86_target_macros_internal with cpp_define.
10875 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10877 * jump.c (delete_related_insns): Keep (use (insn))s.
10878 * reorg.c (redundant_insn): Check for barriers too.
10880 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10882 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10884 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10886 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10887 call to $$dyncall when TARGET_LONG_CALLS is true.
10889 2014-01-17 Jeff Law <law@redhat.com>
10891 * ree.c (combine_set_extension): Temporarily disable test for
10892 changing number of hard registers.
10894 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10896 PR middle-end/58125
10897 * ipa-inline-analysis.c (inline_free_summary):
10898 Do not free summary of aliases.
10900 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10902 PR middle-end/59706
10903 * gimplify.c (gimplify_expr): Use create_tmp_var
10904 instead of create_tmp_var_raw. If cond doesn't have
10905 integral type, don't add the IFN_ANNOTATE builtin at all.
10907 2014-01-17 Martin Jambor <mjambor@suse.cz>
10910 * ipa-cp.c (prev_edge_clone): New variable.
10911 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10912 Also resize prev_edge_clone vector.
10913 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10914 (ipcp_edge_removal_hook): New function.
10915 (ipcp_driver): Register ipcp_edge_removal_hook.
10917 2014-01-17 Andrew Pinski <apinski@cavium.com>
10918 Steve Ellcey <sellcey@mips.com>
10921 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10924 2014-01-17 Jeff Law <law@redhat.com>
10926 PR middle-end/57904
10927 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10928 so that pass_ccp runs first.
10930 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10932 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10933 (ix86_adjust_cost): Use !TARGET_XXX.
10934 (do_reorder_for_imul): Likewise.
10935 (swap_top_of_ready_list): Likewise.
10936 (ix86_sched_reorder): Likewise.
10938 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10940 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10941 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10942 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10943 (intel_memset): New. Duplicate slm_memset.
10944 (intel_cost): New. Duplicate slm_cost.
10945 (m_INTEL): New macro.
10946 (processor_target_table): Add "intel".
10947 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10948 with PROCESSOR_INTEL for "intel".
10949 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10950 PROCESSOR_SILVERMONT.
10951 (ix86_issue_rate): Likewise.
10952 (ix86_adjust_cost): Likewise.
10953 (ia32_multipass_dfa_lookahead): Likewise.
10954 (swap_top_of_ready_list): Likewise.
10955 (ix86_sched_reorder): Likewise.
10956 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10957 instead of TARGET_OPT_AGU.
10958 * config/i386/i386.h (TARGET_INTEL): New.
10959 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10960 (processor_type): Add PROCESSOR_INTEL.
10961 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10962 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10964 2014-01-17 Marek Polacek <polacek@redhat.com>
10967 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10970 2014-01-17 Richard Biener <rguenther@suse.de>
10972 PR tree-optimization/46590
10973 * opts.c (default_options_table): Add entries for
10974 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10975 all enabled at -O1 but not for -Og.
10976 * common.opt (fbranch-count-reg): Remove Init(1).
10977 (fmove-loop-invariants): Likewise.
10978 (ftree-pta): Likewise.
10980 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10982 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10983 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10984 decls to at least the GCC 4.8 used alignments.
10987 * tree-nested.c (convert_nonlocal_reference_stmt,
10988 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10989 of GIMPLE_BIND stmts, adjust associated decls.
10991 2014-01-17 Richard Biener <rguenther@suse.de>
10993 PR tree-optimization/46590
10994 * vec.h (vec<>::bseach): New member function implementing
10995 binary search according to C89 bsearch.
10996 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10997 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10998 bitmap pointer again. Make accesses_in_loop a flat array.
10999 (mem_ref_obstack): New global.
11000 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11001 (mark_ref_stored): Likewise.
11002 (ref_indep_loop_p_2): Likewise.
11003 (set_ref_stored_in_loop): New helper function.
11004 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11005 (memref_free): Adjust.
11006 (record_mem_ref_loc): Simplify.
11007 (gather_mem_refs_stmt): Adjust.
11008 (sort_locs_in_loop_postorder_cmp): New function.
11009 (analyze_memory_references): Sort accesses_in_loop after
11010 loop postorder number.
