1 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
3 * config/rs6000/rs6000.md (type): Reorder, reformat.
5 2014-05-23 Martin Jambor <mjambor@suse.cz>
7 PR tree-optimization/53787
8 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
9 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
10 analysis_done, update all uses.
11 * ipa-prop.c: Include domwalk.h
12 (param_analysis_info): Removed.
13 (param_aa_status): New type.
14 (ipa_bb_info): Likewise.
15 (func_body_info): Likewise.
16 (ipa_get_bb_info): New function.
17 (aa_overwalked): Likewise.
18 (find_dominating_aa_status): Likewise.
19 (parm_bb_aa_status_for_bb): Likewise.
20 (parm_preserved_before_stmt_p): Changed to use new param AA info.
21 (load_from_unmodified_param): Accept func_body_info as a parameter
22 instead of parms_ainfo.
23 (parm_ref_data_preserved_p): Changed to use new param AA info.
24 (parm_ref_data_pass_through_p): Likewise.
25 (ipa_load_from_parm_agg_1): Likewise. Update callers.
26 (compute_complex_assign_jump_func): Changed to use new param AA info.
27 (compute_complex_ancestor_jump_func): Likewise.
28 (ipa_compute_jump_functions_for_edge): Likewise.
29 (ipa_compute_jump_functions): Removed.
30 (ipa_compute_jump_functions_for_bb): New function.
31 (ipa_analyze_indirect_call_uses): Likewise, moved variable
33 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
34 and info, moved variable declarations down.
35 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
37 (ipa_analyze_stmt_uses): Likewise.
38 (ipa_analyze_params_uses): Removed.
39 (ipa_analyze_params_uses_in_bb): New function.
40 (ipa_analyze_controlled_uses): Likewise.
41 (free_ipa_bb_info): Likewise.
42 (analysis_dom_walker): New class.
43 (ipa_analyze_node): Handle node-specific forbidden analysis,
44 initialize and free func_body_info, use dominator walker.
45 (ipcp_modif_dom_walker): New class.
46 (ipcp_transform_function): Create and free func_body_info, use
47 ipcp_modif_dom_walker, moved a lot of functionality there.
49 2014-05-23 Marek Polacek <polacek@redhat.com>
50 Jakub Jelinek <jakub@redhat.com>
52 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
53 * gcc.c (sanitize_spec_function): Likewise.
54 * convert.c (convert_to_integer): Include "ubsan.h". Add
55 floating-point to integer instrumentation.
56 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
57 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
59 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
60 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
61 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
62 * ubsan.c: Include "realmpfr.h" and "dfp.h".
63 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
64 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
65 float/double/long double.
66 (ubsan_instrument_float_cast): New function.
67 * ubsan.h (ubsan_instrument_float_cast): Declare.
69 2014-05-23 Jiong Wang <jiong.wang@arm.com>
71 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
73 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
74 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
75 Adjust for tailcalling through registers.
76 * config/aarch64/aarch64.h (enum reg_class): New caller save
78 (REG_CLASS_NAMES): Likewise.
79 (REG_CLASS_CONTENTS): Likewise.
80 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
81 Allow tailcalling without decls.
83 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
85 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
86 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
88 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
89 gsi, and variables v_* to v*.
91 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
93 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
95 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
97 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
98 * omp-low.c: Update accordingly.
100 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
101 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
102 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
103 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
104 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
105 GF_OMP_TARGET_KIND_UPDATE.
107 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
108 Explicitly enumerate the expected region types.
110 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
113 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
114 documentation; the old documentation didn't clearly state the
115 constraints on the contents of the pointed-to storage.
117 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
119 Fix bootstrap error on ia64
120 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
121 Return default value.
123 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
125 PR tree-optimization/54733
126 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
128 (find_bswap_or_nop_load): New.
129 (find_bswap_1): Renamed to ...
130 (find_bswap_or_nop_1): This. Also add support for memory source.
131 (find_bswap): Renamed to ...
132 (find_bswap_or_nop): This. Also add support for memory source and
133 detection of bitwise operations equivalent to load in host endianness.
134 (execute_optimize_bswap): Likewise. Also move its leading comment back
135 in place and split statement transformation into ...
136 (bswap_replace): This.
138 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
140 PR rtl-optimization/61215
141 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
142 simplify_gen_subreg until final substitution.
144 2014-05-23 Alan Modra <amodra@gmail.com>
147 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
148 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
149 Use "Y" constraint rather than "m".
151 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
153 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
155 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
156 New function declaration.
157 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
158 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
159 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
160 (aarch64_init_builtins) : Initialize builtins
161 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
162 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
163 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
164 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
165 and __builtins_aarch64_set_fpsr.
166 (aarch64_atomic_assign_expand_fenv): New function.
167 * config/aarch64/aarch64.md (set_fpcr): New pattern.
168 (get_fpcr) : Likewise.
169 (set_fpsr) : Likewise.
170 (get_fpsr) : Likewise.
171 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
172 and UNSPECV_SET_FPSR.
173 * doc/extend.texi (AARCH64 Built-in Functions) : Document
174 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
175 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
177 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
179 PR rtl-optimization/60969
180 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
181 constraints. Set up mem cost for NO_REGS case.
183 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
185 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
187 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
189 * config/darwin.c: Include "lto-section-names.h".
190 (LTO_SEGMENT_NAME): Don't define.
191 * config/i386/winnt.c: Include "lto-section-names.h".
192 * lto-streamer.c: Include "lto-section-names.h".
193 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
194 * lto-wrapper.c: Include "lto-section-names.h".
195 (LTO_SECTION_NAME_PREFIX): Don't define.
196 * lto-section-names.h: New file.
197 * cgraphunit.c: Include "lto-section-names.h".
199 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
201 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
203 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
206 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
208 2014-05-22 Nick Clifton <nickc@redhat.com>
210 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted
213 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
215 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
216 -> (T)A transformation to integer types.
218 2014-05-22 Teresa Johnson <tejohnson@google.com>
220 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
221 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
222 (gcov_rewrite): Use gcov_nonruntime_assert.
224 (gcov_write_words): Ditto.
225 (gcov_write_length): Ditto.
226 (gcov_read_words): Use gcov_nonruntime_assert, and remove
227 gcc_assert from IN_LIBGCOV code.
228 (gcov_read_summary): Use gcov_error to flag profile corruption.
229 (gcov_sync): Use gcov_nonruntime_assert.
230 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
231 (gcov_histo_index): Use gcov_nonruntime_assert.
232 (static void gcov_histogram_merge): Ditto.
233 (compute_working_sets): Ditto.
234 * gcov-io.h (gcov_nonruntime_assert): Define.
235 (gcov_error): Define for !IN_LIBGCOV
237 2014-05-22 Richard Biener <rguenther@suse.de>
239 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
240 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
241 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
242 and deallocation site.
243 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
244 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
245 passing through the incoming points-to set.
246 (handle_lhs_call): Use flags argument instead of recomputing it.
247 (find_func_aliases_for_call): Call handle_lhs_call with proper
250 2014-05-22 Jakub Jelinek <jakub@redhat.com>
252 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
253 all padding bits in REAL_VALUE_TYPE are cleared.
255 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
257 Cleanup and improve multipass_dfa_lookahead_guard
258 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
259 (core2i7_first_cycle_multipass_begin,)
260 (core2i7_first_cycle_multipass_issue,)
261 (core2i7_first_cycle_multipass_backtrack): Update signature.
263 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
264 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
265 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
267 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
268 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
270 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
272 * doc/tm.texi: Regenerate.
274 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
275 * haifa-sched.c (ready_try): Make signed to allow negative values.
276 (rebug_ready_list_1): Update.
277 (choose_ready): Simplify.
278 (sched_extend_ready_list): Update.
280 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
282 Remove IA64 speculation tweaking flags
283 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
284 speculation tuning flags.
285 (msched-prefer-non-data-spec-insns,)
286 (msched-prefer-non-control-spec-insns): Obsolete options.
287 * haifa-sched.c (choose_ready): Remove handling of
288 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
289 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
290 and PREFER_NON_DATA_SPEC.
291 * sel-sched.c (process_spec_exprs): Remove handling of
292 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
294 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
296 Improve scheduling debug output
297 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
298 (advance_one_cycle): Update.
299 (schedule_insn, queue_to_ready): Add debug printouts.
300 (debug_ready_list_1): New static function.
301 (debug_ready_list): Update.
302 (max_issue): Add debug printouts.
303 (dump_insn_stream): New static function.
304 (schedule_block): Use it. Also better indent printouts.
306 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
308 Fix sched_insn debug counter
309 * haifa-sched.c (schedule_insn): Update.
310 (struct haifa_saved_data): Add nonscheduled_insns_begin.
311 (save_backtrack_point, restore_backtrack_point): Update.
312 (first_nonscheduled_insn): New static function.
313 (queue_to_ready, choose_ready): Use it.
314 (schedule_block): Init nonscheduled_insns_begin.
315 (sched_emit_insn): Update.
318 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
320 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
322 (aarch64_secondary_reload) : LikeWise.
323 (aarch64_class_max_nregs) : Remove CORE_REGS.
324 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
325 (REG_CLASS_NAMES) : Likewise.
326 (REG_CLASS_CONTENTS) : LikeWise.
327 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
329 2014-05-21 Guozhi Wei <carrot@google.com>
332 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
334 (vqdmulhq_n_s16): Likewise.
336 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
338 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
340 2014-05-21 Marek Polacek <polacek@redhat.com>
343 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
345 2014-05-21 Martin Jambor <mjambor@suse.cz>
347 * doc/invoke.texi (Optimize Options): Document parameters
348 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
349 ipa-cp-array-index-hint-bonus.
351 2014-05-21 Mark Wielaard <mjw@redhat.com>
354 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
355 version >= 3 or not strict DWARF.
356 * langhooks.h (struct lang_hooks_for_types): Add
357 enum_underlying_base_type.
358 * langhooks.c (lhd_enum_underlying_base_type): New function.
359 * gcc/langhooks.h (struct lang_hooks_for_types): Add
360 enum_underlying_base_type.
361 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
362 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
363 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
365 2014-05-21 Richard Biener <rguenther@suse.de>
367 * doc/invoke.texi (-flto-partition=): Document one and
370 2014-05-21 John Marino <gnugcc@marino.st>
372 * config.gcc (*-*-dragonfly*): New target.
373 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
374 * configure: Regenerate.
375 * config/dragonfly-stdint.h: New.
376 * config/dragonfly.h: New.
377 * config/dragonfly.opt: New.
378 * config/i386/dragonfly.h: New.
379 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
381 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
383 * tree.def (VOID_CST): New.
384 * tree-core.h (TI_VOID): New.
385 * tree.h (void_node): New.
386 * tree.c (tree_node_structure_for_code, tree_code_size)
387 (iterative_hash_expr): Handle VOID_CST.
388 (build_common_tree_nodes): Initialize void_node.
390 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
392 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
394 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
396 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
399 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
400 flag_reorder_blocks_and_partition.
401 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
403 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
406 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
408 (*addc_r_t): Add new insn_and_split.
410 2014-05-21 Jakub Jelinek <jakub@redhat.com>
413 * omp-low.c (handle_simd_reference): New function.
414 (lower_rec_input_clauses): Use it. Defer adding reference
415 initialization even for reduction without placeholder if in simd,
416 handle it properly later on.
418 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
420 PR tree-optimization/60899
421 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
422 assume all static symbols will have definition wile parsing and
423 check the do have definition later in compilation; check that
424 variable referring symbol will be output before concluding that
425 reference is safe; be conservative for referring local statics;
426 be more precise about when comdat is output in other partition.
428 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
431 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED parameter.
432 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
433 (ipa_inline): Loop inline_to_all_callers until no more aliases are removed.
435 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
437 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
438 set writeonly flag only for vars actually written to.
440 2014-05-20 Dehao Chen <dehao@google.com>
442 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
443 and callee count to get clone count.
444 * tree-inline.c (expand_call_inline): Use callee count instead of bb
447 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
449 PR rtl-optimization/61243
450 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
452 2014-05-20 Xinliang David Li <davidxl@google.com>
454 * cgraphunit.c (walk_polymorphic_call_targets): Add
455 dbgcnt and fopt-info support.
456 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
457 * ipa-devirt.c (ipa_devirt): Ditto.
458 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
460 * ipa.c (walk_polymorphic_call_targets): Ditto.
461 * gimple-fold.c (fold_gimple_assign): Ditto.
462 (gimple_fold_call): Ditto.
463 * dbgcnt.def: New counter.
465 2014-05-20 DJ Delorie <dj@redhat.com>
467 * config/msp430/msp430.md (split): Don't allow subregs when
468 splitting SImode adds.
469 (andneghi): Fix subtraction logic.
470 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
472 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
474 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
476 * except.c (switch_to_exception_section, resolve_unique_section,
477 get_named_text_section, default_function_rodata_section,
478 align_variable, get_block_for_decl, default_section_type_flags):
479 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
480 * symtab.c (symtab_add_to_same_comdat_group,
481 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
482 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
484 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
485 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
486 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
487 (c6x_function_in_section_p): Likewise.
488 * config/darwin.c (machopic_select_section): Likewise.
489 * config/arm/arm.c (arm_function_in_section_p): Likewise.
490 * config/mips/mips.c (mips_function_rodata_section): Likewise.
491 * config/mep/mep.c (mep_select_section): LIkewise.
492 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
494 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
496 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
497 EH region of calls to pure functions that can throw an exception.
498 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
499 (copy_reference_ops_from_call): Also copy the EH region of the call if
500 it can throw an exception.
502 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
504 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
505 nested VEC_SELECTs that are inverses of each other.
507 2014-05-20 Richard Biener <rguenther@suse.de>
509 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
511 (extract_and_process_scc_for_name): not here.
512 (cond_dom_walker::before_dom_children): Only process
513 stmts that end the BB in interesting ways.
514 (run_scc_vn): Mark param uses as visited.
516 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
518 * config/arm/arm.md (arith_shiftsi): Do not predicate for
521 2014-05-20 Nick Clifton <nickc@redhat.com>
523 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
524 (msp430_gimplify_va_arg_expr): New function.
525 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
527 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
528 operand 0 in order to prevent confusion about the number of
531 2014-05-20 Richard Biener <rguenther@suse.de>
533 PR tree-optimization/61221
534 * tree-ssa-pre.c (el_to_update): Remove.
535 (eliminate_dom_walker::before_dom_children): Handle released
536 VDEFs by value-numbering them to the associated VUSE. Update
537 stmt immediately for substituted call address.
538 (eliminate): Remove delayed stmt updating code.
539 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
540 possibly late re-numbered vuses.
541 (vn_reference_lookup_2): Adjust.
542 (vn_reference_lookup_pieces): Likewise.
543 (vn_reference_lookup): Likewise.
545 2014-05-20 Richard Biener <rguenther@suse.de>
547 * config.gcc: Remove need_64bit_hwint.
548 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
549 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
551 * config.in: Regenerate.
552 * configure: Likewise.
554 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
556 * doc/extend.texi: Create Label Attributes section,
557 move all label attributes into it and reference it.
559 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
561 * arm.c (thumb1_reorg): When scanning backwards skip anything
562 that's not a proper insn.
564 2014-05-19 Richard Biener <rguenther@suse.de>
566 PR tree-optimization/61221
567 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
568 Do nothing for unreachable blocks.
569 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
570 Improve unreachability detection.
572 2014-05-19 Richard Biener <rguenther@suse.de>
574 PR tree-optimization/61209
575 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
577 2014-05-19 Nick Clifton <nickc@redhat.com>
579 * except.c (init_eh): Fix computation of builtin setjmp buffer
580 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
582 2014-05-19 Richard Biener <rguenther@suse.de>
584 PR tree-optimization/61184
585 * tree-vrp.c (is_negative_overflow_infinity): Use
586 TREE_OVERFLOW_P and do that check first.
587 (is_positive_overflow_infinity): Likewise.
588 (is_overflow_infinity): Likewise.
589 (vrp_operand_equal_p): Properly treat operands with
590 differing overflow as not equal.
592 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
594 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
595 shift simplification where it was intended.
597 2014-05-19 Christian Bruel <christian.bruel@st.com>
600 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
602 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
605 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
608 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
610 * reg-notes.def (CROSSING_JUMP): Likewise.
611 * rtl.h (rtx_def): Update comment for jump flag.
612 (CROSSING_JUMP_P): Define.
613 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
614 of a REG_CROSSING_JUMP note.
615 * cfghooks.c (tidy_fallthru_edges): Likewise.
616 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
617 * emit-rtl.c (try_split): Likewise.
618 * haifa-sched.c (sched_create_recovery_edges): Likewise.
619 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
620 * jump.c (redirect_jump_2): Likewise.
621 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
622 (relax_delay_slots): Likewise.
623 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
625 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
626 * config/sh/sh.md (jump_compact): Likewise.
627 * bb-reorder.c (rotate_loop): Likewise.
628 (pass_duplicate_computed_gotos::execute): Likewise.
629 (add_reg_crossing_jump_notes): Rename to...
630 (update_crossing_jump_flags): ...this.
631 (pass_partition_blocks::execute): Update accordingly.
633 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
635 * tree.h: Remove extraneous template <>.
637 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
639 * ipa.c (symtab_remove_unreachable_nodes): Remove
640 symbol from comdat group if its body was eliminated.
641 (comdat_can_be_unshared_p_1): Static symbols can always
643 * symtab.c (symtab_remove_from_same_comdat_group): Break out
645 (symtab_unregister_node): ... this one.
646 (verify_symtab_base): More strict checking of comdats.
647 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
649 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
651 * tree-pass.h (make_pass_ipa_comdats): New pass.
652 * timevar.def (TV_IPA_COMDATS): New timevar.
653 * passes.def (pass_ipa_comdats): Add.
654 * Makefile.in (OBJS): Add ipa-comdats.o
655 * ipa-comdats.c: New file.
657 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
659 * ipa.c (update_visibility_by_resolution_info): New function.
660 (function_and_variable_visibility): Use it.
662 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
664 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
666 (FOR_EACH_DEFINED_SYMBOL): New macro.
667 (varpool_first_static_initializer, varpool_next_static_initializer,
668 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
669 (symtab_in_same_comdat_p): Correctly deal with inline functions.
671 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
673 * ggc-page.c (ggc_handle_finalizers): Add comment.
675 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
677 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
678 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
679 (ggc_internal_cleared_alloc): Likewise.
680 * ggc-page.c (finalizer): New class.
681 (vec_finalizer): Likewise.
682 (globals::finalizers): New member.
683 (globals::vec_finalizers): Likewise.
684 (ggc_internal_alloc): Record the finalizer if any for the block being
686 (ggc_handle_finalizers): New function.
687 (ggc_collect): Call ggc_handle_finalizers.
688 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
690 (ggc_internal_cleared_alloc): Likewise.
691 (finalize): New function.
692 (need_finalization_p): Likewise.
693 (ggc_alloc): Install the type's destructor as the finalizer if it
695 (ggc_cleared_alloc): Likewise.
696 (ggc_vec_alloc): Likewise.
697 (ggc_cleared_vec_alloc): Likewise.
699 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
701 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
703 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
705 * alias.c (record_alias_subset): Adjust.
706 * bitmap.c (bitmap_element_allocate): Likewise.
707 (bitmap_gc_alloc_stat): Likewise.
708 * cfg.c (init_flow): Likewise.
709 (alloc_block): Likewise.
710 (unchecked_make_edge): Likewise.
711 * cfgloop.c (alloc_loop): Likewise.
712 (flow_loops_find): Likewise.
713 (rescan_loop_exit): Likewise.
714 * cfgrtl.c (init_rtl_bb_info): Likewise.
715 * cgraph.c (insert_new_cgraph_node_version): Likewise.
716 (cgraph_allocate_node): Likewise.
717 (cgraph_create_edge_1): Likewise.
718 (cgraph_allocate_init_indirect_info): Likewise.
719 * cgraphclones.c (cgraph_clone_edge): Likewise.
720 * cgraphunit.c (add_asm_node): Likewise.
721 (init_lowered_empty_function): Likewise.
722 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
723 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
724 (alpha_use_linkage): Likewise.
725 * config/arc/arc.c (arc_init_machine_status): Likewise.
726 * config/arm/arm.c (arm_init_machine_status): Likewise.
727 * config/avr/avr.c (avr_init_machine_status): Likewise.
728 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
729 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
730 * config/cris/cris.c (cris_init_machine_status): Likewise.
731 * config/darwin.c (machopic_indirection_name): Likewise.
732 (darwin_build_constant_cfstring): Likewise.
733 (darwin_enter_string_into_cfstring_table): Likewise.
734 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
735 * config/frv/frv.c (frv_init_machine_status): Likewise.
736 * config/i386/i386.c (get_dllimport_decl): Likewise.
737 (ix86_init_machine_status): Likewise.
738 (assign_386_stack_local): Likewise.
739 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
740 (i386_pe_maybe_record_exported_symbol): Likewise.
741 (i386_pe_record_stub): Likewise.
742 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
743 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
744 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
745 (m32c_note_pragma_address): Likewise.
746 * config/mep/mep.c (mep_init_machine_status): Likewise.
747 (mep_note_pragma_flag): Likewise.
748 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
749 (mips16_local_alias): Likewise.
750 (mips_init_machine_status): Likewise.
751 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
752 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
753 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
754 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
755 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
756 * config/pa/pa.c (pa_init_machine_status): Likewise.
757 (pa_get_deferred_plabel): Likewise.
758 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
759 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
760 (rs6000_init_machine_status): Likewise.
761 (output_toc): Likewise.
762 * config/s390/s390.c (s390_init_machine_status): Likewise.
763 * config/score/score.c (score_output_external): Likewise.
764 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
765 * config/spu/spu.c (spu_init_machine_status): Likewise.
766 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
767 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
768 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
769 * coverage.c (coverage_end_function): Likewise.
770 * dbxout.c (dbxout_init): Likewise.
771 * doc/gty.texi: Don't mention variable_size attribute.
772 * dwarf2cfi.c (new_cfi): Adjust.
773 (new_cfi_row): Likewise.
774 (copy_cfi_row): Likewise.
775 (create_cie_data): Likewise.
776 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
777 (new_loc_descr): Likewise.
778 (find_AT_string_in_table): Likewise.
779 (add_addr_table_entry): Likewise.
781 (add_var_loc_to_decl): Likewise.
782 (clone_die): Likewise.
783 (clone_as_declaration): Likewise.
784 (break_out_comdat_types): Likewise.
785 (new_loc_list): Likewise.
786 (add_loc_descr_to_each): Likewise.
787 (add_location_or_const_value_attribute): Likewise.
788 (add_linkage_name): Likewise.
789 (lookup_filename): Likewise.
790 (dwarf2out_var_location): Likewise.
791 (new_line_info_table): Likewise.
792 (dwarf2out_init): Likewise.
793 (mem_loc_descriptor): Likewise.
794 (loc_descriptor): Likewise.
795 (add_const_value_attribute): Likewise.
796 (tree_add_const_value_attribute): Likewise.
797 (comp_dir_string): Likewise.
798 (dwarf2out_vms_debug_main_pointer): Likewise.
799 (string_cst_pool_decl): Likewise.
800 * emit-rtl.c (set_mem_attrs): Likewise.
801 (get_reg_attrs): Likewise.
802 (start_sequence): Likewise.
803 (init_emit): Likewise.
804 (init_emit_regs): Likewise.
805 * except.c (init_eh_for_function): Likewise.
806 (gen_eh_region): Likewise.
807 (gen_eh_region_catch): Likewise.
808 (gen_eh_landing_pad): Likewise.
809 (add_call_site): Likewise.
810 * function.c (add_frame_space): Likewise.
811 (insert_temp_slot_address): Likewise.
812 (assign_stack_temp_for_type): Likewise.
813 (get_hard_reg_initial_val): Likewise.
814 (allocate_struct_function): Likewise.
815 (prepare_function_start): Likewise.
816 (types_used_by_var_decl_insert): Likewise.
817 * gengtype.c (variable_size_p): Remove function.
818 (enum alloc_quantity): Remove enum.
819 (write_typed_alloc_def): Remove function.
820 (write_typed_struct_alloc_def): Likewise.
821 (write_typed_typedef_alloc_def): Likewise.
822 (write_typed_alloc_defns): Likewise.
824 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
825 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
826 * ggc.h (ggc_alloc): new function.
827 (ggc_cleared_alloc): Likewise.
828 (ggc_vec_alloc): Template on type of vector element, and remove
829 element size argument.
830 (ggc_cleared_vec_alloc): Likewise.
831 * gimple.c (gimple_build_omp_for): Adjust.
832 (gimple_copy): Likewise.
833 * ipa-cp.c (get_replacement_map): Likewise.
834 (find_aggregate_values_for_callers_subset): Likewise.
835 (known_aggs_to_agg_replacement_list): Likewise.
836 * ipa-devirt.c (get_odr_type): Likewise.
837 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
838 (read_agg_replacement_chain): Likewise.
839 * loop-iv.c (get_simple_loop_desc): Likewise.
840 * lto-cgraph.c (input_node_opt_summary): Likewise.
841 * lto-section-in.c (lto_new_in_decl_state): Likewise.
842 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
843 (input_eh_region): Likewise.
844 (input_eh_lp): Likewise.
845 (input_cfg): Likewise.
846 * optabs.c (set_optab_libfunc): Likewise.
847 (init_tree_optimization_optabs): Likewise.
848 (set_conv_libfunc): Likewise.
849 * passes.c (do_per_function_toporder): Likewise.
850 * rtl.h: Don't use variable_size gty attribute.
851 * sese.c (if_region_set_false_region): Adjust.
852 * stringpool.c (gt_pch_save_stringpool): Likewise.
853 * target-globals.c (save_target_globals): Likewise.
854 * toplev.c (general_init): Likewise.
855 * trans-mem.c (record_tm_replacement): Likewise.
856 (split_bb_make_tm_edge): Likewise.
857 * tree-cfg.c (move_sese_region_to_fn): Likewise.
858 * tree-data-ref.h (lambda_vector_new): Likewise.
859 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
860 * tree-iterator.c (tsi_link_before): Likewise.
861 (tsi_link_after): Likewise.
862 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
863 * tree-ssa-loop-niter.c (record_estimate): Likewise.
864 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
865 * tree-ssa-operands.h: Don't use variable_size gty attribute.
866 * tree-ssa.c (init_tree_ssa): Adjust.
867 * tree-ssanames.c (set_range_info): Likewise.
868 (get_ptr_info): Likewise.
869 (duplicate_ssa_name_ptr_info): Likewise.
870 (duplicate_ssa_name_range_info): Likewise.
871 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
872 (unpack_ts_fixed_cst_value_fields): Likewise.
873 * tree.c (build_fixed): Likewise.
874 (build_real): Likewise.
875 (build_string): Likewise.
876 (decl_priority_info): Likewise.
877 (decl_debug_expr_insert): Likewise.
878 (decl_value_expr_insert): Likewise.
879 (decl_debug_args_insert): Likewise.
880 (type_hash_add): Likewise.
881 (build_omp_clause): Likewise.
882 * ubsan.c (decl_for_type_insert): Likewise.
883 * varasm.c (get_unnamed_section): Likewise.
884 (get_noswitch_section): Likewise.
885 (get_section): Likewise.
886 (get_block_for_section): Likewise.
887 (create_block_symbol): Likewise.
888 (build_constant_desc): Likewise.
889 (create_constant_pool): Likewise.
890 (force_const_mem): Likewise.
891 (record_tm_clone_pair): Likewise.
892 * varpool.c (varpool_create_empty_node): Likewise.
894 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
896 * dwarf2out.c (tree_add_const_value_attribute): Call
897 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
898 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
899 instead of ggc_internal_<x>alloc_stat.
900 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
901 (ggc_realloc): Likewise.
902 * ggc-none.c (ggc_internal_alloc): Likewise.
903 (ggc_internal_cleared_alloc): Likewise.
904 * ggc-page.c: Likewise.
905 * ggc.h (ggc_internal_alloc_stat): Likewise.
906 (ggc_internal_alloc): Remove macro.
907 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
908 (ggc_internal_cleared_alloc): Remove macro.
909 (GGC_RESIZEVEC): Adjust.
910 (ggc_resizevar): Remove macro.
911 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
912 (ggc_internal_cleared_vec_alloc_stat): Likewise.
913 (ggc_internal_vec_cleared_alloc): Remove macro.
914 (ggc_alloc_atomic_stat): Drop _stat suffix.
915 (ggc_alloc_atomic): Remove macro.
916 (ggc_alloc_cleared_atomic): Remove macro.
917 (ggc_alloc_string_stat): Drop _stat suffix.
918 (ggc_alloc_string): Remove macro.
919 (ggc_alloc_rtx_def_stat): Adjust.
920 (ggc_alloc_tree_node_stat): Likewise.
921 (ggc_alloc_cleared_tree_node_stat): Likewise.
922 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
923 (ggc_alloc_cleared_simd_clone_stat): Likewise.
924 * gimple.c (gimple_build_omp_for): Likewise.
925 (gimple_copy): Likewise.
926 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
927 * toplev.c (realloc_for_line_map): Adjust.
928 * tree-data-ref.h (lambda_vector_new): Likewise.
929 * tree-phinodes.c (allocate_phi_node): Likewise.
930 * tree.c (grow_tree_vec_stat): Likewise.
931 * vec.h (va_gc::reserve): Adjust.
933 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
935 * config/microblaze/microblaze.c (break_handler): New Declaration.
936 (microblaze_break_function_p,microblaze_is_break_handler): New.
937 (compute_frame_size): Use microblaze_break_function_p.
938 Add the test of break_handler.
939 (microblaze_function_prologue) : Add the test of variable
940 break_handler. Check the fnname by BREAK_HANDLER_NAME.
941 (microblaze_function_epilogue) : Add the test of break_handler.
942 (microblaze_globalize_label) : Add the test of break_handler.
943 Check the name by BREAK_HANDLER_NAME.
945 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
947 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
948 microblaze_is_break_handler test.
949 (call_internal1,call_value_intern): Use microblaze_break_function_p.
952 * config/microblaze/microblaze-protos.h
953 (microblaze_break_function_p,microblaze_is_break_handler):
956 * doc/extend.texi (MicroBlaze break_handler Functions): Document
957 new MicroBlaze break_handler functions.
959 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
961 * doc/extend.texi (Size of an asm): Move node text according
962 to its @menu entry position.
964 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
966 PR tree-optimization/61140
967 PR tree-optimization/61150
968 PR tree-optimization/61197
969 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
971 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
973 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
975 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
977 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
978 __SIZEOF_INT128__ is defined.
980 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
982 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
983 (rs6000_delegitimize_address): Use it.
985 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
987 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
988 inplace argument. Store the new address in the original MEM when true.
989 * emit-rtl.c (change_address_1): Likewise.
990 (adjust_address_1, adjust_automodify_address_1, offset_address):
992 * rtl.h (plus_constant): Add an inplace argument.
993 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
994 when true. Avoid generating (plus X (const_int 0)).
995 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
996 in-place. Pass true to plus_constant.
997 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
999 2014-05-16 Dehao Chen <dehao@google.com>
1001 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1003 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1006 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1008 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1010 2014-05-16 Dehao Chen <dehao@google.com>
1012 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1013 optimize_function_for_size_p.
1014 * regs.h (REG_FREQ_FROM_BB): Likewise.
1016 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1019 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1020 negt_reg_operand cases.
1021 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1023 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1025 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1027 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1030 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1033 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1035 * tree-cfg.c (dump_function_to_file): Dump the return type of
1036 functions, in a line to itself before the function body, mimicking
1037 the layout of a C function.
1039 2014-05-16 Dehao Chen <dehao@google.com>
1041 * cfghooks.c (make_forwarder_block): Use direct computation to
1042 get fall-through edge's count and frequency.
1044 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1046 * config/arc/arc.c (arc_init): Fix typo in error message.
1047 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1048 (split_stack_prologue_scratch_regno): Likewise.
1049 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1050 word from error message.
1052 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1054 * ira-costs.c: Fix typo in comment.
1056 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1058 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1060 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1062 * varpool.c (dump_varpool_node): Dump write-only flag.
1063 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1065 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1066 write-only variables.
1067 * ipa.c (process_references): New function.
1068 (set_readonly_bit): New function.
1069 (set_writeonly_bit): New function.
1070 (clear_addressable_bit): New function.
1071 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1072 fix handling of aliases.
1073 * cgraph.h (struct varpool_node): Add writeonly flag.
1075 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1077 PR rtl-optimization/60969
1078 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1079 Calculate costs for this case.
1081 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1083 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1084 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1086 2014-05-16 Richard Biener <rguenther@suse.de>
1088 PR tree-optimization/61194
1089 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1090 bool patterns ending in a COND_EXPR.
1092 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1094 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1096 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1098 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1099 where we were unable to cost an RTX.
1101 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1103 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1106 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1107 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1109 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1111 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1112 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1114 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1115 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1117 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1118 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1120 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1123 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1124 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1126 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1129 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1130 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1132 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1133 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1135 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1136 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1138 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1141 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1142 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1144 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1145 ZERO_EXTEND and SIGN_EXTEND better.
1147 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1148 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1150 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1153 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1154 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1156 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1157 costs when costing loads and stores to memory.
1159 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1160 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1162 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1165 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1167 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1169 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1170 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1172 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1174 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1176 (aarch64_rtx_mult_cost): New.
1177 (aarch64_rtx_costs): Use it, refactor as appropriate.
1179 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1180 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1182 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1183 emit instructions, return number of instructions which would
1185 (aarch64_add_constant): Update call to aarch64_build_constant.
1186 (aarch64_output_mi_thunk): Likewise.
1187 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1190 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1192 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1193 (TARGET_RTX_COSTS): Call it.
1195 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1197 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1198 (cortexa57_vector_cost): Likewise.
1199 (cortexa57_tunings): Use them.
1201 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1203 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1204 (cpu_addrcost_table): Use it.
1205 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1206 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1209 2014-05-16 Richard Biener <rguenther@suse.de>
1211 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1212 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1213 (visit_phi): Ignore edges marked as not executable.
1214 (class cond_dom_walker): New.
1215 (cond_dom_walker::before_dom_children): Value-number
1216 control statements and mark successor edges as not
1217 executable if possible.
1218 (run_scc_vn): First walk all control statements in
1219 dominator order, marking edges as not executable.
1220 * tree-inline.c (copy_edges_for_bb): Be not confused
1221 about random edge flags.
1223 2014-05-16 Richard Biener <rguenther@suse.de>
1225 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1227 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1230 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1231 (__TM_simple_begin): Use it.
1232 (__TM_begin): Likewise.
1234 2014-05-15 Martin Jambor <mjambor@suse.cz>
1237 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1238 type_preserved flag when the indirect edge is polymorphic.
1240 2014-05-15 Martin Jambor <mjambor@suse.cz>
1242 PR tree-optimization/61090
1243 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1244 build_ref_for_model.
1246 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1248 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1249 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1251 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1253 PR tree-optimization/61158
1254 * fold-const.c (fold_binary_loc): If X is zero-extended and
1255 shiftc >= prec, make sure zerobits is all ones instead of
1256 invoking undefined behavior.
1258 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1260 * regcprop.h: New file.
1261 * regcprop.c (skip_debug_insn_p): New decl.
1262 (replace_oldest_value_reg): Check skip_debug_insn_p.
1263 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1264 * shrink-wrap.c: Include regcprop.h.
1265 (prepare_shrink_wrap): Call
1266 copyprop_hardreg_forward_bb_without_debug_insn.
1268 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1270 * shrink-wrap.h: Update comment.
1271 * shrink-wrap.c: Update comment.
1272 (next_block_for_reg): Rename to live_edge_for_reg.
1273 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1274 (move_insn_for_shrink_wrap): Split live_edge.
1275 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1277 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1279 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1281 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1282 * config/sparc/sparc.md (fptype_ut699): New attribute.
1283 (in_branch_delay): Return false if -mfix-ut699 is specified and
1284 fptype_ut699 is set to single.
1285 (truncdfsf2): Add fptype_ut699 attribute.
1286 (fix_truncdfsi2): Likewise.
1287 (floatsisf2): Change fptype attribute.
1288 (fix_truncsfsi2): Likewise.
1289 (negtf2_notv9): Delete.
1290 (negtf2_v9): Likewise.
1291 (negtf2_hq): New instruction.
1292 (negtf2): New instruction and splitter.
1293 (negdf2_notv9): Rewrite.
1294 (abstf2_notv9): Delete.
1295 (abstf2_hq_v9): Likewise.
1296 (abstf2_v9): Likewise.
1297 (abstf2_hq): New instruction.
1298 (abstf2): New instruction and splitter.
1299 (absdf2_notv9): Rewrite.
1301 2014-05-14 Cary Coutant <ccoutant@google.com>
1304 * opts.c (common_handle_option): Don't special-case "-g".
1305 (set_debug_level): Default to at least level 2 with "-g".
1307 2014-05-14 DJ Delorie <dj@redhat.com>
1309 * config/msp430/msp430.c (msp430_builtin): Add
1310 MSP430_BUILTIN_DELAY_CYCLES.
1311 (msp430_init_builtins): Register void __delay_cycles(long long).
1312 (msp430_builtin_decl): Add it.
1313 (cg_magic_constant): New.
1314 (msp430_expand_delay_cycles): New.
1315 (msp430_expand_builtin): Call it.
1316 (msp430_print_operand_raw): Change integer printing from "int" to
1318 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1319 (delay_cycles_start): New.
1320 (delay_cycles_end): New.
1321 (delay_cycles_32): New.
1322 (delay_cycles_32x): New.
1323 (delay_cycles_16): New.
1324 (delay_cycles_16x): New.
1325 (delay_cycles_2): New.
1326 (delay_cycles_1): New.
1327 * doc/extend.texi: Document __delay_cycles().
1329 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1331 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1332 length attribute computation.
1334 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1337 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1339 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1342 * config/sparc/sparc.md: Fix types of low and high in DI constant
1343 splitter. Use gen_int_mode in some other splitters.
1345 2014-05-14 Martin Jambor <mjambor@suse.cz>
1348 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1350 2014-05-14 James Norris <jnorris@codesourcery.com>
1352 * omp-low.c (expand_parallel_call): Remove shadow variable.
1353 (expand_omp_taskreg): Likewise.
1355 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1357 * common/config/i386/i386-common.c
1358 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1359 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1360 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1361 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1362 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1363 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1364 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1365 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1366 xsavecintrin.h, xsavesintrin.h.
1367 (x86_64-*-*): Ditto.
1368 * config/i386/clflushoptintrin.h: New.
1369 * config/i386/xsavecintrin.h: Ditto.
1370 * config/i386/xsavesintrin.h: Ditto.
1371 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1372 (bit_XSAVES): Ditto.
1373 (bit_XSAVES): Ditto.
1374 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1375 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1377 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1378 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1379 OPTION_MASK_ISA_XSAVES.
1380 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1382 (PTA_CLFLUSHOPT) Define.
1383 (PTA_XSAVEC): Ditto.
1384 (PTA_XSAVES): Ditto.
1385 (ix86_option_override_internal): Handle new options.
1386 (ix86_valid_target_attribute_inner_p): Ditto.
1387 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1388 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1389 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1390 (bdesc_special_args): Add __builtin_ia32_xsaves,
1391 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1392 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1393 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1394 (ix86_expand_builtin): Handle new builtins.
1395 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1396 (TARGET_CLFLUSHOPT_P): Ditto.
1397 (TARGET_XSAVEC): Ditto.
1398 (TARGET_XSAVEC_P): Ditto.
1399 (TARGET_XSAVES): Ditto.
1400 (TARGET_XSAVES_P): Ditto.
1401 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1402 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1403 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1405 (ANY_XRSTOR64): Ditto.
1407 (xrstor): Change into <xrstor>.
1408 (xrstor_rex64): Change into <xrstor>_rex64.
1409 (xrstor64): Change into <xrstor>64
1411 * config/i386/i386.opt (mclflushopt): New.
1414 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1416 * doc/invoke.texi: Document new options.
1418 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1420 PR rtl-optimization/60866
1421 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1422 Default it to -1. Pass it down to init_simplejump_data.
1423 (init_simplejump_data): New parameter old_seqno. Pass it down
1424 to get_seqno_for_a_jump.
1425 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1426 initializing new jump seqno as a last resort. Add comment.
1427 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1428 jump and pass it down to sel_init_new_insn.
1429 (sel_redirect_edge_and_branch_force): Likewise.
1431 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1433 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1434 shifted values to avoid build warning.
1436 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1438 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1439 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1440 (cfg_layout_merge_blocks): Likewise.
1441 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1443 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1445 PR rtl-optimization/60901
1446 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1447 bb predecessor belongs to the same scheduling region. Adjust comment.
1449 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1451 * doc/sourcebuild.texi: (dfp_hw): Document.
1452 (p8vector_hw): Likewise.
1453 (powerpc_eabi_ok): Likewise.
1454 (powerpc_elfv2): Likewise.
1455 (powerpc_htm_ok): Likewise.
1456 (ppc_recip_hw): Likewise.
1459 2014-05-13 Cary Coutant <ccoutant@google.com>
1461 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1463 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1465 * gengtype-parse.c (require3): Eliminate in favor of...
1467 (require_template_declaration): Update to support optional single *
1470 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1471 (create_user_defined_type): Handle a single level of explicit
1472 pointerness within template arguments.
1473 (struct write_types_data): Add field "kind".
1474 (filter_type_name): Handle "*" character.
1475 (write_user_func_for_structure_ptr): Require a write_types_data
1476 rather than just a prefix string, so that we can look up the kind
1477 of the wtd and use it as an index into wrote_user_func_for_ptr,
1478 ensuring that such functions are written at most once. Support
1479 subclasses by invoking the marking function of the ultimate base class.
1480 (write_user_func_for_structure_body): Require a write_types_data
1481 rather than just a prefix string, so that we can pass this to
1482 write_user_func_for_structure_ptr.
1483 (write_func_for_structure): Likewise.
1484 (ggc_wtd): Add initializer of new "kind" field.
1485 (pch_wtd): Likewise.
1487 * gengtype.h (enum write_types_kinds): New.
1488 (struct type): Add field wrote_user_func_for_ptr to the "s"
1491 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1493 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1494 instead of const_binop.
1495 (fold_binary_loc): Likewise.
1497 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1499 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1500 calculation to match get_ref_base_and_extent.
1502 2014-05-13 Catherine Moore <clm@codesourcery.com>
1503 Sandra Loosemore <sandra@codesourcery.com>
1505 * configure.ac: Fix assembly for explicit JALR relocation check.
1506 * configure: Regenerate.
1508 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1510 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1511 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1512 Remove associated type declarations and initialisations.
1513 (arm_expand_neon_builtin): Likewise.
1514 (neon_emit_pair_result_insn): Delete.
1515 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1516 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1517 (neon_vzip<mode>): Likewise.
1518 (neon_vuzp<mode>): Likewise.
1520 2014-05-13 Richard Biener <rguenther@suse.de>
1523 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1524 it needs revisiting whether the call still may be tail-called.
1526 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1528 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1529 * rtl.h (block_symbol): Reduce number of fields to 2.
1530 (rtx_def): Add u2.symbol_ref_flags.
1531 (SYMBOL_REF_FLAGS): Use it.
1532 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1533 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1534 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1535 Lower index of SYMBOL_REF_DATA.
1536 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1537 Print SYMBOL_REF_FLAGS at the same time.
1538 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1540 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1542 * rtl.def (VAR_LOCATION): Remove "i" field.
1543 * rtl.h (rtx_def): Add u2.var_location_status.
1544 (PAT_VAR_LOCATION_STATUS): Use it.
1545 (gen_rtx_VAR_LOCATION): Declare.
1546 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1547 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1548 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1550 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1552 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1553 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1555 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1557 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1558 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1559 * rtl.h (rtx_def): Add insn_uid to u2 field.
1560 (RTX_FLAG_CHECK8): Delete in favor of...
1561 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1562 (INSN_DELETED_P): Update accordingly.
1563 (INSN_UID): Use u2.insn_uid.
1564 (INSN_CHAIN_CODE_P): Define.
1565 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1566 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1567 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1568 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1569 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1570 indices accordingly.
1571 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1572 Update indices for insn-chain rtxes.
1573 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1574 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1575 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1576 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1577 * combine.c (try_combine): Likewise.
1578 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1580 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1582 * rtl.def (REG): Remove middle field.
1583 * rtl.h (rtx_def): Add orignal_regno to u2.
1584 (ORIGINAL_REGNO): Use it instead of field 1.
1585 (REG_ATTRS): Lower field index accordingly.
1586 * gengtype.c (adjust_field_rtx_def): Remove handling of
1587 ORIGINAL_REGNO. Move REG_ATTRS index down.
1588 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1589 code that prints the REGNO.
1591 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1593 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1596 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1598 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1600 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1602 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1603 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1605 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1607 * config/aarch64/aarch64-protos.h
1608 (aarch64_hard_regno_caller_save_mode): New prototype.
1609 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1611 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1613 2014-05-13 Christian Bruel <christian.bruel@st.com>
1615 * target.def (mode_switching): New hook vector.
1616 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1617 (mode_exit, modepriority_to_mode): Likewise.
1618 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1619 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1620 * target.h: Include tm.h and hard-reg-set.h.
1621 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1622 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1623 * doc/tm.texi Regenerate.
1624 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1625 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1626 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1627 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1628 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1629 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1630 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1631 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1632 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1633 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1634 (ix86_emit_mode_set): Hookify.
1635 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1637 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1638 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1639 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1640 (epiphany_mode_priority_to_mode): Remove declaration.
1641 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1642 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1643 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1645 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1646 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1647 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1649 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1652 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1653 is const0_rtx, return immediately. Don't test count == 0 when
1656 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1658 * Makefile.in: add shrink-wrap.o.
1659 * config/i386/i386.c: include "shrink-wrap.h"
1660 * function.c: Likewise.
1661 (requires_stack_frame_p, next_block_for_reg,
1662 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1663 dup_block_and_redirect): Move to shrink-wrap.c
1664 (thread_prologue_and_epilogue_insns): Extract three code segments
1665 as functions in shrink-wrap.c
1666 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1668 * shrink-wrap.c: New file.
1669 * shrink-wrap.h: New file.
1671 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1673 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1674 reference to Solaris.
1676 2014-05-12 Mike Stump <mikestump@comcast.net>
1679 * genattrtab.c (filename): Add.
1680 (convert_set_attr_alternative): Improve error message.
1681 (check_defs): Restore read_md_filename for error messages.
1682 (gen_insn): Save filename.
1684 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1686 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1687 -fno-local-ivars and -fivar-visibility.
1688 * c-family/c.opt: Make -Wshadow also implicitly enable
1691 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1693 * doc/tm.texi: Remove reference to deleted macro.
1694 * doc/tm.texi.in: Likewise.
1696 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1699 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1702 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1704 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1705 * config/arm/aout.h (License): Same.
1706 * config/arm/bpabi.h (License): Same.
1707 * config/arm/elf.h (License): Same.
1708 * config/arm/linux-elf.h (License): Same.
1709 * config/arm/linux-gas.h (License): Same.
1710 * config/arm/netbsd-elf.h (License): Same.
1711 * config/arm/uclinux-eabi.h (License): Same.
1712 * config/arm/uclinux-elf.h (License): Same.
1713 * config/arm/vxworks.h (License): Same.
1715 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1717 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1718 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1719 number of operands to 3.
1720 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1721 * tree-nested.c (convert_nonlocal_omp_clauses,
1722 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1723 * gimplify.c (gimplify_scan_omp_clauses): Handle
1724 OMP_CLAUSE_LINEAR_STMT.
1725 * omp-low.c (lower_rec_input_clauses): Fix typo.
1726 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1727 cast between Fortran boolean_type_node and C _Bool if
1730 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1732 PR tree-optimization/61136
1733 * wide-int.h (multiple_of_p): Define a version that doesn't return
1735 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1736 integer_zerop/const_binop pair.
1737 (multiple_of_p): Likewise, converting both operands to widest_int
1740 2014-05-09 Teresa Johnson <tejohnson@google.com>
1742 * cgraphunit.c (analyze_functions): Use correct dump file.
1744 2014-05-09 Florian Weimer <fweimer@redhat.com>
1746 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1748 (stack_protect_return_slot_p): New function.
1749 (expand_used_vars): Call stack_protect_decl_p and
1750 stack_protect_return_slot_p for -fstack-protector-strong.
1752 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1755 * fold-const.c (fold_binary_loc): Changed width of mask.
1757 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1759 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1760 unsigned int initializers for regno_in, regno_out.
1762 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1765 * config/avr/avr.md (cc): Add new attribute set_vzn.
1766 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1767 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1768 with INC, DEC or NEG.
1769 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1770 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1771 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1773 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1776 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1778 * wide-int.cc (UTItype): Define.
1779 (UDWtype): Define for appropriate W_TYPE_SIZE.
1781 2014-05-09 Richard Biener <rguenther@suse.de>
1783 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1784 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1785 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1786 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1787 ssa_propagate): Adjust.
1789 2014-05-08 Jeff Law <law@redhat.com>
1791 PR tree-optimization/61009
1792 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1793 tri-state rather than a boolean. When a block is too big to
1794 thread through, inform caller via negative return value.
1795 (thread_across_edge): If a block was too big for normal threading,
1796 then it's too big for a joiner too, so remove temporary equivalences
1797 and return immediately.
1799 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1800 Matthias Klose <doko@ubuntu.com>
1803 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1805 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1808 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1810 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1813 * config/alpha/alpha.c: Include gimple-iterator.h.
1814 (alpha_gimple_fold_builtin): New function. Move
1815 ALPHA_BUILTIN_UMULH folding from ...
1816 (alpha_fold_builtin): ... here.
1817 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1819 2014-05-08 Wei Mi <wmi@google.com>
1822 * config/i386/i386.c (ix86_compute_frame_layout): Update
1823 preferred_stack_boundary for call, expanded from tls descriptor.
1824 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1825 to depend on SP register.
1826 (*tls_local_dynamic_base_32_gnu): Ditto.
1827 (*tls_local_dynamic_32_once): Ditto.
1828 (tls_global_dynamic_64_<mode>): Set
1829 ix86_tls_descriptor_calls_expanded_in_cfun.
1830 (tls_local_dynamic_base_64_<mode>): Ditto.
1831 (tls_global_dynamic_32): Set
1832 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1833 to depend on SP register.
1834 (tls_local_dynamic_base_32): Ditto.
1836 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1838 * config/arm/arm_neon.h: Update comment.
1839 * config/arm/neon-docgen.ml: Delete.
1840 * config/arm/neon-gen.ml: Delete.
1841 * doc/arm-neon-intrinsics.texi: Update comment.
1843 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1845 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1847 (vand, vorr, veor, vorn, vbic): Remove.
1848 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1850 (neon_vsub_unspec): Likewise.
1851 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1853 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1855 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1856 (vadd_s16): Likewise.
1857 (vadd_s32): Likewise.
1858 (vadd_f32): Likewise.
1859 (vadd_u8): Likewise.
1860 (vadd_u16): Likewise.
1861 (vadd_u32): Likewise.
1862 (vadd_s64): Likewise.
1863 (vadd_u64): Likewise.
1864 (vaddq_s8): Likewise.
1865 (vaddq_s16): Likewise.
1866 (vaddq_s32): Likewise.
1867 (vaddq_s64): Likewise.
1868 (vaddq_f32): Likewise.
1869 (vaddq_u8): Likewise.
1870 (vaddq_u16): Likewise.
1871 (vaddq_u32): Likewise.
1872 (vaddq_u64): Likewise.
1873 (vmul_s8): Likewise.
1874 (vmul_s16): Likewise.
1875 (vmul_s32): Likewise.
1876 (vmul_f32): Likewise.
1877 (vmul_u8): Likewise.
1878 (vmul_u16): Likewise.
1879 (vmul_u32): Likewise.
1880 (vmul_p8): Likewise.
1881 (vmulq_s8): Likewise.
1882 (vmulq_s16): Likewise.
1883 (vmulq_s32): Likewise.
1884 (vmulq_f32): Likewise.
1885 (vmulq_u8): Likewise.
1886 (vmulq_u16): Likewise.
1887 (vmulq_u32): Likewise.
1888 (vsub_s8): Likewise.
1889 (vsub_s16): Likewise.
1890 (vsub_s32): Likewise.
1891 (vsub_f32): Likewise.
1892 (vsub_u8): Likewise.
1893 (vsub_u16): Likewise.
1894 (vsub_u32): Likewise.
1895 (vsub_s64): Likewise.
1896 (vsub_u64): Likewise.
1897 (vsubq_s8): Likewise.
1898 (vsubq_s16): Likewise.
1899 (vsubq_s32): Likewise.
1900 (vsubq_s64): Likewise.
1901 (vsubq_f32): Likewise.
1902 (vsubq_u8): Likewise.
1903 (vsubq_u16): Likewise.
1904 (vsubq_u32): Likewise.
1905 (vsubq_u64): Likewise.
1906 (vand_s8): Likewise.
1907 (vand_s16): Likewise.
1908 (vand_s32): Likewise.
1909 (vand_u8): Likewise.
1910 (vand_u16): Likewise.
1911 (vand_u32): Likewise.
1912 (vand_s64): Likewise.
1913 (vand_u64): Likewise.
1914 (vandq_s8): Likewise.
1915 (vandq_s16): Likewise.
1916 (vandq_s32): Likewise.
1917 (vandq_s64): Likewise.
1918 (vandq_u8): Likewise.
1919 (vandq_u16): Likewise.
1920 (vandq_u32): Likewise.
1921 (vandq_u64): Likewise.
1922 (vorr_s8): Likewise.
1923 (vorr_s16): Likewise.
1924 (vorr_s32): Likewise.
1925 (vorr_u8): Likewise.