11011 (find_ref_loc_in_loop_cmp): New function.
11012 (for_all_locs_in_loop): Find relevant cluster of locs in
11013 accesses_in_loop and iterate without recursion.
11014 (execute_sm): Avoid uninit warning.
11015 (struct ref_always_accessed): Simplify.
11016 (ref_always_accessed::operator ()): Likewise.
11017 (ref_always_accessed_p): Likewise.
11018 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11019 loop postorder numbers here.
11020 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11023 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11026 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11027 on decls for which assemble_alias has been called.
11029 2014-01-17 Nick Clifton <nickc@redhat.com>
11031 * config/msp430/msp430.opt: (mcpu): New option.
11032 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11033 (msp430_option_override): Parse target_cpu. If the MCU name
11034 matches a generic string, clear target_mcu.
11035 (msp430_attr): Allow numeric interrupt values up to 63.
11036 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11037 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11039 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11041 * config/msp430/msp430.md (popm): Use %J rather than %I.
11042 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11043 (addhi_cy_i): Use immediate_operand for operand 2.
11044 * doc/invoke.texi: Document -mcpu option.
11046 2014-01-17 Richard Biener <rguenther@suse.de>
11048 PR rtl-optimization/38518
11049 * df.h (df_analyze_loop): Declare.
11050 * df-core.c: Include cfgloop.h.
11051 (df_analyze_1): Split out main part of df_analyze.
11052 (df_analyze): Adjust.
11053 (loop_inverted_post_order_compute): New function.
11054 (loop_post_order_compute): Likewise.
11055 (df_analyze_loop): New function avoiding whole-function
11056 postorder computes.
11057 * loop-invariant.c (find_defs): Use df_analyze_loop.
11058 (find_invariants): Adjust.
11059 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11061 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11063 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11064 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11066 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11068 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11069 traversal when removing references.
11071 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11074 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11076 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11078 PR middle-end/56791
11079 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11080 pushing a reload for an autoinc when we had previously reloaded an
11081 inner part of the address.
11083 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11085 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11087 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11088 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11089 when not giving up or versioning for alias only because of
11091 (vect_analyze_data_ref_dependences): Set to true.
11092 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11094 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11095 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11098 PR middle-end/58344
11099 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11102 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11103 operand 0 predicate for gathers, use a new pseudo as subtarget.
11105 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11107 PR middle-end/59609
11108 * lra-constraints.c (process_alt_operands): Add printing debug info.
11109 Check absence of input/output reloads for matched operands too.
11111 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11113 PR rtl-optimization/59835
11114 * ira.c (ira_init_register_move_cost): Increase cost for
11117 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11119 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11121 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11124 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11125 non-register objects. Use gen_(high/low)part more consistently.
11128 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11131 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11132 endian support, remove tests for WORDS_BIG_ENDIAN.
11133 (p8_mfvsrd_3_<mode>): Likewise.
11134 (reload_gpr_from_vsx<mode>): Likewise.
11135 (reload_gpr_from_vsxsf): Likewise.
11136 (p8_mfvsrd_4_disf): Likewise.
11138 2014-01-16 Richard Biener <rguenther@suse.de>
11140 PR rtl-optimization/46590
11141 * lcm.c (compute_antinout_edge): Use postorder iteration.
11142 (compute_laterin): Use inverted postorder iteration.
11144 2014-01-16 Nick Clifton <nickc@redhat.com>
11146 PR middle-end/28865
11147 * varasm.c (output_constant): Return the number of bytes actually
11149 (output_constructor_array_range): Update the field size with the
11150 number of bytes emitted by output_constant.
11151 (output_constructor_regular_field): Likewise. Also do not
11152 complain if the total number of bytes emitted is now greater
11153 than the expected fieldpos.
11154 * output.h (output_constant): Update prototype and descriptive comment.
11156 2014-01-16 Marek Polacek <polacek@redhat.com>
11158 PR middle-end/59827
11159 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11160 it is error_mark_node.
11162 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11164 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11165 VALID_AVX256_REG_OR_OI_MODE.
11167 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11169 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11170 current procedure should be profiled.