1926 (vorr_u16): Likewise.
1927 (vorr_u32): Likewise.
1928 (vorr_s64): Likewise.
1929 (vorr_u64): Likewise.
1930 (vorrq_s8): Likewise.
1931 (vorrq_s16): Likewise.
1932 (vorrq_s32): Likewise.
1933 (vorrq_s64): Likewise.
1934 (vorrq_u8): Likewise.
1935 (vorrq_u16): Likewise.
1936 (vorrq_u32): Likewise.
1937 (vorrq_u64): Likewise.
1938 (veor_s8): Likewise.
1939 (veor_s16): Likewise.
1940 (veor_s32): Likewise.
1941 (veor_u8): Likewise.
1942 (veor_u16): Likewise.
1943 (veor_u32): Likewise.
1944 (veor_s64): Likewise.
1945 (veor_u64): Likewise.
1946 (veorq_s8): Likewise.
1947 (veorq_s16): Likewise.
1948 (veorq_s32): Likewise.
1949 (veorq_s64): Likewise.
1950 (veorq_u8): Likewise.
1951 (veorq_u16): Likewise.
1952 (veorq_u32): Likewise.
1953 (veorq_u64): Likewise.
1954 (vbic_s8): Likewise.
1955 (vbic_s16): Likewise.
1956 (vbic_s32): Likewise.
1957 (vbic_u8): Likewise.
1958 (vbic_u16): Likewise.
1959 (vbic_u32): Likewise.
1960 (vbic_s64): Likewise.
1961 (vbic_u64): Likewise.
1962 (vbicq_s8): Likewise.
1963 (vbicq_s16): Likewise.
1964 (vbicq_s32): Likewise.
1965 (vbicq_s64): Likewise.
1966 (vbicq_u8): Likewise.
1967 (vbicq_u16): Likewise.
1968 (vbicq_u32): Likewise.
1969 (vbicq_u64): Likewise.
1970 (vorn_s8): Likewise.
1971 (vorn_s16): Likewise.
1972 (vorn_s32): Likewise.
1973 (vorn_u8): Likewise.
1974 (vorn_u16): Likewise.
1975 (vorn_u32): Likewise.
1976 (vorn_s64): Likewise.
1977 (vorn_u64): Likewise.
1978 (vornq_s8): Likewise.
1979 (vornq_s16): Likewise.
1980 (vornq_s32): Likewise.
1981 (vornq_s64): Likewise.
1982 (vornq_u8): Likewise.
1983 (vornq_u16): Likewise.
1984 (vornq_u32): Likewise.
1985 (vornq_u64): Likewise.
1987 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1989 * wide-int.cc (UTItype): Define.
1990 (UDWtype): Define for appropriate W_TYPE_SIZE.
1992 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1994 PR tree-optimization/59100
1995 * tree-ssa-phiopt.c: Include tree-inline.h.
1996 (neutral_element_p, absorbing_element_p): New functions.
1997 (value_replacement): Handle conditional binary operations with a
1998 neutral or absorbing element.
2000 2014-05-08 Richard Biener <rguenther@suse.de>
2002 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2003 folding the expression.
2004 (valueize_expr): Remove.
2005 (visit_reference_op_load): Do not valueize the result of
2007 (simplify_binary_expression): Likewise.
2008 (simplify_unary_expression): Likewise.
2010 2014-05-08 Richard Biener <rguenther@suse.de>
2012 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2013 looking at TYPE_ARG_TYPES.
2015 2014-05-08 Richard Biener <rguenther@suse.de>
2017 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2018 pointer propagation special-case.
2020 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2022 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2023 core part of address expressions.
2025 2014-05-08 Alan Modra <amodra@gmail.com>
2028 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2029 loads and stores when -mno-strict-align at any alignment.
2030 (expand_block_clear): Similarly. Also correct calculation of
2033 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2036 * tree-complex.c (expand_complex_move): Keep line info when expanding
2038 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2039 of complex expression. Use new argument to display correct location
2040 for values coming from phi statement.
2041 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2042 (warn_uninitialized_phi): Pass location of phi argument to
2044 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2045 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2047 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2049 * config/rs6000/predicates.md (indexed_address_mem): New.
2050 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2051 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2052 fpstore_ux, fpstore_u.
2053 (sign_extend, indexed, update): New.
2054 (cell_micro): Adjust.
2055 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2056 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2057 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2058 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2059 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2060 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2061 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2062 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2063 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2064 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2065 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2066 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2067 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2068 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2069 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2071 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2072 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2073 *vsx_extract_<mode>_store): Adjust.
2074 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2075 is_cracked_insn, insn_must_be_first_in_group,
2076 insn_must_be_last_in_group): Adjust.
2078 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2080 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2081 ppc440-fpstore): Adjust.
2082 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2083 ppc476-fpstore): Adjust.
2084 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2085 ppc601-fpstore): Adjust.
2086 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2088 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2090 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2091 ppc7450-fpstore): Adjust.
2092 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2093 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2094 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2096 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2097 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2098 cell-fpstore, cell-fpstore-update): Adjust.
2099 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2100 ppce300c3_store, ppce300c3_fpstore): Adjust.
2101 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2102 e500mc_fpstore): Adjust.
2103 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2104 e500mc64_store, e500mc64_fpstore): Adjust.
2105 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2106 e5500_fpstore): Adjust.
2107 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2108 e6500_fpstore): Adjust.
2109 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2111 * config/rs6000/power4.md (power4-load, power4-load-ext,
2112 power4-load-ext-update, power4-load-ext-update-indexed,
2113 power4-load-update-indexed, power4-load-update, power4-fpload,
2114 power4-fpload-update, power4-store, power4-store-update,
2115 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2117 * config/rs6000/power5.md (power5-load, power5-load-ext,
2118 power5-load-ext-update, power5-load-ext-update-indexed,
2119 power5-load-update-indexed, power5-load-update, power5-fpload,
2120 power5-fpload-update, power5-store, power5-store-update,
2121 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2123 * config/rs6000/power6.md (power6-load, power6-load-ext,
2124 power6-load-update, power6-load-update-indexed,
2125 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2126 power6-fpload-update, power6-store, power6-store-update,
2127 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2129 * config/rs6000/power7.md (power7-load, power7-load-ext,
2130 power7-load-update, power7-load-update-indexed,
2131 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2132 power7-fpload-update, power7-store, power7-store-update,
2133 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2135 * config/rs6000/power8.md (power8-load, power8-load-update,
2136 power8-load-ext, power8-load-ext-update, power8-fpload,
2137 power8-fpload-update, power8-store, power8-store-update-indexed,
2138 power8-fpstore, power8-fpstore-update): Adjust.
2139 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2141 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2142 titan_lsu_store, titan_lsu_fpstore): Adjust.
2143 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2145 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2148 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2149 unrolled byte insns. Emit address increments after move insns.
2151 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2153 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2154 const_gimple, rather than a gimple.
2155 (gimple_call_builtin_p): Likewise, for the three variants.
2157 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2158 (gimple_call_builtin_p): Likewise, for the three variants.
2160 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2162 PR tree-optimization/61095
2163 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2165 2014-05-07 Richard Biener <rguenther@suse.de>
2167 PR tree-optimization/61034
2168 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2169 (maybe_skip_until): Use translate to take into account
2170 lattices when trying to do disambiguations.
2171 (get_continuation_for_phi_1): Likewise.
2172 (get_continuation_for_phi): Adjust for added translate arguments.
2173 (walk_non_aliased_vuses): Likewise.
2174 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2175 (walk_non_aliased_vuses): Likewise.
2176 (call_may_clobber_ref_p_1): Declare.
2177 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2178 calls. Stop early if we are only supposed to disambiguate.
2179 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2181 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2183 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2184 Emit an error when the function has arguments.
2186 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2188 * cfgloop.h (unswitch_loops): Remove.
2189 * doc/passes.texi: Remove references to loop-unswitch.c
2190 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2192 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2194 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2195 check for loads group of length 3.
2196 (vect_permute_load_chain): New permutations for loads group of
2198 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2199 of vec_perm_shuffle for the new permutations.
2201 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2203 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2204 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2205 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2206 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2207 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2208 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2209 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2210 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2212 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2214 * loop-unswitch.c: Delete.
2216 2014-05-07 Richard Biener <rguenther@suse.de>
2218 * config.gcc: Always set need_64bit_hwint to yes.
2220 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2222 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2223 of using optimize_size.
2225 2014-05-06 Mike Stump <mikestump@comcast.net>
2227 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2229 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2231 * config/i386/sse.md (*mov<mode>_internal)
2232 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2233 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2234 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2235 (*<code><mode>3, *andnot<mode>3<mask_name>)
2236 (<mask_codefor><code><mode>3<mask_name>): Only consider
2237 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2239 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2242 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2244 * lra-constraints.c (valid_address_p): Move earlier in file.
2245 Add a constraint argument to the address_info version.
2246 (satisfies_memory_constraint_p): New function.
2247 (satisfies_address_constraint_p): Likewise.
2248 (process_alt_operands, curr_insn_transform): Use them.
2249 (process_address): Pass the constraint to valid_address_p when
2250 checking address operands.
2252 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2254 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2255 to their respective blocks. Fix inadvertent use of "node".
2257 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2259 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2261 (init_emit_once): ...here.
2262 * rtl.h (init_derived_machine_modes): Declare.
2263 * toplev.c (do_compile): Call it even if no_backend.
2265 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2266 Mike Stump <mikestump@comcast.net>
2267 Richard Sandiford <rdsandiford@googlemail.com>
2268 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2270 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2271 (rtx_equal_for_memref_p): Update comment.
2272 (adjust_offset_for_component_ref): Use wide-int interfaces.
2273 * builtins.c (get_object_alignment_2): Likewise.
2274 (c_readstr): Likewise.
2275 (target_char_cast): Add comment.
2276 (determine_block_size): Use wide-int interfaces.
2277 (expand_builtin_signbit): Likewise.
2278 (fold_builtin_int_roundingfn): Likewise.
2279 (fold_builtin_bitop): Likewise.
2280 (fold_builtin_bswap): Likewise.
2281 (fold_builtin_logarithm): Use signop.
2282 (fold_builtin_pow): Likewise.
2283 (fold_builtin_memory_op): Use wide-int interfaces.
2284 (fold_builtin_object_size): Likewise.
2285 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2286 nb_iterations_estimate.
2287 (record_niter_bound): Use wide-int interfaces.
2288 (get_estimated_loop_iterations_int): Likewise.
2289 (get_estimated_loop_iterations): Likewise.
2290 (get_max_loop_iterations): Likewise.
2291 * cfgloop.h: Include wide-int.h.
2292 (struct nb_iter_bound): Change bound to widest_int.
2293 (struct loop): Change nb_iterations_upper_bound and
2294 nb_iterations_estimate to widest_int.
2295 (record_niter_bound): Switch to use widest_int.
2296 (get_estimated_loop_iterations): Likewise.
2297 (get_max_loop_iterations): Likewise.
2298 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2299 update for wide-int.
2300 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2301 * combine.c (try_combine): Likewise.
2302 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2303 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2305 (aarch64_float_const_representable_p): Likewise.
2306 * config/arc/arc.c: Include wide-int.h.
2307 (arc_can_use_doloop_p): Use wide-int interfaces.
2308 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2309 (vfp3_const_double_index): Likewise.
2310 * config/avr/avr.c (avr_out_round): Likewise.
2311 (avr_fold_builtin): Likewise.
2312 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2313 (bfin_can_use_doloop_p): Likewise.
2314 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2315 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2316 * config/i386/i386.c: Include wide-int.h.
2317 (ix86_data_alignment): Use wide-int interfaces.
2318 (ix86_local_alignment): Likewise.
2319 (ix86_emit_swsqrtsf): Update real_from_integer.
2320 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2321 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2322 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2323 (zero_constant): Likewise.
2324 (input_operand): Likewise.
2325 (splat_input_operand): Likewise.
2326 (non_logical_cint_operand): Change const_double to const_wide_int.
2327 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2328 (easy_altivec_constant): Remove comment.
2329 (paired_expand_vector_init): Use CONSTANT_P.
2330 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2331 (rs6000_emit_move): Update checks.
2332 (rs6000_aggregate_candidate): Use wide-int interfaces.
2333 (rs6000_expand_ternop_builtin): Likewise.
2334 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2335 (rs6000_assemble_integer): Likewise.
2336 (rs6000_hash_constant): Likewise.
2337 (output_toc): Likewise.
2338 (rs6000_rtx_costs): Likewise.
2339 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2340 * config/rs6000/rs6000-c.c: Include wide-int.h.
2341 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2342 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2343 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2344 Handle CONST_WIDE_INT.
2345 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2346 Use tree_fits_uhwi_p.
2347 * config/sparc/sparc.c: Include wide-int.h.
2348 (sparc_fold_builtin): Use wide-int interfaces.
2349 * config/vax/vax.c: Include wide-int.h.
2350 (vax_float_literal): Use real_from_integer.
2351 * coretypes.h (struct hwivec_def): New.
2353 (const_hwivec): New.
2354 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2355 (equiv_constant): Handle CONST_WIDE_INT.
2356 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2357 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2358 * dbxout.c (stabstr_U): Use wide-int interfaces.
2359 (dbxout_type): Update to use cst_fits_shwi_p.
2360 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2361 (TARGET_SUPPORTS_WIDE_INT): Add default.
2362 * dfp.c: Include wide-int.h.
2363 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2364 decimal_real_to_integer.
2365 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2366 decimal_real_to_integer.
2367 * doc/generic.texi (Constant expressions): Update for wide_int.
2368 * doc/rtl.texi (const_double): Likewise.
2369 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2370 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2371 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2372 (REAL_VALUE_FROM_INT): Remove.
2373 (TARGET_SUPPORTS_WIDE_INT): New.
2374 * doc/tm.texi: Regenerate.
2375 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2376 * double-int.h: Include wide-int.h.
2377 (struct wi::int_traits): New.
2378 * dwarf2out.c (get_full_len): New.
2379 (dw_val_equal_p): Add case dw_val_class_wide_int.
2380 (size_of_loc_descr): Likewise.
2381 (output_loc_operands): Likewise.
2382 (insert_double): Remove.
2383 (insert_wide_int): New.
2385 (print_die): Add case dw_val_class_wide_int.
2386 (attr_checksum): Likewise.
2387 (attr_checksum_ordered): Likewise.
2388 (same_dw_val_p): Likewise.
2389 (size_of_die): Likewise.
2390 (value_format): Likewise.
2391 (output_die): Likewise.
2392 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2394 (clz_loc_descriptor): Use wide-int interfaces.
2395 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2396 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2397 (round_up_to_align): Use wide-int interfaces.
2398 (field_byte_offset): Likewise.
2399 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2400 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2401 CONST_DOUBLE handling. Use wide-int interfaces.
2402 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2403 (gen_enumeration_type_die): Use add_AT_wide.
2404 (hash_loc_operands): Add case dw_val_class_wide_int.
2405 (compare_loc_operands): Likewise.
2406 * dwarf2out.h: Include wide-int.h.
2407 (wide_int_ptr): New.
2408 (enum dw_val_class): Add dw_val_class_wide_int.
2409 (struct dw_val_struct): Add val_wide.
2410 * emit-rtl.c (const_wide_int_htab): New.
2411 (const_wide_int_htab_hash): New.
2412 (const_wide_int_htab_eq): New.
2413 (lookup_const_wide_int): New.
2414 (const_double_htab_hash): Use wide-int interfaces.
2415 (const_double_htab_eq): Likewise.
2416 (rtx_to_double_int): Conditionally compile for wide-int.
2417 (immed_double_int_const): Rename to immed_wide_int_const and
2418 update for wide-int.
2419 (immed_double_const): Conditionally compile for wide-int.
2420 (init_emit_once): Use wide-int interfaces.
2421 * explow.c (plus_constant): Likewise.
2422 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2423 (lshift_value): Use wide-int interfaces.
2424 (expand_mult): Likewise.
2425 (choose_multiplier): Likewise.
2426 (expand_smod_pow2): Likewise.
2427 (make_tree): Likewise.
2428 * expr.c (convert_modes): Consolidate handling of constants.
2429 Use wide-int interfaces.
2430 (emit_group_load_1): Add note.
2431 (store_expr): Update comment.
2432 (get_inner_reference): Use wide-int interfaces.
2433 (expand_constructor): Update comment.
2434 (expand_expr_real_2): Use wide-int interfaces.
2435 (expand_expr_real_1): Likewise.
2436 (reduce_to_bit_field_precision): Likewise.
2437 (const_vector_from_tree): Likewise.
2438 * final.c: Include wide-int-print.h.
2439 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2440 * fixed-value.c: Include wide-int.h.
2441 (fixed_from_string): Use wide-int interfaces.
2442 (fixed_to_decimal): Likewise.
2443 (fixed_convert_from_real): Likewise.
2444 (real_convert_from_fixed): Likewise.
2445 * fold-const.h (mem_ref_offset): Return an offset_int.
2446 (div_if_zero_remainder): Remove code parameter.
2447 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2448 Use wide-int interfaces.
2449 (may_negate_without_overflow_p): Use wide-int interfaces.
2450 (negate_expr_p): Likewise.
2451 (fold_negate_expr): Likewise.
2452 (int_const_binop_1): Likewise.
2453 (const_binop): Likewise.
2454 (fold_convert_const_int_from_int): Likewise.
2455 (fold_convert_const_int_from_real): Likewise.
2456 (fold_convert_const_int_from_fixed): Likewise.
2457 (fold_convert_const_fixed_from_int): Likewise.
2458 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2459 (sign_bit_p): Use wide-int interfaces.
2460 (make_range_step): Likewise.
2461 (build_range_check): Likewise. Pass an integer of the correct type
2462 instead of using integer_one_node.
2463 (range_predecessor): Pass an integer of the correct type instead
2464 of using integer_one_node.
2465 (range_successor): Likewise.
2466 (merge_ranges): Likewise.
2467 (unextend): Use wide-int interfaces.
2468 (extract_muldiv_1): Likewise.
2469 (fold_div_compare): Likewise.
2470 (fold_single_bit_test): Likewise.
2471 (fold_sign_changed_comparison): Likewise.
2472 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2473 (fold_plusminus_mult_expr): Use wide-int interfaces.
2474 (native_encode_int): Likewise.
2475 (native_interpret_int): Likewise.
2476 (fold_unary_loc): Likewise.
2477 (pointer_may_wrap_p): Likewise.
2478 (size_low_cst): Likewise.
2479 (mask_with_tz): Likewise.
2480 (fold_binary_loc): Likewise.
2481 (fold_ternary_loc): Likewise.
2482 (multiple_of_p): Likewise.
2483 (tree_call_nonnegative_warnv_p): Update calls to
2484 tree_int_cst_min_precision and real_from_integer.
2485 (fold_negate_const): Use wide-int interfaces.
2486 (fold_abs_const): Likewise.
2487 (fold_relational_const): Use tree_int_cst_lt.
2488 (round_up_loc): Use wide-int interfaces.
2489 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2490 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2491 * gengtype.c: Remove include of double-int.h.
2492 (do_typedef): Use wide-int interfaces.
2493 (open_base_files): Add wide-int.h.
2494 (main): Add offset_int and widest_int typedefs.
2495 * gengtype-lex.l: Handle "^".
2496 (CXX_KEYWORD): Add "static".
2497 * gengtype-parse.c (require3): New.
2498 (require_template_declaration): Handle constant template arguments
2499 and nested templates.
2500 * gengtype-state.c: Don't include "double-int.h".
2501 * genpreds.c (write_one_predicate_function): Update comment.
2502 (write_tm_constrs_h): Add check for hval and lval use in
2504 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2505 (add_to_sequence): Likewise.
2506 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2507 and const_double_operand.
2508 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2510 * gimple-fold.c (get_base_constructor): Likewise.
2511 (fold_array_ctor_reference): Likewise.
2512 (fold_nonarray_ctor_reference): Likewise.
2513 (fold_const_aggregate_ref_1): Likewise.
2514 (gimple_val_nonnegative_real_p): Likewise.
2515 (gimple_fold_indirect_ref): Likewise.
2516 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2517 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2518 (struct slsr_cand_d): Change index to be widest_int.
2519 (struct incr_info_d): Change incr to be widest_int.
2520 (alloc_cand_and_find_basis): Use wide-int interfaces.
2521 (slsr_process_phi): Likewise.
2522 (backtrace_base_for_ref): Likewise. Return a widest_int.
2523 (restructure_reference): Take a widest_int instead of a double_int.
2524 (slsr_process_ref): Use wide-int interfaces.
2525 (create_mul_ssa_cand): Likewise.
2526 (create_mul_imm_cand): Likewise.
2527 (create_add_ssa_cand): Likewise.
2528 (create_add_imm_cand): Take a widest_int instead of a double_int.
2529 (slsr_process_add): Use wide-int interfaces.
2530 (slsr_process_cast): Likewise.
2531 (slsr_process_copy): Likewise.
2532 (dump_candidate): Likewise.
2533 (dump_incr_vec): Likewise.
2534 (replace_ref): Likewise.
2535 (cand_increment): Likewise. Return a widest_int.
2536 (cand_abs_increment): Likewise.
2537 (replace_mult_candidate): Take a widest_int instead of a double_int.
2538 (replace_unconditional_candidate): Use wide-int interfaces.
2539 (incr_vec_index): Take a widest_int instead of a double_int.
2540 (create_add_on_incoming_edge): Likewise.
2541 (create_phi_basis): Use wide-int interfaces.
2542 (replace_conditional_candidate): Likewise.
2543 (record_increment): Take a widest_int instead of a double_int.
2544 (record_phi_increments): Use wide-int interfaces.
2545 (phi_incr_cost): Take a widest_int instead of a double_int.
2546 (lowest_cost_path): Likewise.
2547 (total_savings): Likewise.
2548 (analyze_increments): Use wide-int interfaces.
2549 (ncd_with_phi): Take a widest_int instead of a double_int.
2550 (ncd_of_cand_and_phis): Likewise.
2551 (nearest_common_dominator_for_cands): Likewise.
2552 (insert_initializers): Use wide-int interfaces.
2553 (all_phi_incrs_profitable): Likewise.
2554 (replace_one_candidate): Likewise.
2555 (replace_profitable_candidates): Likewise.
2556 * godump.c: Include wide-int-print.h.
2557 (go_output_typedef): Use wide-int interfaces.
2558 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2559 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2560 (build_loop_iteration_domains): Likewise.
2561 * hooks.h: Include wide-int.h rather than double-int.h.
2562 (hook_bool_dint_dint_uint_bool_true): Delete.
2563 (hook_bool_wint_wint_uint_bool_true): Declare.
2564 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2565 (hook_bool_wint_wint_uint_bool_true): New.
2566 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2568 (ubsan_expand_si_overflow_mul_check): Likewise.
2569 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2570 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2571 (get_ancestor_addr_info): Likewise.
2572 (ipa_modify_call_arguments): Likewise.
2573 * loop-doloop.c (doloop_modify): Likewise.
2574 (doloop_optimize): Likewise.
2575 * loop-iv.c (iv_number_of_iterations): Likewise.
2576 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2577 (unroll_loop_constant_iterations): Likewise.
2578 (decide_unroll_runtime_iterations): Likewise.
2579 (unroll_loop_runtime_iterations): Likewise.
2580 (decide_peel_simple): Likewise.
2581 (decide_unroll_stupid): Likewise.
2582 * lto-streamer-in.c (streamer_read_wi): Add.
2583 (input_cfg): Use wide-int interfaces.
2584 (lto_input_tree_1): Likewise.
2585 * lto-streamer-out.c (streamer_write_wi): Add.
2586 (hash_tree): Use wide-int interfaces.
2587 (output_cfg): Likewise.
2588 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2589 (GTFILES): Add wide-int.h and signop.h.
2590 (TAGS): Look for .cc files too.
2591 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2592 * optabs.c (expand_subword_shift): Likewise.
2593 (expand_doubleword_shift): Likewise.
2594 (expand_absneg_bit): Likewise.
2595 (expand_copysign_absneg): Likewise.
2596 (expand_copysign_bit): Likewise.
2597 * postreload.c (reload_cse_simplify_set): Likewise.
2598 * predict.c (predict_iv_comparison): Likewise.
2599 * pretty-print.h: Include wide-int-print.h.
2601 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2602 * print-tree.c: Include wide-int-print.h.
2603 (print_node_brief): Use wide-int interfaces.
2604 (print_node): Likewise.
2605 * read-rtl.c (validate_const_wide_int): New.
2606 (read_rtx_code): Add CONST_WIDE_INT case.
2607 * real.c: Include wide-int.h.
2608 (real_to_integer2): Delete.
2609 (real_to_integer): New function, returning a wide_int.
2610 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2611 (ten_to_ptwo): Update call to real_from_integer.
2612 (real_digit): Likewise.
2613 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2614 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2615 (REAL_VALUE_TO_INT): Delete.
2616 (real_to_integer): Declare a wide-int form.
2617 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2618 * recog.c (const_int_operand): Improve comment.
2619 (const_scalar_int_operand): New.
2620 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2621 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2622 (split_double): Likewise.
2623 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2624 (rtx_size): Likewise.
2625 (rtx_alloc_stat_v): New.
2626 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2627 (cwi_output_hex): New.
2628 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2629 (cwi_check_failed_bounds): New.
2630 * rtl.def (CONST_WIDE_INT): New.
2631 * rtl.h: Include <utility> and wide-int.h.
2632 (struct hwivec_def): New.
2633 (CWI_GET_NUM_ELEM): New.
2634 (CWI_PUT_NUM_ELEM): New.
2635 (struct rtx_def): Add num_elem and hwiv.
2636 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2637 (CASE_CONST_UNIQUE): Likewise.
2638 (CASE_CONST_ANY): Likewise.
2639 (CONST_SCALAR_INT_P): Likewise.
2640 (CONST_WIDE_INT_P): New.
2642 (HWIVEC_CHECK): New.
2643 (cwi_check_failed_bounds): New.
2645 (HWIVEC_CHECK): New.
2646 (CONST_WIDE_INT_VEC) New.
2647 (CONST_WIDE_INT_NUNITS) New.
2648 (CONST_WIDE_INT_ELT) New.
2649 (rtx_mode_t): New type.
2650 (wi::int_traits <rtx_mode_t>): New.
2652 (wi::min_value): New.
2653 (wi::max_value): New.
2655 (const_wide_int_alloc): New.
2656 (immed_wide_int_const): New.
2657 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2658 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2659 * signop.h: New file.
2660 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2661 (simplify_const_unary_operation): Use wide-int interfaces.
2662 (simplify_binary_operation_1): Likewise.
2663 (simplify_const_binary_operation): Likewise.
2664 (simplify_const_relational_operation): Likewise.
2665 (simplify_immed_subreg): Likewise.
2666 * stmt.c (expand_case): Likewise.
2667 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2668 signop rather than a bool.
2669 * stor-layout.c (layout_type): Use wide-int interfaces.
2670 (initialize_sizetypes): Update calls to
2671 set_min_and_max_values_for_integral_type.
2672 (set_min_and_max_values_for_integral_type): Take a signop rather
2673 than a bool. Use wide-int interfaces.
2674 (fixup_signed_type): Update accordingly. Remove
2675 HOST_BITS_PER_DOUBLE_INT limit.
2676 (fixup_unsigned_type): Likewise.
2677 * system.h (STATIC_CONSTANT_P): New.
2678 (STATIC_ASSERT): New.
2679 * target.def (can_use_doloop_p): Take widest_ints rather than
2681 * target.h: Include wide-int.h rather than double-int.h.
2682 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2684 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2685 rather than INT_CST_LT_UNSIGNED.
2686 (can_use_doloop_if_innermost): Take widest_ints rather than
2688 * tree-affine.c: Include wide-int-print.h.
2689 (double_int_ext_for_comb): Delete.
2690 (wide_int_ext_for_comb): New.
2691 (aff_combination_zero): Use wide-int interfaces.
2692 (aff_combination_const): Take a widest_int instead of a double_int.
2693 (aff_combination_elt): Use wide-int interfaces.
2694 (aff_combination_scale): Take a widest_int instead of a double_int.
2695 (aff_combination_add_elt): Likewise.
2696 (aff_combination_add_cst): Likewise.
2697 (aff_combination_add): Use wide-int interfaces.
2698 (aff_combination_convert): Likewise.
2699 (tree_to_aff_combination): Likewise.
2700 (add_elt_to_tree): Take a widest_int instead of a double_int.
2701 (aff_combination_to_tree): Use wide-int interfaces.
2702 (aff_combination_remove_elt): Likewise.
2703 (aff_combination_add_product): Take a widest_int instead of
2705 (aff_combination_mult): Use wide-int interfaces.
2706 (aff_combination_expand): Likewise.
2707 (double_int_constant_multiple_p): Delete.
2708 (wide_int_constant_multiple_p): New.
2709 (aff_combination_constant_multiple_p): Take a widest_int pointer
2710 instead of a double_int pointer.
2711 (print_aff): Use wide-int interfaces.
2712 (get_inner_reference_aff): Take a widest_int pointer
2713 instead of a double_int pointer.
2714 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2715 * tree-affine.h: Include wide-int.h.
2716 (struct aff_comb_elt): Change type of coef to widest_int.
2717 (struct affine_tree_combination): Change type of offset to widest_int.
2718 (double_int_ext_for_comb): Delete.
2719 (wide_int_ext_for_comb): New.
2720 (aff_combination_const): Use widest_int instead of double_int.
2721 (aff_combination_scale): Likewise.
2722 (aff_combination_add_elt): Likewise.
2723 (aff_combination_constant_multiple_p): Likewise.
2724 (get_inner_reference_aff): Likewise.
2725 (aff_comb_cannot_overlap_p): Likewise.
2726 (aff_combination_zero_p): Use wide-int interfaces.
2727 * tree.c: Include tree.h.
2728 (init_ttree): Use make_int_cst.
2729 (tree_code_size): Removed code for INTEGER_CST case.
2730 (tree_size): Add INTEGER_CST case.
2731 (make_node_stat): Update comment.
2732 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2733 (build_int_cst_type): Use wide-int interfaces.
2734 (double_int_to_tree): Likewise.
2735 (double_int_fits_to_tree_p): Delete.
2736 (force_fit_type_double): Delete.
2737 (force_fit_type): New.
2738 (int_cst_hash_hash): Use wide-int interfaces.
2739 (int_cst_hash_eq): Likewise.
2740 (build_int_cst_wide): Delete.
2741 (wide_int_to_tree): New.
2742 (cache_integer_cst): Use wide-int interfaces.
2743 (build_low_bits_mask): Likewise.
2744 (cst_and_fits_in_hwi): Likewise.
2745 (real_value_from_int_cst): Likewise.
2746 (make_int_cst_stat): New.
2747 (integer_zerop): Use wide_int interfaces.
2748 (integer_onep): Likewise.
2749 (integer_all_onesp): Likewise.
2750 (integer_pow2p): Likewise.
2751 (integer_nonzerop): Likewise.
2752 (tree_log2): Likewise.
2753 (tree_floor_log2): Likewise.
2754 (tree_ctz): Likewise.
2755 (int_size_in_bytes): Likewise.
2756 (mem_ref_offset): Return an offset_int rather than a double_int.
2757 (build_type_attribute_qual_variant): Use wide_int interfaces.
2758 (type_hash_eq): Likewise
2759 (tree_int_cst_equal): Likewise.
2760 (tree_int_cst_lt): Delete.
2761 (tree_int_cst_compare): Likewise.
2762 (tree_fits_shwi_p): Use wide_int interfaces.
2763 (tree_fits_uhwi_p): Likewise.
2764 (tree_int_cst_sign_bit): Likewise.
2765 (tree_int_cst_sgn): Likewise.
2766 (tree_int_cst_min_precision): Take a signop rather than a bool.
2767 (simple_cst_equal): Use wide_int interfaces.
2768 (compare_tree_int): Likewise.
2769 (iterative_hash_expr): Likewise.
2770 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2772 (get_type_static_bounds): Use wide_int interfaces.
2773 (tree_int_cst_elt_check_failed): New.
2774 (build_common_tree_nodes): Reordered to set prec before filling in
2776 (int_cst_value): Check cst_and_fits_in_hwi.
2777 (widest_int_cst_value): Use wide_int interfaces.
2778 (upper_bound_in_type): Likewise.
2779 (lower_bound_in_type): Likewise.
2780 (num_ending_zeros): Likewise.
2781 (drop_tree_overflow): Likewise.
2782 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2783 (gen_conditions_for_pow_cst_base): Likewise.
2784 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2785 (group_case_labels_stmt): Use wide-int interfaces.
2786 (verify_gimple_assign_binary): Likewise.
2787 (print_loop): Likewise.
2788 * tree-chrec.c (tree_fold_binomial): Likewise.
2789 * tree-core.h (struct tree_base): Add int_length.
2790 (struct tree_int_cst): Change rep of value.
2791 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2792 (dr_may_alias_p): Likewise.
2793 (max_stmt_executions_tree): Likewise.
2794 * tree.def (INTEGER_CST): Update comment.
2795 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2796 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2797 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2798 (dequeue_and_dump): Use wide-int interfaces.
2799 * tree.h: Include wide-int.h.
2800 (NULL_TREE): Moved to earlier loc in file.
2801 (TREE_INT_CST_ELT_CHECK): New.
2802 (tree_int_cst_elt_check_failed): New.
2804 (TREE_INT_CST): Delete.
2805 (TREE_INT_CST_LOW): Use wide-int interfaces.
2806 (TREE_INT_CST_HIGH): Delete.
2807 (TREE_INT_CST_NUNITS): New.
2808 (TREE_INT_CST_EXT_NUNITS): Likewise.
2809 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2810 (TREE_INT_CST_ELT): Likewise.
2811 (INT_CST_LT): Delete.
2812 (tree_int_cst_elt_check): New (two forms).
2813 (type_code_size): Update comment.
2814 (make_int_cst_stat, make_int_cst): New.
2815 (tree_to_double_int): Delete.
2816 (double_int_fits_to_tree_p): Delete.
2817 (force_fit_type_double): Delete.
2818 (build_int_cstu): Replace with out-of-line function.
2819 (build_int_cst_wide): Delete.
2820 (tree_int_cst_lt): Define inline.
2821 (tree_int_cst_le): New.
2822 (tree_int_cst_compare): Define inline.
2823 (tree_int_cst_min_precision): Take a signop rather than a bool.
2824 (wi::int_traits <const_tree>): New.
2825 (wi::int_traits <tree>): New.
2826 (wi::extended_tree): New.
2827 (wi::int_traits <wi::extended_tree>): New.
2828 (wi::to_widest): New.
2829 (wi::to_offset): New.
2830 (wi::fits_to_tree_p): New.
2831 (wi::min_value): New.
2832 (wi::max_value): New.
2833 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2834 (copy_tree_body_r): Likewise.
2835 * tree-object-size.c (compute_object_offset): Likewise.
2836 (addr_object_size): Likewise.
2837 * tree-predcom.c: Include wide-int-print.h.
2838 (struct dref_d): Change type of offset to widest_int.
2839 (dump_dref): Call wide-int printer.
2840 (aff_combination_dr_offset): Use wide-int interfaces.
2841 (determine_offset): Take a widest_int pointer rather than a
2843 (split_data_refs_to_components): Use wide-int interfaces.
2844 (suitable_component_p): Likewise.
2845 (order_drefs): Likewise.
2846 (add_ref_to_chain): Likewise.
2847 (valid_initializer_p): Likewise.
2848 (determine_roots_comp): Likewise.
2849 * tree-pretty-print.c: Include wide-int-print.h.
2850 (dump_generic_node): Use wide-int interfaces.
2851 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2852 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2853 (move_fixed_address_to_symbol): Likewise.
2854 (move_hint_to_base): Likewise.
2855 (move_pointer_to_base): Likewise.
2856 (move_variant_to_index): Likewise.
2857 (most_expensive_mult_to_index): Likewise.
2858 (addr_to_parts): Likewise.
2859 (copy_ref_info): Likewise.
2860 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2861 (indirect_refs_may_alias_p): Likewise.
2862 (stmt_kills_ref_p_1): Likewise.
2863 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2864 * tree-ssa-ccp.c: Update comment at top of file. Include
2866 (struct prop_value_d): Change type of mask to widest_int.
2867 (extend_mask): New function.
2868 (dump_lattice_value): Use wide-int interfaces.
2869 (get_default_value): Likewise.
2870 (set_constant_value): Likewise.
2871 (set_value_varying): Likewise.
2872 (valid_lattice_transition): Likewise.
2873 (set_lattice_value): Likewise.
2874 (value_to_double_int): Delete.
2875 (value_to_wide_int): New.
2876 (get_value_from_alignment): Use wide-int interfaces.
2877 (get_value_for_expr): Likewise.
2878 (do_dbg_cnt): Likewise.
2879 (ccp_finalize): Likewise.
2880 (ccp_lattice_meet): Likewise.
2881 (bit_value_unop_1): Use widest_ints rather than double_ints.
2882 (bit_value_binop_1): Likewise.
2883 (bit_value_unop): Use wide-int interfaces.
2884 (bit_value_binop): Likewise.
2885 (bit_value_assume_aligned): Likewise.
2886 (evaluate_stmt): Likewise.
2887 (ccp_fold_stmt): Likewise.
2888 (visit_cond_stmt): Likewise.
2889 (ccp_visit_stmt): Likewise.
2890 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2891 (constant_pointer_difference): Likewise.
2892 (associate_pointerplus): Likewise.
2893 (combine_conversions): Likewise.
2894 * tree-ssa-loop.h: Include wide-int.h.
2895 (struct tree_niter_desc): Change type of max to widest_int.
2896 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2897 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2898 (remove_redundant_iv_tests): Likewise.
2899 (canonicalize_loop_induction_variables): Likewise.
2900 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2901 (constant_multiple_of): Take a widest_int pointer instead of
2902 a double_int pointer.
2903 (get_computation_aff): Use wide-int interfaces.
2904 (ptr_difference_cost): Likewise.
2905 (difference_cost): Likewise.
2906 (get_loop_invariant_expr_id): Likewise.
2907 (get_computation_cost_at): Likewise.
2908 (iv_elimination_compare_lt): Likewise.
2909 (may_eliminate_iv): Likewise.
2910 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2911 instead of double_int.
2912 (max_loop_iterations): Likewise.
2913 (max_stmt_executions): Likewise.
2914 (estimated_stmt_executions): Likewise.
2915 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2916 (split_to_var_and_offset): Use wide-int interfaces.
2917 (determine_value_range): Likewise.
2918 (bound_difference_of_offsetted_base): Likewise.
2919 (bounds_add): Take a widest_int instead of a double_int.
2920 (number_of_iterations_ne_max): Use wide-int interfaces.
2921 (number_of_iterations_ne): Likewise.
2922 (number_of_iterations_lt_to_ne): Likewise.
2923 (assert_loop_rolls_lt): Likewise.
2924 (number_of_iterations_lt): Likewise.
2925 (number_of_iterations_le): Likewise.
2926 (number_of_iterations_cond): Likewise.
2927 (number_of_iterations_exit): Likewise.
2928 (finite_loop_p): Likewise.
2929 (derive_constant_upper_bound_assign): Likewise.
2930 (derive_constant_upper_bound): Return a widest_int.
2931 (derive_constant_upper_bound_ops): Likewise.
2932 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2933 (record_estimate): Take a widest_int rather than a double_int.
2934 (record_nonwrapping_iv): Use wide-int interfaces.
2935 (double_int_cmp): Delete.
2936 (wide_int_cmp): New.
2937 (bound_index): Take a widest_int rather than a double_int.
2938 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2939 (maybe_lower_iteration_bound): Likewise.
2940 (estimate_numbers_of_iterations_loop): Likewise.
2941 (estimated_loop_iterations): Take a widest_int pointer than than
2942 a double_int pointer.
2943 (estimated_loop_iterations_int): Use wide-int interfaces.
2944 (max_loop_iterations): Take a widest_int pointer than than
2945 a double_int pointer.
2946 (max_loop_iterations_int): Use wide-int interfaces.
2947 (max_stmt_executions): Take a widest_int pointer than than
2948 a double_int pointer.
2949 (estimated_stmt_executions): Likewise.
2950 (n_of_executions_at_most): Use wide-int interfaces.
2951 (scev_probably_wraps_p): Likewise.
2952 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2954 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2956 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2957 double_ints. Adjust for trailing_wide_ints <3> representation.
2958 (set_nonzero_bits): Likewise.
2959 (get_range_info): Return wide_ints rather than double_ints.
2960 Adjust for trailing_wide_ints <3> representation.
2961 (get_nonzero_bits): Likewise.
2962 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2964 * tree-ssanames.h (struct range_info_def): Replace min, max and
2965 nonzero_bits with a trailing_wide_ints <3>.
2966 (set_range_info): Use wide_int_refs rather than double_ints.
2967 (set_nonzero_bits): Likewise.
2968 (get_range_info): Return wide_ints rather than double_ints.
2969 (get_nonzero_bits): Likewise.
2970 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2971 * tree-ssa-pre.c (phi_translate_1): Likewise.
2972 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2973 (acceptable_pow_call): Likewise.
2974 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2976 (vn_reference_fold_indirect): Likewise.
2977 (vn_reference_maybe_forwprop_address): Likewise.
2978 (valueize_refs_1): Likewise.
2979 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2980 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2981 tree_int_cst_lt and tree_int_cst_le.
2982 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2984 (streamer_alloc_tree): Likewise.
2985 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2986 (streamer_write_tree_header): Likewise.
2987 (streamer_write_integer_cst): Likewise.
2988 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2989 (build_constructors): Likewise.
2990 (array_value_type): Likewise.
2991 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2992 (vect_check_gather): Likewise.
2993 * tree-vect-generic.c (build_replicated_const): Likewise.
2994 (expand_vector_divmod): Likewise.
2995 * tree-vect-loop.c (vect_transform_loop): Likewise.
2996 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2997 (vect_do_peeling_for_alignment): Likewise.
2998 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2999 * tree-vrp.c: Include wide-int.h.
3000 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3001 (extract_range_from_assert): Use wide-int interfaces.
3002 (vrp_int_const_binop): Likewise.
3003 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3004 double_int pointers.
3005 (ranges_from_anti_range): Use wide-int interfaces.
3006 (quad_int_cmp): Delete.
3007 (quad_int_pair_sort): Likewise.
3008 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3009 (extract_range_from_unary_expr_1): Likewise.
3010 (adjust_range_with_scev): Likewise.
3011 (masked_increment): Take and return wide_ints rather than double_ints.
3012 (register_edge_assert_for_2): Use wide-int interfaces.
3013 (check_array_ref): Likewise.
3014 (search_for_addr_array): Likewise.
3015 (maybe_set_nonzero_bits): Likewise.
3016 (union_ranges): Pass an integer of the correct type instead of
3017 using integer_one_node.
3018 (intersect_ranges): Likewise.
3019 (simplify_truth_ops_using_ranges): Likewise.
3020 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3021 (range_fits_type_p): Likewise.
3022 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3024 (simplify_conversion_using_ranges): Use wide-int interfaces.
3025 (simplify_float_conversion_using_ranges): Likewise.
3026 (vrp_finalize): Likewise.
3027 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3028 (gimple_stringops_transform): Likewise.
3029 * varasm.c (decode_addr_const): Likewise.
3030 (const_hash_1): Likewise.
3031 (const_rtx_hash_1): Likewise
3032 (output_constant): Likewise.
3033 (array_size_for_constructor): Likewise.
3034 (output_constructor_regular_field): Likewise.
3035 (output_constructor_bitfield): Likewise.
3036 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3037 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3039 * gencheck.c: Define BITS_PER_UNIT.
3042 * wide-int-print.cc: New.
3043 * wide-int-print.h: New.
3045 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3047 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3049 2014-05-06 Richard Biener <rguenther@suse.de>
3051 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3052 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3053 (TODO_verify_all): Adjust.
3054 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3055 TODO_verify_stmts and TODO_verify_rtl_sharing.
3056 * bb-reorder.c: Likewise.
3057 * cfgexpand.c: Likewise.
3058 * cprop.c: Likewise.
3060 * function.c: Likewise.
3061 * fwprop.c: Likewise.
3063 * gimple-ssa-isolate-paths.c: Likewise.
3064 * gimple-ssa-strength-reduction.c: Likewise.
3065 * ipa-split.c: Likewise.
3066 * loop-init.c: Likewise.
3067 * loop-unroll.c: Likewise.
3068 * lower-subreg.c: Likewise.
3069 * modulo-sched.c: Likewise.
3070 * postreload-gcse.c: Likewise.
3071 * predict.c: Likewise.
3072 * recog.c: Likewise.
3073 * sched-rgn.c: Likewise.
3074 * store-motion.c: Likewise.
3075 * tracer.c: Likewise.
3076 * trans-mem.c: Likewise.
3077 * tree-call-cdce.c: Likewise.
3078 * tree-cfg.c: Likewise.
3079 * tree-cfgcleanup.c: Likewise.
3080 * tree-complex.c: Likewise.
3081 * tree-eh.c: Likewise.
3082 * tree-emutls.c: Likewise.
3083 * tree-if-conv.c: Likewise.
3084 * tree-into-ssa.c: Likewise.
3085 * tree-loop-distribution.c: Likewise.
3086 * tree-object-size.c: Likewise.
3087 * tree-parloops.c: Likewise.
3088 * tree-pass.h: Likewise.
3089 * tree-sra.c: Likewise.
3090 * tree-ssa-ccp.c: Likewise.
3091 * tree-ssa-copy.c: Likewise.
3092 * tree-ssa-copyrename.c: Likewise.
3093 * tree-ssa-dce.c: Likewise.
3094 * tree-ssa-dom.c: Likewise.
3095 * tree-ssa-dse.c: Likewise.
3096 * tree-ssa-forwprop.c: Likewise.
3097 * tree-ssa-ifcombine.c: Likewise.
3098 * tree-ssa-loop-ch.c: Likewise.
3099 * tree-ssa-loop-ivcanon.c: Likewise.
3100 * tree-ssa-loop.c: Likewise.
3101 * tree-ssa-math-opts.c: Likewise.
3102 * tree-ssa-phiopt.c: Likewise.
3103 * tree-ssa-phiprop.c: Likewise.
3104 * tree-ssa-pre.c: Likewise.
3105 * tree-ssa-reassoc.c: Likewise.
3106 * tree-ssa-sink.c: Likewise.
3107 * tree-ssa-strlen.c: Likewise.
3108 * tree-ssa-tail-merge.c: Likewise.
3109 * tree-ssa-uncprop.c: Likewise.
3110 * tree-switch-conversion.c: Likewise.
3111 * tree-tailcall.c: Likewise.
3112 * tree-vect-generic.c: Likewise.
3113 * tree-vectorizer.c: Likewise.
3114 * tree-vrp.c: Likewise.
3116 * var-tracking.c: Likewise.
3117 * bt-load.c: Likewise.
3118 * cfgcleanup.c: Likewise.
3119 * combine-stack-adj.c: Likewise.
3120 * combine.c: Likewise.
3121 * compare-elim.c: Likewise.
3122 * config/epiphany/resolve-sw-modes.c: Likewise.
3123 * config/i386/i386.c: Likewise.
3124 * config/mips/mips.c: Likewise.
3125 * config/s390/s390.c: Likewise.
3126 * config/sh/sh_treg_combine.cc: Likewise.
3127 * config/sparc/sparc.c: Likewise.
3130 * final.c: Likewise.
3131 * ifcvt.c: Likewise.
3132 * mode-switching.c: Likewise.
3133 * passes.c: Likewise.
3134 * postreload.c: Likewise.
3136 * reg-stack.c: Likewise.
3137 * regcprop.c: Likewise.
3138 * regrename.c: Likewise.
3141 2014-05-06 Richard Biener <rguenther@suse.de>
3144 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3145 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3147 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3150 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3152 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3153 Tom de Vries <tom@codesourcery.com>
3155 * target.def (call_fusage_contains_non_callee_clobbers): New
3157 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3159 (@node Miscellaneous Register Hooks): New node.
3160 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3161 * doc/tm.texi: Regenerate.
3163 2014-05-05 Marek Polacek <polacek@redhat.com>
3166 * opts.c (common_handle_option): Call error_at instead of warning_at.
3168 2014-05-05 Richard Biener <rguenther@suse.de>
3170 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3171 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3172 under the TODO_verify_il umbrella.
3174 2014-05-05 Richard Biener <rguenther@suse.de>
3176 * passes.c (execute_function_todo): Move TODO_verify_flow under
3177 the TODO_verify_ul umbrella.
3179 2014-05-05 Richard Biener <rguenther@suse.de>
3182 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3183 X & CST away from a CST that is the mask of a mode.
3185 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3187 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3188 int argument to enum machine_mode.
3189 (picochip_class_max_nregs): Ditto.
3190 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3191 (picochip_class_max_nregs): Ditto.
3193 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3195 * target.def: Add new target hook.
3196 * doc/tm.texi: Regenerate.
3197 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3198 * targhooks.c (default_keep_leaf_when_profiled): New function.
3200 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3201 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3203 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3205 PR tree-optimization/60363
3206 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3207 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3208 (update_destination_phis): New parameter.
3209 (create_edge_and_update_destination_phis): Ditto.
3210 (ssa_fix_duplicate_block_edges): Pass new arguments.
3211 (thread_single_edge): Ditto.
3213 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3215 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3216 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3217 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3218 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3219 Use RS6000_BTM_HARD_FLOAT.
3220 (BU_MISC_2): Likewise.
3221 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3222 RS6000_BTM_HARD_FLOAT.
3223 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3225 (rs6000_invalid_builtin): Add hard floating builtin support.
3226 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3227 hard float builtins.
3228 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3230 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3232 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3233 Add missing function* argument.
3235 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3237 * lra-constraints.c (valid_address_p): Move earlier in file.
3238 Add a constraint argument to the address_info version.
3239 (satisfies_memory_constraint_p): New function.
3240 (satisfies_address_constraint_p): Likewise.
3241 (process_alt_operands, curr_insn_transform): Use them.
3242 (process_address): Pass the constraint to valid_address_p when
3243 checking address operands.
3245 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3247 * config/mips/mips.c (mips_isa_rev): New variable.
3248 (mips_set_architecture): Set it.
3249 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3251 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3252 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3253 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3254 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3255 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3256 conditions in terms of mips_isa_rev.
3257 (mips_isa_rev): Declare.
3259 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3261 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3262 (prob_unlikely, prob_likely): Make variables const.
3264 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3266 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3268 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3270 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3272 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3274 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3275 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3276 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3278 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3279 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3281 Replace usage of ROUND_REG with sh_round_reg.
3282 Use CEIL instead of ROUND_ADVANCE.
3284 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3287 * config/sh/sh.c: Include stdlib headers before everything else.
3289 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3291 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3292 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3293 (gimplify_adjust_omp_clauses): Simd region is never
3294 directly nested in combined parallel. Instead, for linear
3295 with copyin/copyout, if in combined for simd loop, make decl
3296 firstprivate/lastprivate on OMP_FOR.
3297 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3298 expand_omp_for_static_chunk): When setting endvar, also set
3299 fd->loop.v to the same value.
3301 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3303 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3305 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3307 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3310 2014-05-02 Marek Polacek <polacek@redhat.com>
3312 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3314 2014-05-02 Kito Cheng <kito@0xlab.org>
3316 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3317 to a C expression marco.
3318 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3319 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3320 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3321 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3322 HONOR_REG_ALLOC_ORDER.
3323 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3325 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3327 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3329 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3331 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3333 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3335 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3336 (convert_scalar_cond_reduction): Likewise.
3337 (predicate_scalar_phi): Add recognition and transformation
3338 of simple conditioanl reduction to be vectorizable.
3340 2014-05-01 Marek Polacek <polacek@redhat.com>
3343 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3345 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3347 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3348 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3349 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3350 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3351 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3352 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3353 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3354 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3356 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3358 * config/arc/arc.opt (mlra): Move comment above option name
3359 to avoid mis-parsing as language options.
3361 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3363 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3364 * config/sol2.h: ... here.
3365 * config/sol2-10.h: Remove.
3367 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3368 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3369 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3370 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3371 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3372 * config/sol2.h: ... here.
3373 (SECTION_NAME_FORMAT): Don't redefine.
3374 (STARTFILE_ARCH32_SPEC): Rename to ...
3375 (STARTFILE_ARCH_SPEC): ... this.
3376 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3377 * config/sparc/sol2.h: ... here.
3378 (SECTION_NAME_FORMAT): Don't undef.
3379 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3380 (SUBTARGET_EXTRA_SPECS): Remove.
3381 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3383 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3384 (MD_STARTFILE_PREFIX): Remove.
3385 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3386 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3387 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3388 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3389 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3390 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3391 * config/i386/sol2.h: ... here.
3392 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3393 * config/i386/sol2-bi.h: Remove.
3394 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3395 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3397 * config/i386/t-sol2-64: Rename to ...
3398 * config/i386/t-sol2: ... this.
3399 * config/sparc/t-sol2-64: Rename to ...
3400 * config/sparc/t-sol2: ... this.
3402 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3403 sol2_tm_file_head, sol2_tm_file_tail.
3404 Include ${cpu_type}/sol2.h before sol2.h.
3406 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3407 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3408 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3409 Reflect i386/t-sol2-64 renaming.
3410 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3411 Reflect sparc/t-sol2-64 renaming.
3413 2014-04-30 Richard Biener <rguenther@suse.de>
3415 * passes.c (execute_function_todo): Move TODO_verify_stmts
3416 and TODO_verify_ssa under the TODO_verify_il umbrella.
3417 * tree-ssa.h (verify_ssa): Adjust prototype.
3418 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3419 we should verify SSA operands.
3420 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3421 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3422 whether we should verify whether not throwing stmts have EH info.
3423 * graphite-scop-detection.c (create_sese_edges): Adjust.