11172 2014-01-15 Andrew Pinski <apinski@cavium.com>
11174 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11175 of moving from/to the STACK_REG register class.
11177 2014-01-15 Richard Henderson <rth@redhat.com>
11180 * reginfo.c (global_regs_decl): Globalize.
11181 * rtl.h (global_regs_decl): Declare.
11182 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11183 reserved via global_regs.
11185 2014-01-15 Teresa Johnson <tejohnson@google.com>
11187 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11189 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11191 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11192 and vmulosh rather than call gen_vec_widen_smult_*.
11193 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11194 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11195 (vec_widen_smult_even_v16qi): Likewise.
11196 (vec_widen_umult_even_v8hi): Likewise.
11197 (vec_widen_smult_even_v8hi): Likewise.
11198 (vec_widen_umult_odd_v16qi): Likewise.
11199 (vec_widen_smult_odd_v16qi): Likewise.
11200 (vec_widen_umult_odd_v8hi): Likewise.
11201 (vec_widen_smult_odd_v8hi): Likewise.
11202 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11203 vmuloub rather than call gen_vec_widen_umult_*.
11204 (vec_widen_umult_lo_v16qi): Likewise.
11205 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11206 vmulosb rather than call gen_vec_widen_smult_*.
11207 (vec_widen_smult_lo_v16qi): Likewise.
11208 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11209 rather than call gen_vec_widen_umult_*.
11210 (vec_widen_umult_lo_v8hi): Likewise.
11211 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11212 rather than call gen_vec_widen_smult_*.
11213 (vec_widen_smult_lo_v8hi): Likewise.
11215 2014-01-15 Jeff Law <law@redhat.com>
11217 PR tree-optimization/59747
11218 * ree.c (find_and_remove_re): Properly handle case where a second
11219 eliminated extension requires widening a copy created for elimination
11220 of a prior extension.
11221 (combine_set_extension): Ensure that the number of hard regs needed
11222 for a destination register does not change when we widen it.
11224 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11226 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11227 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11228 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11229 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11230 (avr-*-rtems*): Likewise.
11231 (bfin*-rtems*): Likewise.
11232 (moxie-*-rtems*): Likewise.
11233 (h8300-*-rtems*): Likewise.
11234 (i[34567]86-*-rtems*): Likewise.
11235 (lm32-*-rtems*): Likewise.
11236 (m32r-*-rtems*): Likewise.
11237 (m68k-*-rtems*): Likewise.
11238 (microblaze*-*-rtems*): Likewise.
11239 (mips*-*-rtems*): Likewise.
11240 (powerpc-*-rtems*): Likewise.
11241 (sh-*-rtems*): Likewise.
11242 (sparc-*-rtems*): Likewise.
11243 (sparc64-*-rtems*): Likewise.
11244 (v850-*-rtems*): Likewise.
11245 (m32c-*-rtems*): Likewise.
11247 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11249 PR rtl-optimization/59511
11250 * ira.c (ira_init_register_move_cost): Use memory costs for some
11251 cases of register move cost calculations.
11252 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11253 instead of BB frequency.
11254 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11255 * lra-assigns.c (find_hard_regno_for): Ditto.
11257 2014-01-15 Richard Biener <rguenther@suse.de>
11259 PR tree-optimization/59822
11260 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11261 (vectorizable_load): Use it to hoist defs of uses of invariant
11262 loads out of the loop.
11264 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11265 Kugan Vivekanandarajah <kuganv@linaro.org>
11268 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11271 2014-01-15 Richard Biener <rguenther@suse.de>
11273 PR rtl-optimization/59802
11274 * lcm.c (compute_available): Use inverted postorder to seed
11275 the initial worklist.
11277 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11280 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11281 ADDR_REGS for invalid symrefs in non-PIC code.
11283 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11286 * builtins.c (determine_block_size): Initialize *probable_max_size
11287 even if len_rtx is CONST_INT.
11289 2014-01-14 Andrew Pinski <apinski@cavium.com>
11291 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11292 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11293 (cortexa53_tunings): Likewise.
11294 (aarch64_sched_issue_rate): New function.
11295 (TARGET_SCHED_ISSUE_RATE): Define.