3424 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3425 * tree-eh.c (lower_try_finally_switch): Do not add the
3426 default case label twice.
3428 2014-04-30 Marek Polacek <polacek@redhat.com>
3430 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3431 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3432 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3433 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3435 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3437 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3438 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3439 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3440 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3441 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3442 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3443 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3444 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3446 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3448 * tree-cfg.c (dump_function_to_file): Dump the return type of
3449 functions, in a line to itself before the function body, mimicking
3450 the layout of a C function.
3452 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3454 PR tree-optimization/60971
3455 * tree-tailcall.c (process_assignment): Reject conversions which
3458 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3460 * calls.c (initialize_argument_information): Always treat
3461 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3462 (expand_call): Likewise.
3463 (emit_library_call_calue_1): Likewise.
3464 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3465 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3468 2014-04-29 Nick Clifton <nickc@redhat.com>
3470 * config/msp430/msp430.md (umulsidi): Fix typo.
3471 (mulhisi3): Enable even inside interrupt handlers.
3472 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3473 bigger return address pushed in large mode.
3475 2014-04-29 Nick Clifton <nickc@redhat.com>
3477 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3478 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3480 * config/m32r/m32r.c (init_reg_tables): Likewise.
3481 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3482 enum to hold the modes.
3484 2014-04-29 Richard Biener <rguenther@suse.de>
3486 * dominance.c (free_dominance_info): Add overload with
3488 (dom_info_state): Likewise.
3489 (dom_info_available_p): Likewise.
3490 * basic-block.h (free_dominance_info, dom_info_state,
3491 dom_info_available_p): Declare overloads.
3492 * passes.c (execute_function_todo): Verify that verifiers
3493 don't change dominator info state. Drop dominator info
3494 for IPA pass invocations.
3495 * cgraph.c (release_function_body): Restore asserts that
3496 dominator information is released.
3498 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3500 * doc/invoke.texi: Fix typo.
3501 * tree-vrp.c: Fix typos.
3502 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3504 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3506 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3508 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3510 * config/aarch64/aarch64-builtins.c
3511 (aarch64_types_storestruct_lane_qualifiers): New.
3512 (TYPES_STORESTRUCT_LANE): Likewise.
3513 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3514 (st3_lane): Likewise.
3515 (st4_lane): Likewise.
3516 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3517 (vec_store_lanesci_lane<mode>): Likewise.
3518 (vec_store_lanesxi_lane<mode>): Likewise.
3519 (aarch64_st2_lane<VQ:mode>): Likewise.
3520 (aarch64_st3_lane<VQ:mode>): Likewise.
3521 (aarch64_st4_lane<VQ:mode>): Likewise.
3522 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3523 * config/aarch64/arm_neon.h
3524 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3525 use new macro arguments.
3526 (__ST3_LANE_FUNC): Likewise.
3527 (__ST4_LANE_FUNC): Likewise.
3528 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3529 (V_THREE_ELEM): Likewise.
3530 (V_FOUR_ELEM): Likewise.
3532 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3534 * doc/gimple.texi: Replace the description of the now-defunct
3535 union gimple_statement_d with a diagram showing the
3536 gimple_statement_base class hierarchy and its relationships to
3537 the GSS_ and GIMPLE_ enums.
3539 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3541 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3542 * config/aarch64/aarch64.c
3543 (aarch64_cannot_change_mode_class): Weaken conditions.
3544 (aarch64_modes_tieable_p): New.
3545 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3547 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3549 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3550 (loadsync_<mode>): Change mode.
3551 (load_quadpti, store_quadpti): New.
3552 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3553 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3555 2014-04-28 Martin Jambor <mjambor@suse.cz>
3557 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3558 same alias type as the original statement.
3559 (subreplacement_assignment_data): New type.
3560 (handle_unscalarized_data_in_subtree): New type of parameter,
3561 generate new memory accesses with same alias type as the original
3563 (load_assign_lhs_subreplacements): Likewise.
3564 (sra_modify_constructor_assign): Generate new memory accesses with
3565 same alias type as the original statement.
3567 2014-04-28 Richard Biener <rguenther@suse.de>
3569 * tree-pass.h (TODO_verify_il): Define.
3570 (TODO_verify_all): Complete properly.
3571 * passes.c (execute_function_todo): Move existing loop-closed
3572 SSA verification under TODO_verify_il.
3573 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3574 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3575 Fix tree sharing issue.
3577 2014-04-28 Richard Biener <rguenther@suse.de>
3580 * builtins.def (DEF_C11_BUILTIN): Add.
3581 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3582 * coretypes.h (enum function_class): Add function_c11_misc.
3583 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3584 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3585 (call_may_clobber_ref_p_1): Likewise.
3586 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3587 (mark_all_reaching_defs_necessary_1): Likewise.
3588 (propagate_necessity): Likewise.
3589 (eliminate_unnecessary_stmts): Likewise.
3590 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3592 2014-04-28 Richard Biener <rguenther@suse.de>
3594 * tree-vrp.c (vrp_var_may_overflow): Remove.
3595 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3596 with overflow immediately bump to one before that value and
3597 let iteration figure out overflow status.
3599 2014-04-28 Richard Biener <rguenther@suse.de>
3601 * configure.ac: Do valgrind header checks unconditionally.
3602 Add --enable-valgrind-annotations.
3603 * system.h: Guard valgrind header inclusion with
3604 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3605 * alloc-pool.c (pool_alloc, pool_free): Use
3606 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3607 to guard possibly dead code.
3608 * config.in: Regenerated.
3609 * configure: Likewise.
3611 2014-04-28 Jeff Law <law@redhat.com>
3613 PR tree-optimization/60902
3614 * tree-ssa-threadedge.c
3615 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3616 over real defs when invalidating outputs from statements that do not
3617 produce useful outputs for threading.
3619 2014-04-28 Richard Biener <rguenther@suse.de>
3621 PR tree-optimization/60979
3622 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3623 SCOPs that end in a block with a successor with abnormal
3626 2014-04-28 Richard Biener <rguenther@suse.de>
3628 * tree-pass.h (execute_pass_list): Adjust prototype.
3629 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3630 (do_per_function): Change callback signature, push all actual
3631 work to the callbals.
3632 (do_per_function_toporder): Likewise.
3633 (execute_function_dump): Adjust.
3634 (execute_function_todo): Likewise.
3635 (clear_last_verified): Likewise.
3636 (verify_curr_properties): Likewise.
3637 (update_properties_after_pass): Likewise.
3638 (execute_pass_list_1): Split out from ...
3639 (execute_pass_list): ... here. Adjust.
3640 (execute_ipa_pass_list): Likewise.
3641 * cgraphunit.c (cgraph_add_new_function): Adjust.
3642 (analyze_function): Likewise.
3643 (expand_function): Likewise.
3644 * cgraph.c (release_function_body): Free dominance info
3645 here instead of asserting it was magically freed elsewhere.
3647 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3649 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3650 * configure: Regenerate.
3651 * config/sparc/sparc.opt (muser-mode): New option.
3652 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3654 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3655 * doc/invoke.texi (SPARC options): Document -muser-mode.
3657 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3659 * cselib.c (find_slot_memmode): Delete.
3660 (cselib_hasher): Change compare_type to a struct.
3661 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3663 (preserve_constants_and_equivs): Adjust for new compare_type.
3664 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3665 (wrap_constant): Delete.
3666 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3668 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3670 * doc/install.texi (Building with profile feedback): Remove
3673 2014-04-26 Tom de Vries <tom@codesourcery.com>
3675 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3678 2014-04-25 Cary Coutant <ccoutant@google.com>
3681 * dwarf2out.c (should_move_die_to_comdat): A type definition
3682 can contain a subprogram definition, but don't move it to a
3684 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3685 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3687 (clone_tree_hash): Rename to...
3688 (clone_tree_partial): ...this; change callers. Copy
3689 DW_TAG_subprogram DIEs as declarations.
3690 (copy_decls_walk): Don't copy children of a declaration into a
3693 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3696 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3699 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3701 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3702 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3704 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3707 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3709 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3711 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3713 PR tree-optimization/60930
3714 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3715 creating a multiply candidate by folding two constant
3716 multiplicands when the result overflows.
3718 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3720 PR tree-optimization/60960
3721 * tree-vect-generic.c (expand_vector_operation): Only call
3722 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3724 2014-04-25 Tom de Vries <tom@codesourcery.com>
3726 * expr.c (clobber_reg_mode): New function.
3727 * expr.h (clobber_reg): New function.
3729 2014-04-25 Tom de Vries <tom@codesourcery.com>
3731 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3734 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3735 Tom de Vries <tom@codesourcery.com>
3737 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3739 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3740 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3741 new argument to find_all_hard_reg_sets call.
3743 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3745 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3746 Use HOST_WIDE_INT_C for mask literal.
3747 (aarch_rev16_shleft_mask_imm_p): Likewise.
3749 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3752 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3754 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3756 PR preprocessor/56540
3757 * config/i386/i386-c.c (ix86_target_macros): Define
3758 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3760 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3762 * configure.ac (tga_func): Remove.
3763 (LIB_TLS_SPEC): Remove.
3764 * configure: Regenerate.
3765 * config.in: Regenerate.
3766 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3768 2014-04-25 Richard Biener <rguenther@suse.de>
3771 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3772 call stmt use/clobber sets during stmt walk instead of
3773 walking the possibly incomplete set of caller edges.
3775 2014-04-25 Richard Biener <rguenther@suse.de>
3778 * passes.c (apply_ipa_transforms): Inline into only caller ...
3779 (execute_one_pass): ... here. Properly bring in function
3780 bodies for nodes we want to apply IPA transforms to.
3782 2014-04-24 Cong Hou <congh@google.com>
3784 PR tree-optimization/60896
3785 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3786 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3787 (vect_mark_pattern_stmts): Set the def type of all statements in
3788 PATTERN_DEF_SEQ as vect_internal_def.
3790 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3792 * doc/extend.texi (PowerPC Built-in Functions): Document new
3793 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3794 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3796 * config/rs6000/predicates.md (const_0_to_3_operand): New
3797 predicate to match 0..3 integer constants.
3799 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3800 to support adding miscellaneous builtin functions.
3801 (BU_DFP_MISC_2): Likewise.
3802 (BU_P7_MISC_1): Likewise.
3803 (BU_P7_MISC_2): Likewise.
3804 (BU_P8V_MISC_3): Likewise.
3805 (BU_MISC_1): Likewise.
3806 (BU_MISC_2): Likewise.
3807 (DIVWE): Add extended divide builtin functions.
3810 (DIVWEUO): Likewise.
3814 (DIVDEUO): Likewise.
3815 (DXEX): Add decimal floating-point builtin functions.
3818 (DDEDPDQ): Likewise.
3820 (DENBCDQ): Likewise.
3827 (CDTBCD): Add new BCD builtin functions.
3831 (BCDADD_LT): Likewise.
3832 (BCDADD_EQ): Likewise.
3833 (BCDADD_GT): Likewise.
3834 (BCDADD_OV): Likewise.
3836 (BCDSUB_LT): Likewise.
3837 (BCDSUB_EQ): Likewise.
3838 (BCDSUB_GT): Likewise.
3839 (BCDSUB_OV): Likewise.
3840 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3841 (UNPACK_TD): Likewise.
3842 (PACK_TF): Likewise.
3843 (UNPACK_TF): Likewise.
3844 (UNPACK_TF_0): Likewise.
3845 (UNPACK_TF_1): Likewise.
3846 (PACK_V1TI): Likewise.
3847 (UNPACK_V1TI): Likewise.
3849 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3850 support for decimal floating point builtin functions.
3851 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3852 functions that take constant arguments.
3853 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3854 (rs6000_init_builtins): Setup long double, _Decimal64, and
3855 _Decimal128 types for new builtin functions.
3856 (builtin_function_type): Set the unsigned flags appropriately for
3857 the new builtin functions.
3858 (rs6000_opt_masks): Add support for decimal floating point builtin
3861 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3862 floating point builtin functions.
3863 (RS6000_BTM_COMMON): Likewise.
3864 (RS6000_BTI_long_double): Likewise.
3865 (RS6000_BTI_dfloat64): Likewise.
3866 (RS6000_BTI_dfloat128): Likewise.
3867 (long_double_type_internal_node): Likewise.
3868 (dfloat64_type_internal_node): Likewise.
3869 (dfloat128_type_internal_node): Likewise.
3871 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3872 2.07 bcd arithmetic instructions.
3873 (UNSPEC_BCDSUB): Likewise.
3874 (UNSPEC_BCD_OVERFLOW): Likewise.
3875 (UNSPEC_BCD_ADD_SUB): Likewise.
3876 (bcd_add_sub): Likewise.
3877 (BCD_TEST): Likewise.
3878 (bcd<bcd_add_sub>): Likewise.
3879 (bcd<bcd_add_sub>_test): Likewise.
3880 (bcd<bcd_add_sub>_test2): Likewise.
3881 (bcd<bcd_add_sub>_<code>): Likewise.
3882 (peephole2 for combined bcd ops): Likewise.
3884 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3885 decimal floating point builtin functions.
3886 (UNSPEC_DENBCD): Likewise.
3887 (UNSPEC_DXEX): Likewise.
3888 (UNSPEC_DIEX): Likewise.
3889 (UNSPEC_DSCLI): Likewise.
3890 (UNSPEC_DSCRI): Likewise.
3891 (D64_D128): Likewise.
3892 (dfp_suffix): Likewise.
3893 (dfp_ddedpd_<mode>): Likewise.
3894 (dfp_denbcd_<mode>): Likewise.
3895 (dfp_dxex_<mode>): Likewise.
3896 (dfp_diex_<mode>): Likewise.
3897 (dfp_dscli_<mode>): Likewise.
3898 (dfp_dscri_<mode>): Likewise.
3900 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3902 (UNSPEC_CDTBCD): Likewise.
3903 (UNSPEC_CBCDTD): Likewise.
3904 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3905 (UNSPEC_DIVEO): Likewise.
3906 (UNSPEC_DIVEU): Likewise.
3907 (UNSPEC_DIVEUO): Likewise.
3908 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3909 pack/unpack 128-bit types.
3910 (UNSPEC_PACK_128BIT): Likewise.
3911 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3912 (udiv<mode>3): Use idiv_ldiv mode attribute.
3913 (div<mode>3): Likewise.
3914 (addg6s): Add new BCD builtin functions.
3917 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3918 (div_extend): Likewise.
3919 (div<div_extend>_<mode>"): Likewise.
3920 (FP128_64): Add support for new builtin functions to pack/unpack
3922 (unpack<mode>): Likewise.
3923 (unpacktf_0): Likewise.
3924 (unpacktf_1): Likewise.
3925 (unpack<mode>_dm): Likewise.
3926 (unpack<mode>_nodm): Likewise.
3927 (pack<mode>): Likewise.
3928 (unpackv1ti): Likewise.
3929 (packv1ti): Likewise.
3931 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3933 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3936 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3938 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3939 * gimplify.c (omp_is_private): Change last argument's type to int.
3940 Only diagnose lastprivate if the simd argument is 1, only diagnose
3941 linear if the simd argument is 2.
3942 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3943 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3944 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3945 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3946 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3947 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3948 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3949 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3950 * tree-nested.c (convert_nonlocal_omp_clauses,
3951 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3953 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3956 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3959 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3961 * flag-types.h (enum ivar_visibility): Add.
3963 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3965 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3966 function * argument.
3968 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3970 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3972 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3973 Tom de Vries <tom@codesourcery.com>
3975 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3976 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3978 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3979 * emit-rtl.c (try_split): Same.
3981 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3982 Tom de Vries <tom@codesourcery.com>
3984 * common.opt (fuse-caller-save): New option.
3986 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3988 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3989 elements for big-endian.
3991 2014-04-24 Richard Biener <rguenther@suse.de>
3993 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3994 during TER and instead use the sepops interface for expanding
3995 non-GIMPLE_SINGLE_RHS.
3997 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3999 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4000 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4002 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4004 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4005 assembler 64-bit option.
4006 * configure: Regenerate.
4008 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4010 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4011 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4012 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4013 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4014 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4016 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4018 * config/aarch64/aarch64-builtins.c
4019 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4020 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4021 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4022 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4024 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4025 (Vrevsuff): New mode attribute.
4027 2014-04-24 Terry Guo <terry.guo@arm.com>
4029 * config/arm/arm.h (machine_function): Define variable
4030 after_arm_reorg here.
4031 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4032 (arm_split_constant): Update the way to access variable
4035 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4037 2014-04-23 Tom de Vries <tom@codesourcery.com>
4039 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4041 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4043 * is-a.h: Update comments to reflect the following changes to the
4044 "pointerness" of the API, making the template parameter match the
4045 return type, allowing use of is-a.h with typedefs of pointers.
4046 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4047 that the return type matches the parameter to the is_a_helper.
4049 (dyn_cast): Likewise.
4051 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4052 pointer from the is-a.h API.
4054 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4055 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4057 (is_a_helper <varpool_node>::test): Likewise, convert to...
4058 (is_a_helper <varpool_node *>::test): ...this.
4060 (varpool_first_variable): Update for removal of implicit pointer
4061 from the is-a.h API.
4062 (varpool_next_variable): Likewise.
4063 (varpool_first_static_initializer): Likewise.
4064 (varpool_next_static_initializer): Likewise.
4065 (varpool_first_defined_variable): Likewise.
4066 (varpool_next_defined_variable): Likewise.
4067 (cgraph_first_defined_function): Likewise.
4068 (cgraph_next_defined_function): Likewise.
4069 (cgraph_first_function): Likewise.
4070 (cgraph_next_function): Likewise.
4071 (cgraph_first_function_with_gimple_body): Likewise.
4072 (cgraph_next_function_with_gimple_body): Likewise.
4073 (cgraph_alias_target): Likewise.
4074 (varpool_alias_target): Likewise.
4075 (cgraph_function_or_thunk_node): Likewise.
4076 (varpool_variable_node): Likewise.
4077 (symtab_real_symbol_p): Likewise.
4078 * cgraphunit.c (referred_to_p): Likewise.
4079 (analyze_functions): Likewise.
4080 (handle_alias_pairs): Likewise.
4081 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4082 * gimple-ssa.h (gimple_vuse_op): Likewise.
4083 (gimple_vdef_op): Likewise.
4084 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4085 * gimple.c (gimple_build_asm_1): Likewise.
4086 (gimple_build_try): Likewise.
4087 (gimple_build_resx): Likewise.
4088 (gimple_build_eh_dispatch): Likewise.
4089 (gimple_build_omp_for): Likewise.
4090 (gimple_omp_for_set_clauses): Likewise.
4092 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4093 (is_a_helper <gimple_statement_asm *>::test): ...this.
4094 (is_a_helper <gimple_statement_bind>::test): Convert to...
4095 (is_a_helper <gimple_statement_bind *>::test): ...this.
4096 (is_a_helper <gimple_statement_call>::test): Convert to...
4097 (is_a_helper <gimple_statement_call *>::test): ...this.
4098 (is_a_helper <gimple_statement_catch>::test): Convert to...
4099 (is_a_helper <gimple_statement_catch *>::test): ...this.
4100 (is_a_helper <gimple_statement_resx>::test): Convert to...
4101 (is_a_helper <gimple_statement_resx *>::test): ...this.
4102 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4103 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4104 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4105 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4106 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4107 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4108 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4109 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4110 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4111 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4112 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4113 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4114 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4115 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4116 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4117 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4118 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4119 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4120 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4121 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4122 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4123 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4124 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4125 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4126 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4127 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4128 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4129 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4130 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4131 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4132 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4133 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4134 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4135 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4136 (is_a_helper <gimple_statement_phi>::test): Convert to...
4137 (is_a_helper <gimple_statement_phi *>::test): ...this.
4138 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4139 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4140 (is_a_helper <gimple_statement_try>::test): Convert to...
4141 (is_a_helper <gimple_statement_try *>::test): ...this.
4142 (is_a_helper <gimple_statement_wce>::test): Convert to...
4143 (is_a_helper <gimple_statement_wce *>::test): ...this.
4144 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4145 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4146 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4147 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4148 (is_a_helper <const gimple_statement_call>::test): Convert to...
4149 (is_a_helper <const gimple_statement_call *>::test): ...this.
4150 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4151 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4152 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4153 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4154 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4155 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4156 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4157 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4158 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4160 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4162 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4164 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4166 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4167 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4168 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4170 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4171 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4173 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4174 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4175 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4176 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4177 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4178 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4180 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4181 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4182 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4183 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4185 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4186 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4187 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4188 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4189 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4190 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4191 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4192 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4193 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4194 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4195 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4196 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4197 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4198 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4199 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4200 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4202 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4204 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4205 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4207 (gimple_use_ops): Update for removal of implicit pointer from the
4209 (gimple_set_use_ops): Likewise.
4210 (gimple_vuse): Likewise.
4211 (gimple_vdef): Likewise.
4212 (gimple_vuse_ptr): Likewise.
4213 (gimple_vdef_ptr): Likewise.
4214 (gimple_set_vuse): Likewise.
4215 (gimple_set_vdef): Likewise.
4216 (gimple_omp_return_set_lhs): Likewise.
4217 (gimple_omp_return_lhs): Likewise.
4218 (gimple_omp_return_lhs_ptr): Likewise.
4219 (gimple_call_fntype): Likewise.
4220 (gimple_call_set_fntype): Likewise.
4221 (gimple_call_set_internal_fn): Likewise.
4222 (gimple_call_use_set): Likewise.
4223 (gimple_call_clobber_set): Likewise.
4224 (gimple_bind_vars): Likewise.
4225 (gimple_bind_set_vars): Likewise.
4226 (gimple_bind_body_ptr): Likewise.
4227 (gimple_bind_set_body): Likewise.
4228 (gimple_bind_add_stmt): Likewise.
4229 (gimple_bind_block): Likewise.
4230 (gimple_bind_set_block): Likewise.
4231 (gimple_asm_ninputs): Likewise.
4232 (gimple_asm_noutputs): Likewise.
4233 (gimple_asm_nclobbers): Likewise.
4234 (gimple_asm_nlabels): Likewise.
4235 (gimple_asm_input_op): Likewise.
4236 (gimple_asm_input_op_ptr): Likewise.
4237 (gimple_asm_output_op): Likewise.
4238 (gimple_asm_output_op_ptr): Likewise.
4239 (gimple_asm_set_output_op): Likewise.
4240 (gimple_asm_clobber_op): Likewise.
4241 (gimple_asm_set_clobber_op): Likewise.
4242 (gimple_asm_label_op): Likewise.
4243 (gimple_asm_set_label_op): Likewise.
4244 (gimple_asm_string): Likewise.
4245 (gimple_catch_types): Likewise.
4246 (gimple_catch_types_ptr): Likewise.
4247 (gimple_catch_handler_ptr): Likewise.
4248 (gimple_catch_set_types): Likewise.
4249 (gimple_catch_set_handler): Likewise.
4250 (gimple_eh_filter_types): Likewise.
4251 (gimple_eh_filter_types_ptr): Likewise.
4252 (gimple_eh_filter_failure_ptr): Likewise.
4253 (gimple_eh_filter_set_types): Likewise.
4254 (gimple_eh_filter_set_failure): Likewise.
4255 (gimple_eh_must_not_throw_fndecl): Likewise.
4256 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4257 (gimple_eh_else_n_body_ptr): Likewise.
4258 (gimple_eh_else_e_body_ptr): Likewise.
4259 (gimple_eh_else_set_n_body): Likewise.
4260 (gimple_eh_else_set_e_body): Likewise.
4261 (gimple_try_eval_ptr): Likewise.
4262 (gimple_try_cleanup_ptr): Likewise.
4263 (gimple_try_set_eval): Likewise.
4264 (gimple_try_set_cleanup): Likewise.
4265 (gimple_wce_cleanup_ptr): Likewise.
4266 (gimple_wce_set_cleanup): Likewise.
4267 (gimple_phi_capacity): Likewise.
4268 (gimple_phi_num_args): Likewise.
4269 (gimple_phi_result): Likewise.
4270 (gimple_phi_result_ptr): Likewise.
4271 (gimple_phi_set_result): Likewise.
4272 (gimple_phi_arg): Likewise.
4273 (gimple_phi_set_arg): Likewise.
4274 (gimple_resx_region): Likewise.
4275 (gimple_resx_set_region): Likewise.
4276 (gimple_eh_dispatch_region): Likewise.
4277 (gimple_eh_dispatch_set_region): Likewise.
4278 (gimple_omp_critical_name): Likewise.
4279 (gimple_omp_critical_name_ptr): Likewise.
4280 (gimple_omp_critical_set_name): Likewise.
4281 (gimple_omp_for_clauses): Likewise.
4282 (gimple_omp_for_clauses_ptr): Likewise.
4283 (gimple_omp_for_set_clauses): Likewise.
4284 (gimple_omp_for_collapse): Likewise.
4285 (gimple_omp_for_index): Likewise.
4286 (gimple_omp_for_index_ptr): Likewise.
4287 (gimple_omp_for_set_index): Likewise.
4288 (gimple_omp_for_initial): Likewise.
4289 (gimple_omp_for_initial_ptr): Likewise.
4290 (gimple_omp_for_set_initial): Likewise.
4291 (gimple_omp_for_final): Likewise.
4292 (gimple_omp_for_final_ptr): Likewise.
4293 (gimple_omp_for_set_final): Likewise.
4294 (gimple_omp_for_incr): Likewise.
4295 (gimple_omp_for_incr_ptr): Likewise.
4296 (gimple_omp_for_set_incr): Likewise.
4297 (gimple_omp_for_pre_body_ptr): Likewise.
4298 (gimple_omp_for_set_pre_body): Likewise.
4299 (gimple_omp_parallel_clauses): Likewise.
4300 (gimple_omp_parallel_clauses_ptr): Likewise.
4301 (gimple_omp_parallel_set_clauses): Likewise.
4302 (gimple_omp_parallel_child_fn): Likewise.
4303 (gimple_omp_parallel_child_fn_ptr): Likewise.
4304 (gimple_omp_parallel_set_child_fn): Likewise.
4305 (gimple_omp_parallel_data_arg): Likewise.
4306 (gimple_omp_parallel_data_arg_ptr): Likewise.
4307 (gimple_omp_parallel_set_data_arg): Likewise.
4308 (gimple_omp_task_clauses): Likewise.
4309 (gimple_omp_task_clauses_ptr): Likewise.
4310 (gimple_omp_task_set_clauses): Likewise.
4311 (gimple_omp_task_child_fn): Likewise.
4312 (gimple_omp_task_child_fn_ptr): Likewise.
4313 (gimple_omp_task_set_child_fn): Likewise.
4314 (gimple_omp_task_data_arg): Likewise.
4315 (gimple_omp_task_data_arg_ptr): Likewise.
4316 (gimple_omp_task_set_data_arg): Likewise.
4317 (gimple_omp_taskreg_clauses): Likewise.
4318 (gimple_omp_taskreg_clauses_ptr): Likewise.
4319 (gimple_omp_taskreg_set_clauses): Likewise.