11297 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11299 * ira-costs.c (find_costs_and_classes): Add missed
11300 ira_init_register_move_cost_if_necessary.
11302 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11305 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11307 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11310 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11311 to indicate if type is used for function return value. Warn ABI
11312 change if the vector mode isn't available for function return value.
11313 (ix86_function_arg_advance): Pass false to type_natural_mode.
11314 (ix86_function_arg): Likewise.
11315 (ix86_gimplify_va_arg): Likewise.
11316 (function_arg_32): Don't warn ABI change.
11317 (ix86_function_value): Pass true to type_natural_mode.
11318 (ix86_return_in_memory): Likewise.
11319 (ix86_struct_value_rtx): Removed.
11320 (TARGET_STRUCT_VALUE_RTX): Likewise.
11322 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11324 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11325 converting a conditional jump into a conditional return.
11327 2014-01-14 Richard Biener <rguenther@suse.de>
11329 PR tree-optimization/58921
11330 PR tree-optimization/59006
11331 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11332 hoisting invariant stmts.
11333 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11334 invariant loads on the preheader edge if possible.
11336 2014-01-14 Joey Ye <joey.ye@arm.com>
11338 * doc/plugin.texi (Building GCC plugins): Update to C++.
11340 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11342 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11343 (_mm_rcp28_round_ss): Ditto.
11344 (_mm_rsqrt28_round_sd): Ditto.
11345 (_mm_rsqrt28_round_ss): Ditto.
11346 (_mm_rcp28_sd): Ditto.
11347 (_mm_rcp28_ss): Ditto.
11348 (_mm_rsqrt28_sd): Ditto.
11349 (_mm_rsqrt28_ss): Ditto.
11350 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11351 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11352 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11353 (IX86_BUILTIN_RCP28SD): Ditto.
11354 (IX86_BUILTIN_RCP28SS): Ditto.
11355 (IX86_BUILTIN_RSQRT28SD): Ditto.
11356 (IX86_BUILTIN_RSQRT28SS): Ditto.
11357 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11358 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11359 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11360 (ix86_expand_special_args_builtin): Expand new FTYPE.
11361 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11362 (srcp14<mode>): Make insn unary.
11363 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11364 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11365 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11366 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11367 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11368 Fix rounding: make it SAE only.
11369 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11371 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11373 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11374 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11375 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11376 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11377 (round_saeonly_mask_scalar_operand4): Ditto.
11378 (round_saeonly_mask_scalar_op3): Ditto.
11379 (round_saeonly_mask_scalar_op4): Ditto.
11381 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11383 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11384 Implement -maltivec=be for vec_insert and vec_extract.
11386 2014-01-10 DJ Delorie <dj@redhat.com>
11388 * config/msp430/msp430.md (call_internal): Don't allow memory
11389 references with SP as the base register.
11390 (call_value_internal): Likewise.
11391 * config/msp430/constraints.md (Yc): New. For memory references
11392 that don't use SP as a base register.
11394 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11395 "an integer without a # prefix"
11396 * config/msp430/msp430.md (epilogue_helper): Use it.
11398 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11401 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11402 AVX512F gather builtins.
11403 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11404 on gather decls with INTEGER_TYPE masktype.
11405 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11406 directly into the builtin rather than hoisting it before loop.
11408 PR tree-optimization/59387
11409 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11410 (scev_const_prop): If folded_casts and type has undefined overflow,
11411 use force_gimple_operand instead of force_gimple_operand_gsi and
11412 for each added stmt if it is assign with
11413 arith_code_with_undefined_signed_overflow, call
11414 rewrite_to_defined_overflow.
11415 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11416 gimple-fold.h instead.
11417 (arith_code_with_undefined_signed_overflow,
11418 rewrite_to_defined_overflow): Moved to ...
11419 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11420 rewrite_to_defined_overflow): ... here. No longer static.
11421 Include gimplify-me.h.
11422 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11423 rewrite_to_defined_overflow): New prototypes.
11425 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11427 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11429 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11431 * builtins.c (get_object_alignment_2): Minor tweak.
11432 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11434 2014-01-13 Christian Bruel <christian.bruel@st.com>
11436 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11437 optimized non constant lengths.
11439 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11442 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11443 load as __atomic_load_N if possible.
11445 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11447 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11449 (rs6000_expand_builtin): Adjust call.
11451 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11454 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11455 * config/rs6000/rs6000.c: Include target-globals.h.
11456 (rs6000_set_current_function): Instead of doing target_reinit
11457 unconditionally, use save_target_globals_default_opts and
11458 restore_target_globals.
11460 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11462 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11463 (rs6000_expand_builtin): Handle mffs and mtfsf.
11464 (rs6000_init_builtins): Define mffs and mtfsf.
11465 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11466 (rs6000_mffs): New pattern.
11467 (rs6000_mtfsf): New pattern.
11469 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11471 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11472 Start narrowing with START. Apply candidate-use pair
11473 and check overall cost in narrowing.
11474 (iv_ca_prune): Pass new argument.
11476 2014-01-10 Jeff Law <law@redhat.com>
11478 PR middle-end/59743
11479 * ree.c (combine_reaching_defs): Ensure the defining statement
11480 occurs before the extension when optimizing extensions with
11481 different source and destination hard registers.
11483 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11486 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11487 vtables into the type inheritance graph.
11489 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11491 PR rtl-optimization/59754
11492 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11493 modes in the REGNO != REGNO case.
11495 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11497 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11499 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11501 PR tree-optimization/59745
11502 * tree-predcom.c (tree_predictive_commoning_loop): Call
11503 free_affine_expand_cache if giving up because components is NULL.
11505 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11506 GC in payload of target_globals struct instead of allocating them on
11507 the heap and the larger structs separately using GC.
11508 * target-globals.h (struct target_globals): Make regs, hard_regs,
11509 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11510 of GTY((skip)) and change type to void *.
11511 (reset_target_globals): Cast loads from those fields to corresponding
11514 2014-01-10 Steve Ellcey <sellcey@mips.com>
11517 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11518 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11519 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11521 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11524 * aarch64-modes.def (CC_Zmode): New flags mode.
11525 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11526 represents an equality.
11527 (aarch64_get_condition_code): Handle CC_Zmode.
11528 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11530 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11532 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11533 extraction in good case.
11535 2014-01-10 Richard Biener <rguenther@suse.de>
11537 PR tree-optimization/59374
11538 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11539 checking after SLP discovery. Mark stmts not participating
11540 in any SLP instance properly.
11542 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11544 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11545 when handling a SET rtx.
11547 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11549 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11550 (cortex-a57): Likewise.
11551 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11553 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11555 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11556 non-iwmmxt builtins.
11558 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11562 * ipa-devirt.c record_target_from_binfo): Take as argument
11563 stack of binfos and lookup matching one for virtual inheritance.
11564 (possible_polymorphic_call_targets_1): Update.
11566 2014-01-10 Huacai Chen <chenhc@lemote.com>
11568 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11569 kernel strings for Loongson-2E/2F/3A.
11571 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11573 PR middle-end/59670
11574 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11575 is_gimple_call before calling gimple_call_internal_p.
11577 2014-01-09 Steve Ellcey <sellcey@mips.com>
11579 * Makefile.in (TREE_FLOW_H): Remove.
11580 (TREE_SSA_H): Add file names from tree-flow.h.
11581 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11582 * tree.h: Remove tree-flow.h reference.
11583 * hash-table.h: Remove tree-flow.h reference.
11584 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11585 reference with tree-ssa-loop.h.
11587 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11589 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11590 default element-order behavior for -maltivec.
11591 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11592 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11593 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11594 when targeting big endian, at least for now.
11595 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11597 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11599 PR middle-end/47735
11600 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11601 var satisfies use_register_for_decl, just take into account type
11602 alignment, rather than decl alignment.
11604 PR tree-optimization/59622
11605 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11606 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11607 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11608 Don't devirtualize for inplace at all. For targets.length () == 1,
11609 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11611 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11613 * config/i386/i386.md (cpu): Remove the unused btver1.
11615 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11617 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11619 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11622 * tree-core.h (struct target_globals): New forward declaration.
11623 (struct tree_target_option): Add globals field.
11624 * tree.h (TREE_TARGET_GLOBALS): Define.