4320 (gimple_omp_taskreg_child_fn): Likewise.
4321 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4322 (gimple_omp_taskreg_set_child_fn): Likewise.
4323 (gimple_omp_taskreg_data_arg): Likewise.
4324 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4325 (gimple_omp_taskreg_set_data_arg): Likewise.
4326 (gimple_omp_task_copy_fn): Likewise.
4327 (gimple_omp_task_copy_fn_ptr): Likewise.
4328 (gimple_omp_task_set_copy_fn): Likewise.
4329 (gimple_omp_task_arg_size): Likewise.
4330 (gimple_omp_task_arg_size_ptr): Likewise.
4331 (gimple_omp_task_set_arg_size): Likewise.
4332 (gimple_omp_task_arg_align): Likewise.
4333 (gimple_omp_task_arg_align_ptr): Likewise.
4334 (gimple_omp_task_set_arg_align): Likewise.
4335 (gimple_omp_single_clauses): Likewise.
4336 (gimple_omp_single_clauses_ptr): Likewise.
4337 (gimple_omp_single_set_clauses): Likewise.
4338 (gimple_omp_target_clauses): Likewise.
4339 (gimple_omp_target_clauses_ptr): Likewise.
4340 (gimple_omp_target_set_clauses): Likewise.
4341 (gimple_omp_target_child_fn): Likewise.
4342 (gimple_omp_target_child_fn_ptr): Likewise.
4343 (gimple_omp_target_set_child_fn): Likewise.
4344 (gimple_omp_target_data_arg): Likewise.
4345 (gimple_omp_target_data_arg_ptr): Likewise.
4346 (gimple_omp_target_set_data_arg): Likewise.
4347 (gimple_omp_teams_clauses): Likewise.
4348 (gimple_omp_teams_clauses_ptr): Likewise.
4349 (gimple_omp_teams_set_clauses): Likewise.
4350 (gimple_omp_sections_clauses): Likewise.
4351 (gimple_omp_sections_clauses_ptr): Likewise.
4352 (gimple_omp_sections_set_clauses): Likewise.
4353 (gimple_omp_sections_control): Likewise.
4354 (gimple_omp_sections_control_ptr): Likewise.
4355 (gimple_omp_sections_set_control): Likewise.
4356 (gimple_omp_for_set_cond): Likewise.
4357 (gimple_omp_for_cond): Likewise.
4358 (gimple_omp_atomic_store_set_val): Likewise.
4359 (gimple_omp_atomic_store_val): Likewise.
4360 (gimple_omp_atomic_store_val_ptr): Likewise.
4361 (gimple_omp_atomic_load_set_lhs): Likewise.
4362 (gimple_omp_atomic_load_lhs): Likewise.
4363 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4364 (gimple_omp_atomic_load_set_rhs): Likewise.
4365 (gimple_omp_atomic_load_rhs): Likewise.
4366 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4367 (gimple_omp_continue_control_def): Likewise.
4368 (gimple_omp_continue_control_def_ptr): Likewise.
4369 (gimple_omp_continue_set_control_def): Likewise.
4370 (gimple_omp_continue_control_use): Likewise.
4371 (gimple_omp_continue_control_use_ptr): Likewise.
4372 (gimple_omp_continue_set_control_use): Likewise.
4373 (gimple_transaction_body_ptr): Likewise.
4374 (gimple_transaction_label): Likewise.
4375 (gimple_transaction_label_ptr): Likewise.
4376 (gimple_transaction_set_body): Likewise.
4377 (gimple_transaction_set_label): Likewise.
4379 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4380 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4381 * ipa-ref.c (ipa_record_reference): Likewise.
4382 * ipa-reference.c (analyze_function): Likewise.
4383 (ipa_reference_write_optimization_summary): Likewise.
4384 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4385 (address_taken_from_non_vtable_p): Likewise.
4386 (comdat_can_be_unshared_p_1): Likewise.
4387 * lto-cgraph.c (lto_output_ref): Likewise.
4388 (add_references): Likewise.
4389 (compute_ltrans_boundary): Likewise.
4390 (output_symtab): Likewise.
4391 (input_ref): Likewise.
4392 (input_cgraph_1): Likewise.
4393 (output_cgraph_opt_summary): Likewise.
4394 * lto-streamer-out.c (lto_output): Likewise.
4395 (output_symbol_p): Likewise.
4396 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4397 (lsei_start_function_in_partition): Likewise.
4398 (lsei_next_variable_in_partition): Likewise.
4399 (lsei_start_variable_in_partition): Likewise.
4400 * symtab.c (insert_to_assembler_name_hash): Likewise.
4401 (unlink_from_assembler_name_hash): Likewise.
4402 (symtab_unregister_node): Likewise.
4403 (symtab_remove_node): Likewise.
4404 (dump_symtab_node): Likewise.
4405 (verify_symtab_base): Likewise.
4406 (verify_symtab_node): Likewise.
4407 (symtab_make_decl_local): Likewise.
4408 (symtab_alias_ultimate_target): Likewise.
4409 (symtab_resolve_alias): Likewise.
4410 (symtab_get_symbol_partitioning_class): Likewise.
4411 * tree-phinodes.c (allocate_phi_node): Likewise.
4412 (reserve_phi_args_for_new_edge): Likewise.
4413 (remove_phi_args): Likewise.
4414 * varpool.c (varpool_node_for_asm): Likewise.
4415 (varpool_remove_unreferenced_decls): Likewise.
4417 2014-04-23 Jeff Law <law@redhat.com>
4419 PR tree-optimization/60902
4420 * tree-ssa-threadedge.c
4421 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4422 invalidate outputs from statements that do not produce useful
4423 outputs for threading.
4425 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4427 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4428 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4429 machine descriptions for Stack Smashing Protector.
4431 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4433 * aarch64.md (<optab>_rol<mode>3): New pattern.
4434 (<optab>_rolsi3_uxtw): Likewise.
4435 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4437 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4439 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4440 (arm_cortex_a12_tune): Likewise.
4442 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4444 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4446 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4448 * config/arm/arm.md (arm_rev16si2): New pattern.
4449 (arm_rev16si2_alt): Likewise.
4450 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4452 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4454 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4455 (rev16<mode>2_alt): Likewise.
4456 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4457 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4458 (aarch_rev16_shleft_mask_imm_p): Likewise.
4459 (aarch_rev16_p_1): Likewise.
4460 (aarch_rev16_p): Likewise.
4461 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4462 (aarch_rev16_shright_mask_imm_p): Likewise.
4463 (aarch_rev16_shleft_mask_imm_p): Likewise.
4465 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4467 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4468 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4470 (cortex_a53_extra_costs): Likewise.
4471 (cortex_a57_extra_costs): Likewise.
4472 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4473 (cortexa7_extra_costs): Likewise.
4474 (cortexa8_extra_costs): Likewise.
4475 (cortexa12_extra_costs): Likewise.
4476 (cortexa15_extra_costs): Likewise.
4477 (v7m_extra_costs): Likewise.
4478 (arm_new_rtx_costs): Handle BSWAP.
4480 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4482 * config/arm/arm.c (cortexa8_extra_costs): New table.
4483 (arm_cortex_a8_tune): New tuning struct.
4484 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4486 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4488 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4490 2014-04-23 Richard Biener <rguenther@suse.de>
4492 * Makefile.in (OBJS): Remove loop-unswitch.o.
4493 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4494 * passes.def (pass_rtl_unswitch): Likewise.
4495 * loop-init.c (gate_rtl_unswitch): Likewise.
4496 (rtl_unswitch): Likewise.
4497 (pass_data_rtl_unswitch): Likewise.
4498 (pass_rtl_unswitch): Likewise.
4499 (make_pass_rtl_unswitch): Likewise.
4500 * rtl.h (reversed_condition): Likewise.
4501 (compare_and_jump_seq): Likewise.
4502 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4504 * loop-unroll.c (compare_and_jump_seq): Likewise.
4506 2014-04-23 Richard Biener <rguenther@suse.de>
4508 PR tree-optimization/60903
4509 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4510 commented code block.
4511 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4512 loop flags to newly created BBs and edges.
4514 2014-04-23 Nick Clifton <nickc@redhat.com>
4516 * config/msp430/msp430.c (msp430_handle_option): Move function
4518 (msp430_option_override): Simplify mcu and mcpu option handling.
4519 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4520 support for -mhwmult command line option.
4521 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4522 -mhwmult command line option.
4523 (msp430_hwmult_enabled): Delete.
4524 (msp43o_output_labelref): Add support for -mhwmult command line option.
4525 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4526 (umulsidi3): Likewise.
4527 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4528 (mcpu, mlarge, msmall): Likewise.
4529 (mhwmult): New option.
4530 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4532 (msp430_is_f5_mcu): Remove prototype.
4533 (msp430_use_f5_series_hwmult): Add prototype.
4534 * config/msp430/msp430-opts.h: New file.
4535 * common/config/msp430: New directory.
4536 * common/config/msp430/msp430-common.c: New file.
4537 * config.gcc (msp430): Remove target_has_targetm_common.
4538 * doc/invoke.texi: Document -mhwmult command line option.
4540 2014-04-23 Nick Clifton <nickc@redhat.com>
4542 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4543 default-manifest.o if it can be found in the search path.
4544 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4546 2014-04-23 Terry Guo <terry.guo@arm.com>
4548 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4550 2014-04-23 Richard Biener <rguenther@suse.de>
4553 * tree-inline.c (declare_return_variable): Use mark_addressable.
4555 2014-04-23 Richard Biener <rguenther@suse.de>
4558 * loop-init.c (loop_optimizer_init): Make sure to apply
4559 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4561 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4564 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4566 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4567 if flag_sanitize_undefined_trap_on_error.
4568 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4569 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4570 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4571 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4572 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4573 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4574 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4575 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4576 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4577 * ubsan.c (ubsan_instrument_unreachable): Return
4578 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4579 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4580 if flag_sanitize_undefined_trap_on_error and
4581 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4582 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4583 instrument_bool_enum_load): Emit __builtin_trap () if
4584 flag_sanitize_undefined_trap_on_error and
4585 __builtin_handle_*_abort () if !flag_sanitize_recover.
4586 * doc/invoke.texi (-fsanitize-recover,
4587 -fsanitize-undefined-trap-on-error): Document.
4589 2014-04-22 Christian Bruel <christian.bruel@st.com>
4591 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4592 Force immediates to SImode.
4594 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4596 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4598 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4599 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4600 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4601 (nios2_fpu_insn): Add entry for round.
4602 (N2FPU_NO_ERRNO_P): Define.
4603 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4605 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4607 2014-04-22 Richard Henderson <rth@redhat.com>
4609 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4610 (add<GPI>3_compare0): Remove leading * from name.
4611 (add<GPI>3_carryin): Likewise.
4612 (sub<GPI>3_compare0): Likewise.
4613 (sub<GPI>3_carryin): Likewise.
4614 (<su_optab>mulditi3): New expander.
4615 (multi3): New expander.
4616 (madd<GPI>): Remove leading * from name.
4618 2014-04-22 Martin Jambor <mjambor@suse.cz>
4620 * cgraphclones.c (cgraph_function_versioning): Copy
4621 ipa_transforms_to_apply instead of asserting it is empty.
4623 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4626 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4627 on count_exp to get mode.
4629 2014-04-22 Andrew Pinski <apinski@cavium.com>
4631 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4632 Handle TLS for ILP32.
4633 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4634 (tlsie_small_<mode>): this and handle PTR.
4635 (tlsie_small_sidi): New pattern.
4636 (tlsle_small): Change to an expand to handle ILP32.
4637 (tlsle_small_<mode>): New pattern.
4638 (tlsdesc_small): Rename to ...
4639 (tlsdesc_small_<mode>): this and handle PTR.
4641 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4643 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4645 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4647 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4648 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4649 (aarch64_types_signed_poly_qualifiers): Likewise.
4650 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4651 (aarch64_types_poly_signed_qualifiers): Likewise.
4652 (TYPES_REINTERP_SS): Type macro added.
4653 (TYPES_REINTERP_SU): Likewise.
4654 (TYPES_REINTERP_SP): Likewise.
4655 (TYPES_REINTERP_US): Likewise.
4656 (TYPES_REINTERP_PS): Likewise.
4657 (aarch64_fold_builtin): New expression folding added.
4658 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4659 Declarations removed.
4660 (REINTERP_SS): Declarations added.
4661 (REINTERP_US): Likewise.
4662 (REINTERP_PS): Likewise.
4663 (REINTERP_SU): Likewise.
4664 (REINTERP_SP): Likewise.
4665 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4666 (vreinterpretq_p8_f64): Likewise.
4667 (vreinterpret_p16_f64): Likewise.
4668 (vreinterpretq_p16_f64): Likewise.
4669 (vreinterpret_f32_f64): Likewise.
4670 (vreinterpretq_f32_f64): Likewise.
4671 (vreinterpret_f64_f32): Likewise.
4672 (vreinterpret_f64_p8): Likewise.
4673 (vreinterpret_f64_p16): Likewise.
4674 (vreinterpret_f64_s8): Likewise.
4675 (vreinterpret_f64_s16): Likewise.
4676 (vreinterpret_f64_s32): Likewise.
4677 (vreinterpret_f64_s64): Likewise.
4678 (vreinterpret_f64_u8): Likewise.
4679 (vreinterpret_f64_u16): Likewise.
4680 (vreinterpret_f64_u32): Likewise.
4681 (vreinterpret_f64_u64): Likewise.
4682 (vreinterpretq_f64_f32): Likewise.
4683 (vreinterpretq_f64_p8): Likewise.
4684 (vreinterpretq_f64_p16): Likewise.
4685 (vreinterpretq_f64_s8): Likewise.
4686 (vreinterpretq_f64_s16): Likewise.
4687 (vreinterpretq_f64_s32): Likewise.
4688 (vreinterpretq_f64_s64): Likewise.
4689 (vreinterpretq_f64_u8): Likewise.
4690 (vreinterpretq_f64_u16): Likewise.
4691 (vreinterpretq_f64_u32): Likewise.
4692 (vreinterpretq_f64_u64): Likewise.
4693 (vreinterpret_s64_f64): Likewise.
4694 (vreinterpretq_s64_f64): Likewise.
4695 (vreinterpret_u64_f64): Likewise.
4696 (vreinterpretq_u64_f64): Likewise.
4697 (vreinterpret_s8_f64): Likewise.
4698 (vreinterpretq_s8_f64): Likewise.
4699 (vreinterpret_s16_f64): Likewise.
4700 (vreinterpretq_s16_f64): Likewise.
4701 (vreinterpret_s32_f64): Likewise.
4702 (vreinterpretq_s32_f64): Likewise.
4703 (vreinterpret_u8_f64): Likewise.
4704 (vreinterpretq_u8_f64): Likewise.
4705 (vreinterpret_u16_f64): Likewise.
4706 (vreinterpretq_u16_f64): Likewise.
4707 (vreinterpret_u32_f64): Likewise.
4708 (vreinterpretq_u32_f64): Likewise.
4710 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4712 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4713 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4714 (vreinterpret_p8_s8): Likewise.
4715 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4716 (vreinterpret_p8_s16): Likewise.
4717 (vreinterpret_p8_s32): Likewise.
4718 (vreinterpret_p8_s64): Likewise.
4719 (vreinterpret_p8_f32): Likewise.
4720 (vreinterpret_p8_u8): Likewise.
4721 (vreinterpret_p8_u16): Likewise.
4722 (vreinterpret_p8_u32): Likewise.
4723 (vreinterpret_p8_u64): Likewise.
4724 (vreinterpret_p8_p16): Likewise.
4725 (vreinterpretq_p8_s8): Likewise.
4726 (vreinterpretq_p8_s16): Likewise.
4727 (vreinterpretq_p8_s32): Likewise.
4728 (vreinterpretq_p8_s64): Likewise.
4729 (vreinterpretq_p8_f32): Likewise.
4730 (vreinterpretq_p8_u8): Likewise.
4731 (vreinterpretq_p8_u16): Likewise.
4732 (vreinterpretq_p8_u32): Likewise.
4733 (vreinterpretq_p8_u64): Likewise.
4734 (vreinterpretq_p8_p16): Likewise.
4735 (vreinterpret_p16_s8): Likewise.
4736 (vreinterpret_p16_s16): Likewise.
4737 (vreinterpret_p16_s32): Likewise.
4738 (vreinterpret_p16_s64): Likewise.
4739 (vreinterpret_p16_f32): Likewise.
4740 (vreinterpret_p16_u8): Likewise.
4741 (vreinterpret_p16_u16): Likewise.
4742 (vreinterpret_p16_u32): Likewise.
4743 (vreinterpret_p16_u64): Likewise.
4744 (vreinterpret_p16_p8): Likewise.
4745 (vreinterpretq_p16_s8): Likewise.
4746 (vreinterpretq_p16_s16): Likewise.
4747 (vreinterpretq_p16_s32): Likewise.
4748 (vreinterpretq_p16_s64): Likewise.
4749 (vreinterpretq_p16_f32): Likewise.
4750 (vreinterpretq_p16_u8): Likewise.
4751 (vreinterpretq_p16_u16): Likewise.
4752 (vreinterpretq_p16_u32): Likewise.
4753 (vreinterpretq_p16_u64): Likewise.
4754 (vreinterpretq_p16_p8): Likewise.
4755 (vreinterpret_f32_s8): Likewise.
4756 (vreinterpret_f32_s16): Likewise.
4757 (vreinterpret_f32_s32): Likewise.
4758 (vreinterpret_f32_s64): Likewise.
4759 (vreinterpret_f32_u8): Likewise.
4760 (vreinterpret_f32_u16): Likewise.
4761 (vreinterpret_f32_u32): Likewise.
4762 (vreinterpret_f32_u64): Likewise.
4763 (vreinterpret_f32_p8): Likewise.
4764 (vreinterpret_f32_p16): Likewise.
4765 (vreinterpretq_f32_s8): Likewise.
4766 (vreinterpretq_f32_s16): Likewise.
4767 (vreinterpretq_f32_s32): Likewise.
4768 (vreinterpretq_f32_s64): Likewise.
4769 (vreinterpretq_f32_u8): Likewise.
4770 (vreinterpretq_f32_u16): Likewise.
4771 (vreinterpretq_f32_u32): Likewise.
4772 (vreinterpretq_f32_u64): Likewise.
4773 (vreinterpretq_f32_p8): Likewise.
4774 (vreinterpretq_f32_p16): Likewise.
4775 (vreinterpret_s64_s8): Likewise.
4776 (vreinterpret_s64_s16): Likewise.
4777 (vreinterpret_s64_s32): Likewise.
4778 (vreinterpret_s64_f32): Likewise.
4779 (vreinterpret_s64_u8): Likewise.
4780 (vreinterpret_s64_u16): Likewise.
4781 (vreinterpret_s64_u32): Likewise.
4782 (vreinterpret_s64_u64): Likewise.
4783 (vreinterpret_s64_p8): Likewise.
4784 (vreinterpret_s64_p16): Likewise.
4785 (vreinterpretq_s64_s8): Likewise.
4786 (vreinterpretq_s64_s16): Likewise.
4787 (vreinterpretq_s64_s32): Likewise.
4788 (vreinterpretq_s64_f32): Likewise.
4789 (vreinterpretq_s64_u8): Likewise.
4790 (vreinterpretq_s64_u16): Likewise.
4791 (vreinterpretq_s64_u32): Likewise.
4792 (vreinterpretq_s64_u64): Likewise.
4793 (vreinterpretq_s64_p8): Likewise.
4794 (vreinterpretq_s64_p16): Likewise.
4795 (vreinterpret_u64_s8): Likewise.
4796 (vreinterpret_u64_s16): Likewise.
4797 (vreinterpret_u64_s32): Likewise.
4798 (vreinterpret_u64_s64): Likewise.
4799 (vreinterpret_u64_f32): Likewise.
4800 (vreinterpret_u64_u8): Likewise.
4801 (vreinterpret_u64_u16): Likewise.
4802 (vreinterpret_u64_u32): Likewise.
4803 (vreinterpret_u64_p8): Likewise.
4804 (vreinterpret_u64_p16): Likewise.
4805 (vreinterpretq_u64_s8): Likewise.
4806 (vreinterpretq_u64_s16): Likewise.
4807 (vreinterpretq_u64_s32): Likewise.
4808 (vreinterpretq_u64_s64): Likewise.
4809 (vreinterpretq_u64_f32): Likewise.
4810 (vreinterpretq_u64_u8): Likewise.
4811 (vreinterpretq_u64_u16): Likewise.
4812 (vreinterpretq_u64_u32): Likewise.
4813 (vreinterpretq_u64_p8): Likewise.
4814 (vreinterpretq_u64_p16): Likewise.
4815 (vreinterpret_s8_s16): Likewise.
4816 (vreinterpret_s8_s32): Likewise.
4817 (vreinterpret_s8_s64): Likewise.
4818 (vreinterpret_s8_f32): Likewise.
4819 (vreinterpret_s8_u8): Likewise.
4820 (vreinterpret_s8_u16): Likewise.
4821 (vreinterpret_s8_u32): Likewise.
4822 (vreinterpret_s8_u64): Likewise.
4823 (vreinterpret_s8_p8): Likewise.
4824 (vreinterpret_s8_p16): Likewise.
4825 (vreinterpretq_s8_s16): Likewise.
4826 (vreinterpretq_s8_s32): Likewise.
4827 (vreinterpretq_s8_s64): Likewise.
4828 (vreinterpretq_s8_f32): Likewise.
4829 (vreinterpretq_s8_u8): Likewise.
4830 (vreinterpretq_s8_u16): Likewise.
4831 (vreinterpretq_s8_u32): Likewise.
4832 (vreinterpretq_s8_u64): Likewise.
4833 (vreinterpretq_s8_p8): Likewise.
4834 (vreinterpretq_s8_p16): Likewise.
4835 (vreinterpret_s16_s8): Likewise.
4836 (vreinterpret_s16_s32): Likewise.
4837 (vreinterpret_s16_s64): Likewise.
4838 (vreinterpret_s16_f32): Likewise.
4839 (vreinterpret_s16_u8): Likewise.
4840 (vreinterpret_s16_u16): Likewise.
4841 (vreinterpret_s16_u32): Likewise.
4842 (vreinterpret_s16_u64): Likewise.
4843 (vreinterpret_s16_p8): Likewise.
4844 (vreinterpret_s16_p16): Likewise.
4845 (vreinterpretq_s16_s8): Likewise.
4846 (vreinterpretq_s16_s32): Likewise.
4847 (vreinterpretq_s16_s64): Likewise.
4848 (vreinterpretq_s16_f32): Likewise.
4849 (vreinterpretq_s16_u8): Likewise.
4850 (vreinterpretq_s16_u16): Likewise.
4851 (vreinterpretq_s16_u32): Likewise.
4852 (vreinterpretq_s16_u64): Likewise.
4853 (vreinterpretq_s16_p8): Likewise.
4854 (vreinterpretq_s16_p16): Likewise.
4855 (vreinterpret_s32_s8): Likewise.
4856 (vreinterpret_s32_s16): Likewise.
4857 (vreinterpret_s32_s64): Likewise.
4858 (vreinterpret_s32_f32): Likewise.
4859 (vreinterpret_s32_u8): Likewise.
4860 (vreinterpret_s32_u16): Likewise.
4861 (vreinterpret_s32_u32): Likewise.
4862 (vreinterpret_s32_u64): Likewise.
4863 (vreinterpret_s32_p8): Likewise.
4864 (vreinterpret_s32_p16): Likewise.
4865 (vreinterpretq_s32_s8): Likewise.
4866 (vreinterpretq_s32_s16): Likewise.
4867 (vreinterpretq_s32_s64): Likewise.
4868 (vreinterpretq_s32_f32): Likewise.
4869 (vreinterpretq_s32_u8): Likewise.
4870 (vreinterpretq_s32_u16): Likewise.
4871 (vreinterpretq_s32_u32): Likewise.
4872 (vreinterpretq_s32_u64): Likewise.
4873 (vreinterpretq_s32_p8): Likewise.
4874 (vreinterpretq_s32_p16): Likewise.
4875 (vreinterpret_u8_s8): Likewise.
4876 (vreinterpret_u8_s16): Likewise.
4877 (vreinterpret_u8_s32): Likewise.
4878 (vreinterpret_u8_s64): Likewise.
4879 (vreinterpret_u8_f32): Likewise.
4880 (vreinterpret_u8_u16): Likewise.
4881 (vreinterpret_u8_u32): Likewise.
4882 (vreinterpret_u8_u64): Likewise.
4883 (vreinterpret_u8_p8): Likewise.
4884 (vreinterpret_u8_p16): Likewise.
4885 (vreinterpretq_u8_s8): Likewise.
4886 (vreinterpretq_u8_s16): Likewise.
4887 (vreinterpretq_u8_s32): Likewise.
4888 (vreinterpretq_u8_s64): Likewise.
4889 (vreinterpretq_u8_f32): Likewise.
4890 (vreinterpretq_u8_u16): Likewise.
4891 (vreinterpretq_u8_u32): Likewise.
4892 (vreinterpretq_u8_u64): Likewise.
4893 (vreinterpretq_u8_p8): Likewise.
4894 (vreinterpretq_u8_p16): Likewise.
4895 (vreinterpret_u16_s8): Likewise.
4896 (vreinterpret_u16_s16): Likewise.
4897 (vreinterpret_u16_s32): Likewise.
4898 (vreinterpret_u16_s64): Likewise.
4899 (vreinterpret_u16_f32): Likewise.
4900 (vreinterpret_u16_u8): Likewise.
4901 (vreinterpret_u16_u32): Likewise.
4902 (vreinterpret_u16_u64): Likewise.
4903 (vreinterpret_u16_p8): Likewise.
4904 (vreinterpret_u16_p16): Likewise.
4905 (vreinterpretq_u16_s8): Likewise.
4906 (vreinterpretq_u16_s16): Likewise.
4907 (vreinterpretq_u16_s32): Likewise.
4908 (vreinterpretq_u16_s64): Likewise.
4909 (vreinterpretq_u16_f32): Likewise.
4910 (vreinterpretq_u16_u8): Likewise.
4911 (vreinterpretq_u16_u32): Likewise.
4912 (vreinterpretq_u16_u64): Likewise.
4913 (vreinterpretq_u16_p8): Likewise.
4914 (vreinterpretq_u16_p16): Likewise.
4915 (vreinterpret_u32_s8): Likewise.
4916 (vreinterpret_u32_s16): Likewise.
4917 (vreinterpret_u32_s32): Likewise.
4918 (vreinterpret_u32_s64): Likewise.
4919 (vreinterpret_u32_f32): Likewise.
4920 (vreinterpret_u32_u8): Likewise.
4921 (vreinterpret_u32_u16): Likewise.
4922 (vreinterpret_u32_u64): Likewise.
4923 (vreinterpret_u32_p8): Likewise.
4924 (vreinterpret_u32_p16): Likewise.
4925 (vreinterpretq_u32_s8): Likewise.