11625 (prepare_target_option_nodes_for_pch): New prototype.
11626 * target-globals.h (struct target_globals): Define even if
11627 !SWITCHABLE_TARGET.
11628 * tree.c (prepare_target_option_node_for_pch,
11629 prepare_target_option_nodes_for_pch): New functions.
11630 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11631 * config/i386/i386.c: Include target-globals.h.
11632 (ix86_set_current_function): Instead of doing target_reinit
11633 unconditionally, use save_target_globals_default_opts and
11634 restore_target_globals.
11636 2014-01-09 Richard Biener <rguenther@suse.de>
11638 PR tree-optimization/59715
11639 * tree-cfg.h (split_critical_edges): Declare.
11640 * tree-cfg.c (split_critical_edges): Export.
11641 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11643 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11645 * cfgexpand.c (expand_stack_vars): Optionally disable
11646 asan stack protection.
11647 (expand_used_vars): Likewise.
11648 (partition_stack_vars): Likewise.
11649 * asan.c (asan_emit_stack_protection): Optionally disable
11650 after return stack usage.
11651 (instrument_derefs): Optionally disable memory access instrumentation.
11652 (instrument_builtin_call): Likewise.
11653 (instrument_strlen_call): Likewise.
11654 (asan_protect_global): Optionally disable global variables protection.
11655 * doc/invoke.texi: Added doc for new options.
11656 * params.def: Added new options.
11657 * params.h: Likewise.
11659 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11661 PR rtl-optimization/59724
11662 * ifcvt.c (cond_exec_process_if_block): Don't call
11663 flow_find_head_matching_sequence with 0 longest_match.
11664 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11665 non-active insns if !stop_after.
11666 (try_head_merge_bb): Revert 2014-01-07 changes.
11668 2014-01-08 Jeff Law <law@redhat.com>
11670 * ree.c (get_sub_rtx): New function, extracted from...
11671 (merge_def_and_ext): Here.
11672 (combine_reaching_defs): Use get_sub_rtx.
11674 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11676 * cgraph.h (varpool_variable_node): Do not choke on null node.
11678 2014-01-08 Catherine Moore <clm@codesourcery.com>
11680 * config/mips/mips.md (simple_return): Attempt to use JRC
11682 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11684 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11686 PR rtl-optimization/59137
11687 * reorg.c (steal_delay_list_from_target): Call update_block for
11689 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11691 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11693 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11694 two duplicate entries.
11696 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11699 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11701 * config/mips/mips.c (mips_truncated_op_cost): New function.
11702 (mips_rtx_costs): Adjust test for BADDU.
11703 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11705 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11707 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11708 (*baddu_si): ...this new pattern.
11710 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11713 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11715 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11717 PR middle-end/57748
11718 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11720 (expand_expr, expand_normal): Adjust.
11721 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11722 inner_reference_p. Use inner_reference_p to expand inner references.
11723 (store_expr): Adjust.
11724 * cfgexpand.c (expand_call_stmt): Adjust.
11726 2014-01-08 Rong Xu <xur@google.com>
11728 * gcov-io.c (gcov_var): Move from gcov-io.h.
11729 (gcov_position): Ditto.
11730 (gcov_is_error): Ditto.
11731 (gcov_rewrite): Ditto.
11732 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11733 only part to libgcc/libgcov.h.
11735 2014-01-08 Marek Polacek <polacek@redhat.com>
11737 PR middle-end/59669
11738 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11740 2014-01-08 Marek Polacek <polacek@redhat.com>
11743 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11745 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11747 PR rtl-optimization/59649
11748 * stor-layout.c (get_mode_bounds): For BImode return
11749 0 and STORE_FLAG_VALUE.
11751 2014-01-08 Richard Biener <rguenther@suse.de>
11753 PR middle-end/59630
11754 * gimple.h (is_gimple_builtin_call): Remove.
11755 (gimple_builtin_call_types_compatible_p): New.
11756 (gimple_call_builtin_p): New overload.
11757 * gimple.c (is_gimple_builtin_call): Remove.
11758 (validate_call): Rename to ...
11759 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11760 check return types.
11761 (validate_type): New static function.
11762 (gimple_call_builtin_p): New overload and adjust.
11763 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11764 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11765 (gimple_fold_stmt_to_constant_1): Likewise.
11766 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11768 2014-01-08 Richard Biener <rguenther@suse.de>
11770 PR middle-end/59471
11771 * gimplify.c (gimplify_expr): Gimplify register-register type
11772 VIEW_CONVERT_EXPRs to separate stmts.
11774 2014-01-07 Jeff Law <law@redhat.com>
11776 PR middle-end/53623
11777 * ree.c (combine_set_extension): Handle case where source
11778 and destination registers in an extension insn are different.
11779 (combine_reaching_defs): Allow source and destination registers
11780 in extension to be different under limited circumstances.
11781 (add_removable_extension): Remove restriction that the
11782 source and destination registers in the extension are the same.
11783 (find_and_remove_re): Emit a copy from the extension's
11784 destination to its source after the defining insn if
11785 the source and destination registers are different.
11787 PR middle-end/59285
11788 * ifcvt.c (merge_if_block): If we are merging a block with more than
11789 one successor with a block with no successors, remove any BARRIER
11790 after the second block.
11792 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11794 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11796 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11799 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11800 for 14-bit register offsets when INT14_OK_STRICT is false.
11802 2014-01-07 Roland Stigge <stigge@antcom.de>
11803 Michael Meissner <meissner@linux.vnet.ibm.com>
11806 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11807 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11809 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11811 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11814 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11816 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11817 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11818 rtx is const0_rtx or not.
11820 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11823 * target-globals.c (save_target_globals): Remove this_fn_optab
11825 * toplev.c: Include optabs.h.
11826 (target_reinit): Temporarily restore the global options if another
11827 set of options are in force.
11829 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11831 PR rtl-optimization/58668
11832 * cfgcleanup.c (flow_find_cross_jump): Don't count
11833 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11834 to determine what is counted.
11835 (flow_find_head_matching_sequence): Use active_insn_p to determine
11837 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11839 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11840 determine what is counted.
11842 PR tree-optimization/59643
11843 * tree-predcom.c (split_data_refs_to_components): If one dr is
11844 read and one write, determine_offset fails and the write isn't
11845 in the bad component, just put the read into the bad component.
11847 2014-01-07 Mike Stump <mikestump@comcast.net>
11848 Jakub Jelinek <jakub@redhat.com>
11851 * tree-core.h (struct tree_optimization_option): Change optabs
11852 type from unsigned char * to void *.
11853 * optabs.c (init_tree_optimization_optabs): Adjust
11854 TREE_OPTIMIZATION_OPTABS initialization.
11856 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11859 * config/i386/i386.h (struct machine_function): Add
11860 no_drap_save_restore field.
11861 * config/i386/i386.c (ix86_save_reg): Use
11862 !cfun->machine->no_drap_save_restore instead of
11863 crtl->stack_realign_needed.
11864 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11865 this function clears frame_pointer_needed. Set
11866 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11867 and DRAP reg is needed.
11869 2014-01-06 Marek Polacek <polacek@redhat.com>
11872 * doc/implement-c.texi: Mention that other integer types are
11873 permitted as bit-field types in strictly conforming mode.
11875 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11877 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11878 is newly allocated.
11880 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11882 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11884 2014-01-06 Martin Jambor <mjambor@suse.cz>
11887 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11889 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11891 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11895 * var-tracking.c (add_stores): Preserve the value of the source even if
11896 we don't record the store.
11898 2014-01-06 Terry Guo <terry.guo@arm.com>
11900 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11902 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11905 * config/darwin.c (darwin_function_section): Adjust return values to
11906 correspond to optimisation changes made in r206070.
11908 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11910 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11911 from prefetch_block tune setting.
11912 (nocona_cost): Correct size of prefetch block to 64.
11914 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11916 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11917 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11918 used to save the static chain register in the computation of the offset
11919 from which the FP registers need to be restored.
11921 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11923 PR tree-optimization/59519
11924 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11925 ICE if get_current_def (current_new_name) is already non-NULL, as long
11926 as it is a phi result of some other phi in *new_exit_bb that has
11929 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11930 or vmovdqu* for misaligned_operand.