4926 (vreinterpretq_u32_s16): Likewise.
4927 (vreinterpretq_u32_s32): Likewise.
4928 (vreinterpretq_u32_s64): Likewise.
4929 (vreinterpretq_u32_f32): Likewise.
4930 (vreinterpretq_u32_u8): Likewise.
4931 (vreinterpretq_u32_u16): Likewise.
4932 (vreinterpretq_u32_u64): Likewise.
4933 (vreinterpretq_u32_p8): Likewise.
4934 (vreinterpretq_u32_p16): Likewise.
4936 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4938 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4940 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4942 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4943 (vqnegd_s64): Likewise.
4944 (vqabs_s64): Likewise.
4945 (vqabsd_s64): Likewise.
4947 2014-04-22 Richard Henderson <rth@redhat.com>
4949 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4950 computation to the top of the loop.
4952 2014-04-22 Renlin <renlin.li@arm.com>
4953 Jiong Wang <jiong.wang@arm.com>
4955 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4956 * config/aarch64/aarch64.c (aarch64_layout_frame)
4957 (aarch64_initial_elimination_offset): Likewise.
4959 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4961 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4964 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4966 * machmode.h (bitwise_mode_for_mode): Declare.
4967 * stor-layout.h (bitwise_type_for_mode): Likewise.
4968 * stor-layout.c (bitwise_mode_for_mode): New function.
4969 (bitwise_type_for_mode): Likewise.
4970 * builtins.c (fold_builtin_memory_op): Use it instead of
4971 int_mode_for_mode and build_nonstandard_integer_type.
4973 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4975 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4976 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4977 (*-*-solaris2*): Simplify.
4978 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4979 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4980 *-*-solaris2.9* handling.
4982 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4984 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4985 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4987 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4988 * configure: Regenerate.
4990 * config/i386/sol2-9.h: Remove.
4992 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4993 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4994 Remove Solaris 9 references.
4996 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4998 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4999 (floatuns<GPI:mode><GPF:mode>2): Remove.
5000 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5001 and floatuns conversions.
5002 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5003 and floatuns conversions.
5004 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5005 (w1,w2): New mode attributes for inequal width conversions.
5007 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5009 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5010 the output asm format.
5012 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5014 * config/aarch64/aarch64-simd.md
5015 (aarch64_cm<optab>di): Always split.
5016 (*aarch64_cm<optab>di): New.
5017 (aarch64_cmtstdi): Always split.
5018 (*aarch64_cmtstdi): New.
5020 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5022 PR tree-optimization/60823
5023 * omp-low.c (ipa_simd_modify_function_body): Go through
5024 all SSA_NAMEs and for those refering to vector arguments
5025 which are going to be replaced adjust SSA_NAME_VAR and,
5026 if it is a default definition, change it into a non-default
5027 definition assigned at the beginning of function from new_decl.
5028 (ipa_simd_modify_stmt_ops): Rewritten.
5029 * tree-dfa.c (set_ssa_default_def): When removing default def,
5030 check for NULL loc instead of NULL *loc.
5032 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5034 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5035 restrictions on core registers for DImode values in Thumb2.
5037 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5039 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5040 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5042 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5044 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5045 (*iordi_notzesidi_di): Likewise.
5046 (*iordi_notsesidi_di): Likewise.
5048 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5050 * config/arm/arm-protos.h (tune_params): New struct members.
5051 * config/arm/arm.c: Initialise tune_params per processor.
5052 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5053 for speed, based on new tune_params.
5055 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5057 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5058 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5059 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5060 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5061 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5062 (vrnda_f64): Likewise.
5063 (vrndi_f64): Likewise.
5064 (vrndm_f64): Likewise.
5065 (vrndn_f64): Likewise.
5066 (vrndp_f64): Likewise.
5067 (vrndx_f64): Likewise.
5069 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5071 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5072 GET_MODE_SIZE argument is enum machine_mode.
5074 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5077 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5078 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5080 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5083 * asan.c (asan_emit_stack_protection): Force the base to align to
5084 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5085 appropriate bits if STRICT_ALIGNMENT.
5086 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5088 (expand_used_vars): Leave a space in the stack frame for alignment
5089 if STRICT_ALIGNMENT.
5091 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5093 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5095 (gimple_store_p): Likewise.
5096 (gimple_assign_load_p): Likewise.
5097 (gimple_assign_cast_p): Likewise.
5098 (gimple_clobber_p): Likewise.
5100 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5101 rather than a gimple.
5102 (gimple_assign_cast_p): Likewise.
5104 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5107 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5108 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5110 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5111 more debug information for E500 if -mdebug=reg.
5113 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5116 * config/i386/i386.c (ix86_expand_builtin)
5117 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5118 register for target RTX.
5119 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5121 2014-04-18 Cong Hou <congh@google.com>
5123 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5124 the widen-mult pattern by handling two operands with different sizes,
5125 and operands whose size is smaller than half of the result type.
5127 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5129 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5130 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5131 (do_estimate_edge_time): Compute it.
5132 * ipa-inline.c (want_inline_small_function_p): Bypass
5133 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5135 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5137 * ipa-inline.c (spec_rem): New static variable.
5138 (dump_overall_stats): New function.
5139 (dump_inline_stats): New function.
5141 2014-04-18 Richard Henderson <rth@redhat.com>
5143 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5144 to GET_MODE_SIZE, not a reg_class_t.
5146 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5148 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5149 (vsx_xxmrglw_<mode>): Likewise.
5151 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5154 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5155 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5156 (rs6000_init_hard_regno_mode_ok): Likewise.
5158 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5160 * ipa-inline.c (inline_small_functions): Account only non-cold
5162 * doc/invoke.texi (inline-unit-growth): Update documentation.
5164 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5166 * config/rs6000/rs6000.md (addti3, subti3): New.
5168 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5171 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5172 comment. Check optimize_insn_for_size_p instead of
5173 optimize_insn_for_speed_p.
5175 2014-04-17 Martin Jambor <mjambor@suse.cz>
5177 * gimple-iterator.c (gsi_start_edge): New function.
5178 * gimple-iterator.h (gsi_start_edge): Declare.
5179 * tree-sra.c (single_non_eh_succ): New function.
5180 (disqualify_ops_if_throwing_stmt): Renamed to
5181 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5182 having one non-EH successor BB.
5183 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5184 generate loads into replacements.
5185 (sra_modify_assign): Likewise and and also use the simple path for
5187 (sra_modify_function_body): Commit statements on edges.
5189 2014-04-17 Richard Biener <rguenther@suse.de>
5192 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5193 comparison results and add clarifying comment.
5195 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5197 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5198 (blank_mode): Initialize it.
5199 (emit_mode_size_inline, emit_mode_nunits_inline,
5200 emit_mode_inner_inline): New functions.
5201 (emit_insn_modes_h): Call them and surround their output with
5202 #if GCC_VERSION >= 4001 ... #endif.
5203 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5204 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5205 mode_* arrays if the argument is __builtin_constant_p.
5206 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5207 is enum machine_mode.
5209 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5211 * passes.c (opt_pass::execute): Adjust.
5212 (pass_manager::execute_pass_mode_switching): Likewise.
5213 (early_local_passes::execute): Likewise.
5214 (execute_one_pass): Pass cfun to the pass's execute method.
5215 * tree-pass.h (opt_pass::execute): Add function * argument.
5216 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5217 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5218 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5219 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5220 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5221 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5222 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5223 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5224 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5225 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5226 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5227 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5228 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5229 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5230 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5231 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5232 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5233 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5234 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5235 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5236 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5237 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5238 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5239 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5240 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5241 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5242 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5243 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5244 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5247 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5249 * passes.c (opt_pass::gate): Take function * argument.
5250 (gate_all_early_local_passes): Merge into
5251 (early_local_passes::gate): this.
5252 (gate_all_early_optimizations): Merge into
5253 (all_early_optimizations::gate): this.
5254 (gate_all_optimizations): Mege into
5255 (all_optimizations::gate): this.
5256 (gate_all_optimizations_g): Merge into
5257 (all_optimizations_g::gate): this.
5258 (gate_rest_of_compilation): Mege into
5259 (rest_of_compilation::gate): this.
5260 (gate_postreload): Merge into
5261 (postreload::gate): this.
5262 (dump_one_pass): Pass cfun to the pass's gate method.
5263 (execute_ipa_summary_passes): Likewise.
5264 (execute_one_pass): Likewise.
5265 (ipa_write_summaries_2): Likewise.
5266 (ipa_write_optimization_summaries_1): Likewise.
5267 (ipa_read_summaries_1): Likewise.
5268 (ipa_read_optimization_summaries_1): Likewise.
5269 (execute_ipa_stmt_fixups): Likewise.
5270 * tree-pass.h (opt_pass::gate): Add function * argument.
5271 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5272 combine-stack-adj.c, combine.c, compare-elim.c,
5273 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5274 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5275 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5276 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5277 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5278 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5279 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5280 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5281 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5282 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5283 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5284 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5285 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5286 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5287 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5288 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5289 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5290 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5291 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5292 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5293 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5294 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5295 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5296 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5297 var-tracking.c, vtable-verify.c, web.c: Adjust.
5299 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5301 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5302 * configure: Regenerate.
5304 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5306 * passes.c (dump_one_pass): don't check pass->has_gate.
5307 (execute_ipa_summary_passes): Likewise.
5308 (execute_one_pass): Likewise.
5309 (ipa_write_summaries_2): Likewise.
5310 (ipa_write_optimization_summaries_1): Likewise.
5311 (ipa_read_optimization_summaries_1): Likewise.
5312 (execute_ipa_stmt_fixups): Likewise.
5313 * tree-pass.h (pass_data::has_gate): Remove.
5314 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5315 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5316 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5317 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5318 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5319 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5320 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5321 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5322 gimple-low.c, gimple-ssa-isolate-paths.c,
5323 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5324 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5325 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5326 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5327 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5328 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5329 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5330 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5331 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5332 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5333 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5334 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5335 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5336 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5337 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5338 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5339 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5340 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5341 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5342 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5343 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5344 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5347 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5349 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5351 * passes.c (pass_manager::register_dump_files_1): Merge into
5352 (pass_manager::register_dump_files): this, and remove its handling of
5353 properties since the pass always has the properties anyway.
5354 (pass_manager::pass_manager): Adjust.
5356 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5358 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5359 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5360 dealing with properties.
5361 (pass_manager::register_dump_files): Adjust.
5363 2014-03-20 Mark Wielaard <mjw@redhat.com>
5365 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5366 then represent the bound as normal constant value.
5368 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5371 Forward port from 4.8 branch
5372 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5374 * config/i386/bmiintrin.h (_blsi_u32): New.
5378 (_blsmsk_u32): Ditto.
5379 (_blsmsk_u64): Ditto.
5380 (_tzcnt_u32): Ditto.
5381 (_tzcnt_u64): Ditto.
5383 2014-04-17 Kito Cheng <kito@0xlab.org>
5385 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5387 2014-04-17 Richard Biener <rguenther@suse.de>
5390 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5391 boolean results for comparisons.
5393 2014-04-17 Richard Biener <rguenther@suse.de>
5395 PR tree-optimization/60836
5396 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5397 initial PHI args to be gimple values.
5399 2014-04-17 Richard Biener <rguenther@suse.de>
5401 PR tree-optimization/60841
5402 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5403 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5404 of stmts to SLP build.
5405 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5406 (vect_analyze_slp): Likewise.
5407 (vect_analyze_slp_instance): Likewise.
5408 (vect_build_slp_tree): Limit overall SLP tree growth.
5409 * tree-vectorizer.h (vect_analyze_data_refs,
5410 vect_analyze_slp): Adjust prototypes.
5412 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5414 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5417 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5419 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5420 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5421 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5422 for TARGET_SLOW_PSHUFB
5424 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5426 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5427 * config/i386/i386.c (intel_cost): Ditto.
5429 2014-04-17 Joey Ye <joey.ye@arm.com>
5431 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5433 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5435 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5438 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5440 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5441 (type_all_derivations_known_p): New predicate.
5442 (type_all_ctors_visible_p): New predicate.
5443 (type_possibly_instantiated_p): New predicate.
5444 (get_odr_type): Compute all_derivations_known.
5445 (dump_odr_type): Dump the flag.
5446 (maybe_record_type): Cleanup.
5447 (record_target_from_binfo): Add bases_to_consider array;
5448 record bases for types w/o instances and skip CXX destructor.
5449 (possible_polymorphic_call_targets_1): Add bases_to_consider
5450 and consider_construction parameters; check if type may have instance.
5451 (get_polymorphic_call_info): Set maybe_in_construction to true
5452 when we know nothing.
5453 (record_targets_from_bases): Skip CXX destructors; they are
5454 never called for types in construction.
5455 (possible_polymorphic_call_targets): Do not record target when
5456 type may not have instance.
5458 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5461 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5462 external aliases alive, too.
5464 2014-04-16 Andrew Pinski <apinski@cavium.com>
5466 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5469 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5471 * final.c (compute_alignments): Do not apply loop alignment to a block
5472 falling through to the exit.
5474 2014-04-16 Catherine Moore <clm@codesourcery.com>
5476 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5477 Adjust constraints for microMIPS store patterns.
5479 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5481 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5483 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5485 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5486 (append_use): Run at -O0.
5487 (append_vdef): Likewise.
5488 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5489 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5491 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5493 PR tree-optimization/60844
5494 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5495 (propagate_op_to_single_use, remove_visited_stmt_chain,
5496 linearize_expr, repropagate_negates, reassociate_bb): Use it
5497 instead of gsi_remove.
5499 2014-04-16 Martin Jambor <mjambor@suse.cz>
5501 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5502 ipa_transforms_to_apply.
5503 (cgraph_function_versioning): Assert that old_node has empty
5504 ipa_transforms_to_apply.
5505 * trans-mem.c (ipa_tm_create_version): Likewise.
5506 * tree-inline.c (tree_function_versioning): Do not duplicate
5507 ipa_transforms_to_apply.
5509 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5512 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5514 Pass necessary as flags on 64-bit Solaris/x86.
5515 Use lowercase relocs for x86_64-*-*.
5516 * configure: Regenerate.
5518 2014-04-15 Jan Hubicka <jh@suse.cz>
5520 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5521 (maybe_record_node, likely_target_p): Use it.
5523 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5526 Revert following patch
5528 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5531 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5532 software floating point or no floating point registers, do not
5533 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5534 in GPRs that occurs after we tested for GPRs that would never be
5537 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5538 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5539 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5540 specifically allow DDmode, since that does not use the SPE SIMD
5543 2014-03-21 Mark Wielaard <mjw@redhat.com>
5545 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5546 as unsigned or int depending on type and value used.
5548 2014-04-15 Richard Biener <rguenther@suse.de>
5550 PR rtl-optimization/56965
5551 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5552 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5554 * alias.c (true_dependence_1): Do not call
5555 nonoverlapping_component_refs_p.
5556 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5557 nonoverlapping_component_refs_p.
5558 (indirect_refs_may_alias_p): Likewise.
5560 2014-04-15 Teresa Johnson <tejohnson@google.com>
5562 * cfg.c (dump_bb_info): Fix flags check.
5563 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5565 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5567 PR rtl-optimization/60663
5568 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5571 2014-04-15 Richard Biener <rguenther@suse.de>
5573 * lto-streamer.h (LTO_major_version): Bump to 4.
5575 2014-04-15 Richard Biener <rguenther@suse.de>
5577 * common.opt (lto_partition_model): New enum.
5578 (flto-partition=): Merge separate options with a single with argument,
5579 add -flto-partition=one support.
5580 * flag-types.h (enum lto_partition_model): Declare.
5581 * opts.c (finish_options): Remove duplicate -flto-partition=
5583 * lto-wrapper.c (run_gcc): Adjust.
5585 2014-04-15 Richard Biener <rguenther@suse.de>
5587 * alias.c (ncr_compar): New function.
5588 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5590 2014-04-15 Richard Biener <rguenther@suse.de>
5592 * alias.c (record_component_aliases): Do not walk BINFOs.
5594 2014-04-15 Richard Biener <rguenther@suse.de>
5596 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5597 Add struct function argument and adjust.
5598 (find_func_aliases_for_call): Likewise.
5599 (find_func_aliases): Likewise.
5600 (find_func_clobbers): Likewise.
5601 (intra_create_variable_infos): Likewise.
5602 (compute_points_to_sets): Likewise.
5603 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5605 2014-04-15 Richard Biener <rguenther@suse.de>
5607 * tree.c (iterative_hash_expr): Use enum tree_code_class
5608 to store TREE_CODE_CLASS.
5609 (tree_block): Likewise.
5610 (tree_set_block): Likewise.
5611 * tree.h (fold_build_pointer_plus_loc): Use
5612 convert_to_ptrofftype_loc.
5614 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5617 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5620 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5622 * cfgloop.h (struct loop): Move force_vectorize down.
5623 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5624 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5625 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5626 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5627 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5628 * tree-core.h (enum annot_expr_kind): Add new kind values.
5629 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5630 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5632 * tree.def (ANNOTATE_EXPR): Tweak comment.
5634 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5636 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5639 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5641 * tree.h (TYPE_IDENTIFIER): Declare.
5642 * tree.c (subrange_type_for_debug_p): Use it.
5643 * godump.c (go_format_type): Likewise.
5644 * dwarf2out.c (is_cxx_auto, modified_type_die,
5645 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5646 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5648 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5651 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5653 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5655 * config/i386/i386.c (examine_argument): Return bool. Return true if
5656 parameter should be passed in memory.
5657 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5658 (construct_container): Update calls to examine_argument.
5659 (function_arg_advance_64): Ditto.
5660 (return_in_memory_32): Merge with ix86_return_in_memory.
5661 (return_in_memory_64): Ditto.
5662 (return_in_memory_ms_64): Ditto.
5664 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5666 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5667 * coverage.c (coverage_compute_profile_id): Handle externally visible
5670 2014-04-14 Martin Jambor <mjambor@suse.cz>
5672 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5673 DECL_DISREGARD_INLINE_LIMITS functions.
5675 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5678 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5680 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5683 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5684 optimize_insn_for_speed_p instead of
5685 optimize_function_for_speed_p.
5687 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5689 * doc/invoke.texi (free): Document AArch64.
5691 2014-04-14 Richard Biener <rguenther@suse.de>
5693 PR tree-optimization/60042
5694 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5695 (insert_into_preds_of_block): Do not prevent PHI insertion
5696 for REFERENCE exprs here ...
5697 (eliminate_dom_walker::before_dom_children): ... but prevent
5698 their use here under similar conditions when applied to the
5699 IL after PRE optimizations.
5701 2014-04-14 Richard Biener <rguenther@suse.de>
5703 * passes.def: Move early points-to after early SRA.
5705 2014-04-14 Richard Biener <rguenther@suse.de>
5707 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5708 check for which sign-changes we allow when forwarding
5709 a converted value into a switch.
5711 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5713 * stor-layout.c (place_field): Finalize non-constant offset for the
5716 2014-04-14 Richard Biener <rguenther@suse.de>
5718 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5720 (expand_switch_using_bit_tests_p): Likewise.
5721 (process_switch): Compute and pass on speed_p based on the
5723 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5724 optimize_bb_for_speed_p.
5726 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5728 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5729 * function.h (struct function): Rename has_force_vect_loops into
5730 has_force_vectorize_loops.
5731 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5732 (input_struct_function_base): Likewise.
5733 * lto-streamer-out.c (output_cfg): Likewise.
5734 (output_struct_function_base): Likewise.
5735 * omp-low.c (expand_omp_simd): Likewise.
5736 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5737 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5738 (version_loop_for_if_conversion): Likewise.
5739 (tree_if_conversion): Likewise.
5740 (main_tree_if_conversion): Likewise.
5741 (gate_tree_if_conversion): Likewise.
5742 * tree-inline.c (copy_loops): Likewise.
5743 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5744 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5745 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5746 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5747 * tree-vectorizer.c (vectorize_loops): Likewise.
5748 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5750 2014-04-14 Richard Biener <rguenther@suse.de>
5753 * lto-streamer-out.c (wrap_refs): New function.
5754 (lto_output): Wrap symbol references in global initializes in
5755 type-preserving MEM_REFs.
5757 2014-04-14 Christian Bruel <christian.bruel@st.com>
5759 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5761 2014-04-14 Christian Bruel <christian.bruel@st.com>
5763 * config/sh/sh.md (setmemqi): New expand pattern.
5764 * config/sh/sh.h (CLEAR_RATIO): Define.
5765 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5766 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5768 2014-04-14 Richard Biener <rguenther@suse.de>
5771 * fold-const.c (negate_expr_p): Don't negate directional rounding
5773 (fold_negate_expr): Likewise.
5775 2014-04-14 Richard Biener <rguenther@suse.de>
5777 PR tree-optimization/59817
5778 PR tree-optimization/60453
5779 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5780 recursion to catch all CHRECs in the scalar evolution and restrict
5781 the predicate for the remains appropriately.
5783 2014-04-12 Catherine Moore <clm@codesourcery.com>
5785 * config/mips/constraints.md: Add new register constraint "kb".
5786 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5787 (*movhi_internal): Likewise.
5788 (*movqi_internal): Likewise.
5789 * config/mips/mips.h (M16_STORE_REGS): New register class.
5790 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5791 (REG_CLASS_CONTENTS): Likewise.
5792 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5794 2014-04-11 Tobias Burnus <burnus@net-b.de>
5797 * doc/invoke.texi (-Wformat-signedness): Document it.
5798 (Wformat=2): Mention that this enables -Wformat-signedness.
5800 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5802 * common/config/epiphany/epiphany-common.c
5803 (epiphany_option_optimization_table): Enable section anchors by
5804 default at -O1 or higher.
5805 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5806 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5807 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5808 carries no extra cost.
5809 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5810 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5811 * config/epiphany/predicates.md (memclob_operand): New predicate.
5812 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5813 Use memclob_operand predicate and X constraint for operand 3.
5815 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5817 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5818 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5821 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5823 PR rtl-optimization/60651
5824 * mode-switching.c (optimize_mode_switching): Make sure to emit
5825 sets of a lower numbered entity before sets of a higher numbered
5826 entity to a mode of the same or lower priority.
5827 When creating a seginfo for a basic block that starts with a code
5828 label, move the insertion point past the code label.
5829 (new_seginfo): Document and enforce requirement that
5830 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5831 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5832 * doc/tm.texi: Regenerate.
5834 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5837 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5839 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5841 * BASE-VER: Set to 4.10.0.
5843 2014-04-11 Tobias Burnus <burnus@net-b.de>
5846 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5847 * doc/gcc.texi (Service): Update description in the @menu
5848 * doc/invoke.texi (Option Summary): Remove misplaced and
5851 2014-04-11 Steve Ellcey <sellcey@mips.com>
5852 Jakub Jelinek <jakub@redhat.com>
5855 * expr.c (convert_move): Use emit_store_flag_force instead of
5856 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5859 2014-04-11 Richard Biener <rguenther@suse.de>
5862 * varasm.c (assemble_alias): Avoid endless error reporting
5863 recursion by setting TREE_ASM_WRITTEN.
5865 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5867 * config/s390/s390.md: Add a splitter for NOT rtx.
5869 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5871 PR rtl-optimization/60663
5872 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5874 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5875 Jakub Jelinek <jakub@redhat.com>
5878 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5879 flag from decl_node to node.
5881 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5884 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5885 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5886 ameliorating the cases where it can be.
5888 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5891 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5893 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5894 (loadsync_<mode>): Change mode.
5895 (load_quadpti, store_quadpti): New.
5896 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5897 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5898 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5900 2014-04-09 Cong Hou <congh@google.com>
5903 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5906 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5908 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5909 instead of vnor to exploit possible fusion opportunity in the
5911 (altivec_expand_vec_perm_const_le): Likewise.
5913 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5915 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5916 (loadsync_<mode>): Change mode.
5917 (load_quadpti, store_quadpti): New.
5918 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5919 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5921 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5924 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5925 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5926 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5928 2014-04-08 Richard Biener <rguenther@suse.de>
5931 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5932 a 64bit widest int print double-int similar to on HWI64 hosts.
5934 2014-04-08 Richard Biener <rguenther@suse.de>
5936 PR tree-optimization/60785
5937 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5938 default defs properly.
5940 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5942 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5943 (Weffc++): Likewise.
5945 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5947 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5948 set completep to false rather than true.
5950 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5953 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5954 ARM_TARGET2_DWARF_FORMAT.
5956 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5959 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5960 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5963 2014-04-07 Richard Biener <rguenther@suse.de>
5965 PR tree-optimization/60766
5966 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5967 (may_eliminate_iv): Convert cand_value_at result to desired type.
5969 2014-04-07 Jason Merrill <jason@redhat.com>
5972 * common.opt (-fno-gnu-unique): Add.
5973 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5975 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5977 * haifa-sched.c: Fix outdated function reference and minor
5978 grammar errors in introductory comment.
5980 2014-04-07 Richard Biener <rguenther@suse.de>
5983 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5985 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5987 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5990 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5991 size accounting for thunks.
5992 (pa_asm_output_mi_thunk): Use final_start_function() and
5993 final_end_function() to output function start and end directives.
5995 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5997 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5998 device specific ISA/ feature information. Remove short_sp and
5999 errata_skip ds. Add avr_device_specific_features enum to have device
6001 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6002 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6003 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6004 updated device specific info.
6005 * config/avr/avr-mcus.def: Merge device specific details to
6006 dev_attribute field.
6007 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6009 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6010 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6011 assembler if RMW isa supported by current device.
6012 * config/avr/genmultilib.awk: Update as device info structure changed.
6013 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6015 2014-04-04 Cong Hou <congh@google.com>
6017 PR tree-optimization/60656
6018 * tree-vect-stmts.c (supportable_widening_operation):
6019 Fix a bug that elements in a vector with vect_used_by_reduction
6020 property are incorrectly reordered when the operation on it is not
6021 consistant with the one in reduction operation.
6023 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6025 PR rtl-optimization/60155
6026 * gcse.c (record_set_data): New function.
6027 (single_set_gcse): New function.
6028 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6029 (hoist_code): Likewise.
6030 (get_pressure_class_and_nregs): Likewise.
6032 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6034 * explow.c (probe_stack_range): Emit a final optimization blockage.
6036 2014-04-04 Anthony Green <green@moxielogic.com>
6038 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6041 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6044 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6045 flags are set only during streaming.
6046 * ipa.c (process_references, walk_polymorphic_call_targets,
6047 symtab_remove_unreachable_nodes): Drop bodies of always inline
6048 after early inlining.
6049 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6051 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6052 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6055 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6058 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6061 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6063 (cortex_a53_fdivd): Likewise.
6065 2014-04-04 Martin Jambor <mjambor@suse.cz>
6068 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6070 * cgraph.c (clone_of_p): Also return true if thunks match.
6071 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6072 cgraph_function_or_thunk_node and an obsolete comment.
6073 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6075 (build_function_decl_skip_args): Likewise.
6076 (set_new_clone_decl_and_node_flags): New function.
6077 (duplicate_thunk_for_node): Likewise.