11931 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11932 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11933 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11934 aligned_mem for AVX512F masked aligned load and store builtins and for
11935 non-temporal moves.
11937 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11939 PR tree-optimization/59651
11940 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11941 Address range for negative step should be added by TYPE_SIZE_UNIT.
11943 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11945 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11946 overlapping registers also for registers other than the stack pointer.
11948 2014-01-03 Marek Polacek <polacek@redhat.com>
11951 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11954 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11957 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11960 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11961 (push splitter): Use <P:MODE_SIZE> instead of
11962 GET_MODE_SIZE (<P:MODE>mode).
11963 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11964 (mov -1, reg peephole2): Likewise.
11965 * config/i386/sse.md (*mov<mode>_internal,
11966 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11967 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11968 *<code><mode>3, *andnot<mode>3<mask_name>,
11969 <mask_codefor><code><mode>3<mask_name>): Likewise.
11970 * config/i386/subst.md (mask_mode512bit_condition,
11971 sd_mask_mode512bit_condition): Likewise.
11973 2014-01-02 Xinliang David Li <davidxl@google.com>
11975 PR tree-optimization/59303
11976 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11977 (dump_predicates): Better output format.
11978 (pred_equal_p): New function.
11979 (is_neq_relop_p): Ditto.
11980 (is_neq_zero_form_p): Ditto.
11981 (pred_expr_equal_p): Ditto.
11982 (pred_neg_p): Ditto.
11983 (simplify_pred): Ditto.
11984 (simplify_preds_2): Ditto.
11985 (simplify_preds_3): Ditto.
11986 (simplify_preds_4): Ditto.
11987 (simplify_preds): Ditto.
11988 (push_pred): Ditto.
11989 (push_to_worklist): Ditto.
11990 (get_pred_info_from_cmp): Ditto.
11991 (is_degenerated_phi): Ditto.
11992 (normalize_one_pred_1): Ditto.
11993 (normalize_one_pred): Ditto.
11994 (normalize_one_pred_chain): Ditto.
11995 (normalize_preds): Ditto.
11996 (normalize_cond_1): Remove function.
11997 (normalize_cond): Ditto.
11998 (is_gcond_subset_of): Ditto.
11999 (is_subset_of_any): Ditto.
12000 (is_or_set_subset_of): Ditto.
12001 (is_and_set_subset_of): Ditto.
12002 (is_norm_cond_subset_of): Ditto.
12003 (pred_chain_length_cmp): Ditto.
12004 (convert_control_dep_chain_into_preds): Type change.
12005 (find_predicates): Ditto.
12006 (find_def_preds): Ditto.
12007 (destroy_predicates_vecs): Ditto.
12008 (find_matching_predicates_in_rest_chains): Ditto.
12009 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12010 (is_pred_expr_subset): Ditto.
12011 (is_pred_chain_subset_of): Ditto.
12012 (is_included_in): Ditto.
12013 (is_superset_of): Ditto.
12015 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12017 Update copyright years.
12019 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12021 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12022 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12023 config/arc/arc.md, config/arc/arc.opt,
12024 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12025 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12026 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12027 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12028 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12029 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12030 vtable-verify.c, vtable-verify.h: Use the standard form for the
12033 2014-01-02 Tobias Burnus <burnus@net-b.de>
12035 * gcc.c (process_command): Update copyright notice dates.
12036 * gcov-dump.c: Ditto.
12038 * doc/cpp.texi: Bump @copying's copyright year.
12039 * doc/cppinternals.texi: Ditto.
12040 * doc/gcc.texi: Ditto.
12041 * doc/gccint.texi: Ditto.
12042 * doc/gcov.texi: Ditto.
12043 * doc/install.texi: Ditto.
12044 * doc/invoke.texi: Ditto.
12046 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12048 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12050 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12052 * config/i386/sse.md (*mov<mode>_internal): Guard
12053 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12055 PR rtl-optimization/59647
12056 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12057 new_rtx into UNSIGNED_FLOAT rtxes.
12059 Copyright (C) 2014 Free Software Foundation, Inc.
12061 Copying and distribution of this file, with or without modification,
12062 are permitted in any medium without royalty provided the copyright
12063 notice and this notice are preserved.