6078 (redirect_edge_duplicating_thunks): Likewise.
6079 (cgraph_clone_node): New parameter args_to_skip, pass it to
6080 redirect_edge_duplicating_thunks which is called instead of
6081 cgraph_redirect_edge_callee.
6082 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6083 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6085 2014-04-04 Jeff Law <law@redhat.com>
6088 * config/arm/predicates.md (const_int_I_operand): New predicate.
6089 (const_int_M_operand): Similarly.
6090 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6092 (insv_t2, extv_reg, extzv_t2): Likewise.
6093 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6094 (pop_multiple_with_writeback_and_return): Likewise.
6095 (vfp_pop_multiple_with_writeback): Likewise
6097 2014-04-04 Richard Biener <rguenther@suse.de>
6100 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6101 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6103 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6104 * gimplify.c (gimple_add_tmp_var_fn): New function.
6105 * gimple-expr.h (create_tmp_reg_fn): Declare.
6106 * gimple-expr.c (create_tmp_reg_fn): New function.
6107 * gimple-low.c (record_vars_into): Don't change cfun.
6108 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6109 code generation without cfun.
6111 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6114 * Makefile.in (install-driver): Fix shell scripting.
6116 2014-04-03 Cong Hou <congh@google.com>
6118 PR tree-optimization/60505
6119 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6120 threshold of number of iterations below which no vectorization
6122 * tree-vect-loop.c (new_loop_vec_info):
6123 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6124 * tree-vect-loop.c (vect_analyze_loop_operations):
6125 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6126 * tree-vect-loop.c (vect_transform_loop):
6127 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6128 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6129 of iterations of the loop and see if we should build the epilogue.
6131 2014-04-03 Richard Biener <rguenther@suse.de>
6133 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6134 (streamer_tree_cache_create): Adjust.
6135 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6136 to allow optional nodes array.
6137 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6138 (streamer_tree_cache_append): Likewise.
6139 (streamer_tree_cache_create): Create nodes array optionally
6140 as specified by parameter.
6141 * lto-streamer-out.c (create_output_block): Avoid maintaining
6142 the node array in the writer cache.
6143 (DFS_write_tree): Remove assertion.
6144 (produce_asm_for_decls): Free the out decl state hash table early.
6145 * lto-streamer-in.c (lto_data_in_create): Adjust for
6146 streamer_tree_cache_create prototype change.
6148 2014-04-03 Richard Biener <rguenther@suse.de>
6150 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6151 set TREE_CHAIN to NULL_TREE.
6153 2014-04-03 Richard Biener <rguenther@suse.de>
6155 PR tree-optimization/60740
6156 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6157 over all GIMPLE_COND operands.
6159 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6161 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6162 (Weffc++): Remove Scott's numbering, merge lists and reference
6165 2014-04-03 Nick Clifton <nickc@redhat.com>
6167 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6170 2014-04-03 Martin Jambor <mjambor@suse.cz>
6172 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6173 mention gcc_unreachable before failing.
6174 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6177 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6180 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6181 inconsistent code and instead mark the context inconsistent.
6182 (possible_polymorphic_call_targets): For inconsistent contexts
6183 return empty complete list.
6185 2014-04-02 Anthony Green <green@moxielogic.com>
6187 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6188 (extendqisi2, extendhisi2): Define.
6189 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6190 (WCHAR_TYPE): Change to unsigned int.
6192 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6194 PR tree-optimization/60733
6195 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6196 insertion point for PHI candidates to be the end of the feeding
6197 block for the PHI argument.
6199 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6201 PR rtl-optimization/60650
6202 * lra-constraints.c (process_alt_operands): Decrease reject for
6203 earlyclobber matching.
6205 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6207 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6209 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6211 * config/spu/spu.c (pad_bb): Do not crash when the last
6212 insn is CODE_FOR_blockage.
6214 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6216 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6217 lies outside the target mode.
6219 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6222 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6223 software floating point or no floating point registers, do not
6224 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6225 in GPRs that occurs after we tested for GPRs that would never be
6228 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6229 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6230 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6231 specifically allow DDmode, since that does not use the SPE SIMD
6234 2014-04-02 Richard Biener <rguenther@suse.de>
6237 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6238 MODE_INTs. Properly use negv_optab.
6239 (expand_abs): Likewise.
6241 2014-04-02 Richard Biener <rguenther@suse.de>
6244 * Makefile.in (install-driver): Guard extra installs with special
6247 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6249 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6252 2014-04-01 Richard Henderson <rth@redhat.com>
6255 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6256 alternative enabled before register allocation.
6258 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6260 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6261 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6263 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6264 (nios2_got_address): Update nios2_large_got_address call site.
6265 (nios2_delegitimize_address): New function.
6266 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6267 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6268 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6270 2014-04-01 Martin Husemann <martin@duskware.de>
6272 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6275 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6277 PR rtl-optimization/60604
6278 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6279 check from register_operand.
6280 (register_operand): Redefine in terms of general_operand.
6281 (nonmemory_operand): Use register_operand for the non-constant cases.
6283 2014-04-01 Richard Biener <rguenther@suse.de>
6285 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6287 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6289 * doc/invoke.texi (mapp-regs): Clarify.
6291 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6293 * config/i386/avx512fintrin.h (__v32hi): Define type.
6294 (__v64qi): Likewise.
6295 (_mm512_set1_epi8): Define.
6296 (_mm512_set1_epi16): Define.
6297 (_mm512_set4_epi32): Define.
6298 (_mm512_set4_epi64): Define.
6299 (_mm512_set4_pd): Define.
6300 (_mm512_set4_ps): Define.
6301 (_mm512_setr4_epi64): Define.
6302 (_mm512_setr4_epi32): Define.
6303 (_mm512_setr4_pd): Define.
6304 (_mm512_setr4_ps): Define.
6305 (_mm512_setzero_epi32): Define.
6307 2014-03-31 Martin Jambor <mjambor@suse.cz>
6310 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6311 callsite_arguments_match_p. Updated all callers. Also check types of
6312 corresponding formal parameters and actual arguments.
6313 (not_all_callers_have_enough_arguments_p) Renamed to
6314 some_callers_have_mismatched_arguments_p.
6316 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6318 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6320 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6323 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6326 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6328 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6329 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6331 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6332 Use FMAMODE_NOVF512 mode iterator.
6333 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6334 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6335 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6336 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6338 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6339 Use VF_128_256 mode iterator.
6340 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6343 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6345 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6346 static chain if needed.
6348 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6351 * lra-constraints.c (index_part_to_reg): New.
6352 (process_address): Use it.
6354 2014-03-27 Jeff Law <law@redhat.com>
6355 Jakub Jelinek <jakub@redhat.com>
6358 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6359 gen_rtx_{PLUS,MULT} to build up the address expression.
6361 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6362 creating non-canonical RTL.
6364 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6367 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6368 functions; reorganize to make cheap checks first.
6369 (inline_small_functions): Do not estimate growth when dumping;
6371 * ipa-inline.h (inline_summary): Add min_size.
6372 (growth_likely_positive): New function.
6373 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6374 (set_cond_stmt_execution_predicate): Cleanup.
6375 (estimate_edge_size_and_time): Compute min_size.
6376 (estimate_calls_size_and_time): Likewise.
6377 (estimate_node_size_and_time): Likewise.
6378 (inline_update_overall_summary): Update min_size.
6379 (do_estimate_edge_time): Likewise.
6380 (do_estimate_edge_size): Update.
6381 (do_estimate_edge_hints): Update.
6382 (growth_likely_positive): New function.
6384 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6387 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6388 also if addr has VOIDmode.
6390 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6392 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6393 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6395 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6396 instructions as well as AdvancedSIMD loads.
6398 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6400 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6401 Use crypto_aese type.
6402 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6403 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6404 crypto_aese, crypto_aesmc. Move to types.md.
6405 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6407 * config/arm/iterators.md (crypto_type): Likewise.
6409 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6411 * cgraph.c: Include expr.h and tree-dfa.h.
6412 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6415 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6418 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6419 regs from checking multi-reg pseudos.
6421 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6423 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6425 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6427 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6428 if it would clobber the stack pointer, even temporarily.
6430 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6432 * mode-switching.c: Make small adjustments to the top comment.
6434 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6436 * config/rs6000/constraints.md (wD constraint): New constraint to
6437 match the constant integer to get the top DImode/DFmode out of a
6438 vector in a VSX register.
6440 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6441 match the constant integer to get the top DImode/DFmode out of a
6442 vector in a VSX register.
6444 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6447 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6450 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6451 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6453 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6454 Optimize vec_extract of 64-bit values, where the value being
6455 extracted is in the top word, where we can use scalar
6456 instructions. Add direct move and store support. Combine the big
6457 endian/little endian vector select load support into a single insn.
6458 (vsx_extract_<mode>_internal1): Likewise.
6459 (vsx_extract_<mode>_internal2): Likewise.
6460 (vsx_extract_<mode>_load): Likewise.
6461 (vsx_extract_<mode>_store): Likewise.
6462 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6463 combined into vsx_extract_<mode>_load.
6464 (vsx_extract_<mode>_one_le): Likewise.
6466 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6467 define the top 64-bit vector element.
6469 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6472 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6473 Document vec_vbpermq builtin.
6476 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6477 enable use of xxsldwi and xxpermdi builtin functions.
6478 (vec_xxpermdi): Likewise.
6480 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6481 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6483 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6485 PR rtl-optimization/60650
6486 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6488 (find_spills_for): New.
6489 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6490 Spill all pseudos on the second iteration.
6492 2014-03-27 Marek Polacek <polacek@redhat.com>
6495 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6498 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6500 * config/s390/s390.c (s390_can_use_return_insn): Check for
6501 call-saved FPRs on 31 bit.
6503 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6506 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6507 if they need regimplification, just drop them instead of
6508 calling gimple_regimplify_operands on them.
6510 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6513 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6514 (aarch64_frame_pointer_required): Adjust logic.
6515 (aarch64_can_eliminate): Adjust logic.
6516 (aarch64_override_options_after_change): Adjust logic.
6518 2014-03-27 Dehao Chen <dehao@google.com>
6520 * ipa-inline.c (early_inliner): Update node's inline info.
6522 2014-03-26 Dehao Chen <dehao@google.com>
6524 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6525 compiler inserted conditional jumps for NAN float check.
6527 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6529 * ubsan.h (ubsan_create_data): Change second argument's type
6530 to const location_t *.
6531 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6533 (ubsan_create_data): Change second argument to const location_t *PLOC.
6534 Create Loc field whenever PLOC is non-NULL.
6535 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6536 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6540 * real.c (real_to_integer2): Change type of low to UHWI.
6542 2014-03-26 Tobias Burnus <burnus@net-b.de>
6544 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6545 (CILK_SELF_SPECS): New define.
6546 (driver_self_specs): Use it.
6548 2014-03-26 Richard Biener <rguenther@suse.de>
6550 * tree-pretty-print.c (percent_K_format): Implement special
6551 case for LTO and its stripped down BLOCK tree.
6553 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6556 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6558 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6559 one range is range_int_cst_p, but not both, at least optimize
6560 addition/subtraction of 0 and multiplication by 0 or 1.
6561 * gimple-fold.c (gimple_fold_call): Fold
6562 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6563 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6564 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6566 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6568 PR rtl-optimization/60452
6569 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6570 <case REG>: Return 1 for invalid offsets from the frame pointer.
6572 2014-03-26 Marek Polacek <polacek@redhat.com>
6575 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6578 2014-03-26 Marek Polacek <polacek@redhat.com>
6581 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6584 2014-03-26 Marek Polacek <polacek@redhat.com>
6587 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6588 multiplication in unsigned type.
6590 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6592 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6594 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6596 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6598 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6601 * cif-code.def (UNREACHABLE) New code.
6602 * ipa-inline.c (inline_small_functions): Skip edges to
6603 __builtlin_unreachable.
6604 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6605 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6606 predicate to __bulitin_unreachable.
6607 (set_cond_stmt_execution_predicate): Fix issue when
6608 invert_tree_comparison returns ERROR_MARK.
6609 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6610 propagate to inline clones.
6611 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6613 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6614 * cgraphclones.c (cgraph_clone_node): If call destination is already
6615 ureachable, do not redirect it back.
6616 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6619 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6621 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6622 Do not modify inline clones.
6624 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6626 * config/i386/i386.md (general_sext_operand): New mode attr.
6627 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6628 don't generate (sign_extend (const_int)).
6629 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6630 operands[2]. Use We constraint instead of <i> and
6631 <general_sext_operand> predicate instead of <general_operand>.
6632 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6633 * config/i386/constraints.md (We): New constraint.
6634 * config/i386/predicates.md (x86_64_sext_operand,
6635 sext_operand): New predicates.
6637 2014-03-25 Martin Jambor <mjambor@suse.cz>
6640 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6641 inconsistent devirtualizations to __builtin_unreachable.
6643 2014-03-25 Marek Polacek <polacek@redhat.com>
6646 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6648 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6650 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6651 order of elements for big-endian.
6653 2014-03-25 Richard Biener <rguenther@suse.de>
6656 * gimplify-me.c (gimple_regimplify_operands): Update the
6659 2014-03-25 Martin Jambor <mjambor@suse.cz>
6662 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6663 (lto_output_varpool_node): Likewise.
6664 (input_overwrite_node): Likewise.
6665 (input_varpool_node): Likewise.
6667 2014-03-25 Richard Biener <rguenther@suse.de>
6669 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6670 (run_gcc): Likewise.
6672 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6674 * combine.c (simplify_compare_const): Add MODE argument.
6675 Handle mode_width 0 as very large mode_width.
6676 (try_combine, simplify_comparison): Adjust callers.
6678 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6679 type to avoid signed integer overflow.
6680 * explow.c (plus_constant): Likewise.
6682 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6684 * doc/generic.texi: Correct typos.
6686 2014-03-24 Tobias Burnus <burnus@net-b.de>
6688 * doc/invoke.texi (-flto): Expand section about
6689 using static libraries with LTO.
6691 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6693 PR rtl-optimization/60501
6694 * optabs.def (addptr3_optab): New optab.
6695 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6696 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6697 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6699 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6701 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6703 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6705 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6708 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6709 (_mm256_undefined_ps): Define.
6710 (_mm256_undefined_pd): Define.
6711 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6712 (_mm_undefined_pd): Define.
6713 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6714 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6715 (_mm512_undefined_ps): Define.
6716 (_mm512_undefined_pd): Define.
6717 Use _mm*_undefined_*.
6718 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6720 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6722 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6723 (lshr_simd): DI mode added.
6724 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6725 (aarch64_ushr_simddi): Likewise.
6726 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6727 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6728 (vshrd_n_u64): Likewise.
6730 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6732 * Makefile.in (s-macro_list): Depend on cc1.
6734 2014-03-23 Teresa Johnson <tejohnson@google.com>
6736 * ipa-utils.c (ipa_print_order): Use specified dump file.
6738 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6740 PR rtl-optimization/60601
6741 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6743 * gcc.c (eval_spec_function): Initialize save_growing_value.
6745 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6748 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6749 code == MINUS_EXPR, never swap op0 with op1.
6751 * toplev.c (init_local_tick): Avoid signed integer multiplication
6753 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6754 shift by first operand's bitsize.
6756 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6759 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6761 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6762 TARGET_ISA_64BIT_P(x).
6764 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6766 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6767 pattern for vector nor instead of subtract from splat(-1).
6768 (altivec_expand_vec_perm_const_le): Likewise.
6770 2014-03-21 Richard Henderson <rth@twiddle.net>
6773 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6774 related insns after epilogue_completed.
6776 2014-03-21 Martin Jambor <mjambor@suse.cz>
6779 * cgraph.h (symtab_node): New flag body_removed.
6780 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6781 when removing bodies.
6782 * symtab.c (dump_symtab_base): Dump body_removed flag.
6783 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6784 had their bodies removed.
6786 2014-03-21 Martin Jambor <mjambor@suse.cz>
6789 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6792 2014-03-21 Richard Biener <rguenther@suse.de>
6794 PR tree-optimization/60577
6795 * tree-core.h (struct tree_base): Document nothrow_flag use
6796 in VAR_DECL_NONALIASED.
6797 * tree.h (VAR_DECL_NONALIASED): New.
6798 (may_be_aliased): Adjust.
6799 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6801 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6803 * expr.c (expand_expr_real_1): Remove outdated comment.
6805 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6808 * ira.c (adjust_cleared_regs): Call copy_rtx on
6809 *reg_equiv[REGNO (loc)].src_p before passing it to
6810 simplify_replace_fn_rtx.
6813 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6814 into CONST, put pic register as first operand of PLUS. Use
6815 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6817 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6819 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6821 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6823 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6824 around for store forwarding issue in the FPU on the UT699.
6825 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6826 loads and operations if -mfix-ut699 is specified.
6827 (divtf3_hq): Tweak attribute.
6828 (sqrttf2_hq): Likewise.
6830 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6832 * calls.c (store_one_arg): Remove incorrect const qualification on the
6833 type of the temporary.
6834 * cfgexpand.c (expand_return): Likewise.
6835 * expr.c (expand_constructor): Likewise.
6836 (expand_expr_real_1): Likewise.
6838 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6840 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6843 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6846 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6848 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6850 * config/arm/aarch-common-protos.h
6851 (alu_cost_table): Fix spelling of "extend".
6852 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6854 2014-03-19 Richard Biener <rguenther@suse.de>
6857 * tree-core.h (tree_type_common): Re-order pointer members
6858 to reduce recursion depth during GC walks.
6860 2014-03-19 Marek Polacek <polacek@redhat.com>
6863 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6864 before accessing it.
6866 2014-03-19 Richard Biener <rguenther@suse.de>
6869 * lto-streamer-in.c (input_function): In WPA stage do not drop
6872 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6874 PR tree-optimization/60559
6875 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6876 with build_zero_cst assignment.
6878 2014-03-18 Kai Tietz <ktietz@redhat.com>
6880 PR rtl-optimization/56356
6881 * sdbout.c (sdbout_parms): Verify that parms'
6882 incoming argument is valid.
6883 (sdbout_reg_parms): Likewise.
6885 2014-03-18 Richard Henderson <rth@redhat.com>
6888 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6889 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6890 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6892 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6894 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6895 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6896 Italicize plugin event names in description. Explain that
6897 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6898 Remind that no GCC functions should be called after PLUGIN_FINISH.
6899 Explain what pragmas with expansion are.
6901 2014-03-18 Martin Liska <mliska@suse.cz>
6903 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6904 gimple call statement is update.
6905 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6906 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6908 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6911 * ubsan.c (ubsan_instrument_unreachable): Call
6912 initialize_sanitizer_builtins.
6913 (ubsan_pass): Likewise.
6916 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6917 varpool_finalize_decl instead of rest_of_decl_compilation.
6919 2014-03-18 Richard Biener <rguenther@suse.de>
6921 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6922 by using bitmap_and_compl instead of bitmap_and_compl_into.
6923 (df_rd_transfer_function): Likewise.
6925 2014-03-18 Richard Biener <rguenther@suse.de>
6927 * doc/lto.texi (fresolution): Fix typo.
6929 2014-03-18 Richard Biener <rguenther@suse.de>
6931 * doc/invoke.texi (flto): Update for changes in 4.9.
6933 2014-03-18 Richard Biener <rguenther@suse.de>
6935 * doc/loop.texi: Remove section on the removed lambda framework.
6936 Update loop docs with recent changes in preserving loop structure.
6938 2014-03-18 Richard Biener <rguenther@suse.de>
6940 * doc/lto.texi (-fresolution): Document.
6942 2014-03-18 Richard Biener <rguenther@suse.de>
6944 * doc/contrib.texi: Adjust my name.
6946 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6949 * internal-fn.c: Include diagnostic-core.h.
6950 (expand_BUILTIN_EXPECT): New function.
6951 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6952 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6953 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6954 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6956 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6957 Revert 3 argument __builtin_expect code.
6958 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6959 * gimple-fold.c (gimple_fold_call): Likewise.
6960 * tree.h (fold_builtin_expect): New prototype.
6961 * builtins.c (build_builtin_expect_predicate): Add predictor
6962 argument, if non-NULL, create 3 argument __builtin_expect.
6963 (fold_builtin_expect): No longer static. Add ARG2 argument,
6964 pass it through to build_builtin_expect_predicate.
6965 (fold_builtin_2): Adjust caller.
6966 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6967 * internal-fn.def (BUILTIN_EXPECT): New.
6969 2014-03-18 Tobias Burnus <burnus@net-b.de>
6972 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6973 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6974 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6976 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6979 * predict.c (combine_predictions_for_bb): Fix up formatting.
6980 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6981 fill what it points to if non-NULL.
6982 (tree_predict_by_opcode): Adjust caller, use the predictor.
6983 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6985 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6987 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6988 proper constant for the store mode.
6990 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6992 * symtab.c (change_decl_assembler_name): Fix transparent alias
6995 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6997 * config/aarch64/aarch64.c: Correct the comments about the
6998 aarch64 stack layout.
7000 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7002 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7003 check for GF_OMP_FOR_KIND_FOR.
7005 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7007 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7008 ymm and zmm register names.
7010 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7013 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7014 note creation for the 2010-08-31 changes.
7016 2014-03-17 Marek Polacek <polacek@redhat.com>
7019 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7020 as -fno-tree-loop-vectorize.
7021 (expand_omp_simd): Likewise.
7023 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7025 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7026 (eligible_for_call_delay): New prototype.
7027 * config/sparc/sparc.c (tls_call_delay): Rename into...
7028 (eligible_for_call_delay): ...this. Return false if the instruction
7029 cannot be put in the delay slot of a branch.
7030 (eligible_for_restore_insn): Simplify.
7031 (eligible_for_return_delay): Return false if the instruction cannot be
7032 put in the delay slot of a branch and simplify.
7033 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7034 put in the delay slot of a branch.
7035 * config/sparc/sparc.md (fix_ut699): New attribute.
7036 (tls_call_delay): Delete.
7037 (in_call_delay): Reimplement.
7038 (eligible_for_sibcall_delay): Rename into...
7039 (in_sibcall_delay): ...this.
7040 (eligible_for_return_delay): Rename into...
7041 (in_return_delay): ...this.
7042 (in_branch_delay): Reimplement.
7043 (in_uncond_branch_delay): Delete.
7044 (in_annul_branch_delay): Delete.
7046 2014-03-14 Richard Henderson <rth@redhat.com>
7049 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7050 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7051 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7052 (floathi splitters): Remove.
7053 (float<SWI48x>xf2): New pattern.
7054 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7055 code that tried to handle DImode for 32-bit, but which was excluded
7056 by the pattern's condition. Drop allocation of stack temporary.
7057 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7058 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7059 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7060 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7061 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7062 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7063 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7064 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7065 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7066 (*float<SWI48x><X87MODEF>2_i387): Remove.
7067 (all float _with_temp splitters): Remove.
7068 (*float<SWI48x><MODEF>2_i387): New pattern.
7069 (*float<SWI48><MODEF>2_sse): New pattern.
7070 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7071 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7073 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7074 Marek Polacek <polacek@redhat.com>
7077 * common.opt (dump_base_name_prefixed): New Variable.
7078 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7079 if x_dump_base_name_prefixed is already set, set it at the end.
7081 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7083 PR rtl-optimization/60508
7084 * lra-constraints.c (get_reload_reg): Add new parameter
7086 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7087 Pass the new parameter values.
7089 2014-03-14 Richard Biener <rguenther@suse.de>
7091 * common.opt: Revert unintented changes from r205065.
7094 2014-03-14 Richard Biener <rguenther@suse.de>
7097 * cfghooks.c (split_block): Properly adjust all loops the
7098 block was a latch of.
7100 2014-03-14 Martin Jambor <mjambor@suse.cz>
7103 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7106 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7109 * config/avr/avr.c (avr_set_current_function): Pass function name
7110 through default_strip_name_encoding before sanity checking instead
7111 of skipping the first char of the assembler name.
7113 2014-03-13 Richard Henderson <rth@redhat.com>
7116 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7117 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7118 * config/i386/i386-protos.h: Likewise.
7119 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7120 in the expander instead of a splitter.
7121 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7122 any possibility of requiring a memory.
7123 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7124 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7125 (fp branch splitters): Update for ix86_split_fp_branch.
7126 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7127 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7128 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7129 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7130 (*fop_<MODEF>_3_i387): Likewise.
7131 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7132 (splitters for the fop_* register patterns): Remove.
7133 (fscalexf4_i387): Rename from *fscalexf4_i387.
7134 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7136 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7138 PR tree-optimization/59779
7139 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7140 type for bitsize and maxsize instead of HOST_WIDE_INT.
7142 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7144 PR rtl-optimization/57320
7145 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7146 the CFG after thread_prologue_and_epilogue_insns.
7148 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7150 PR rtl-optimization/57189
7151 * lra-constraints.c (process_alt_operands): Disfavor spilling
7154 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7156 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7158 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7160 PR tree-optimization/59025
7162 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7163 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7165 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7168 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7169 calls of avr_out_plus_1.
7171 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7173 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7174 BB's single pred and update the father loop's latch info later.
7176 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7178 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7182 (VEC_base): Likewise.
7183 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7184 registers, we need to swap double words in little endian mode.
7186 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7187 to be a container mode for 128-bit integer operations added in ISA
7188 2.07. Unlike TImode and PTImode, the preferred register set is
7189 the Altivec/VMX registers for the 128-bit operations.
7191 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7193 (rs6000_split_128bit_ok_p): Likewise.
7195 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7196 macros for creating ISA 2.07 normal and overloaded builtin
7197 functions with 3 arguments.
7198 (BU_P8V_OVERLOAD_3): Likewise.
7199 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7200 for use as overloaded functions.
7201 (VPERM_1TI_UNS): Likewise.
7202 (VSEL_1TI): Likewise.
7203 (VSEL_1TI_UNS): Likewise.
7204 (ST_INTERNAL_1ti): Likewise.
7205 (LD_INTERNAL_1ti): Likewise.
7206 (XXSEL_1TI): Likewise.
7207 (XXSEL_1TI_UNS): Likewise.
7208 (VPERM_1TI): Likewise.
7209 (VPERM_1TI_UNS): Likewise.
7210 (XXPERMDI_1TI): Likewise.
7211 (SET_1TI): Likewise.
7212 (LXVD2X_V1TI): Likewise.
7213 (STXVD2X_V1TI): Likewise.
7214 (VEC_INIT_V1TI): Likewise.
7215 (VEC_SET_V1TI): Likewise.
7216 (VEC_EXT_V1TI): Likewise.
7217 (EQV_V1TI): Likewise.
7218 (NAND_V1TI): Likewise.
7219 (ORC_V1TI): Likewise.
7220 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7221 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7223 (VADDUQM): Likewise.
7224 (VSUBCUQ): Likewise.
7225 (VADDEUQM): Likewise.
7226 (VADDECUQ): Likewise.
7227 (VSUBEUQM): Likewise.
7228 (VSUBECUQ): Likewise.
7230 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7231 __int128_t and __uint128_t types.
7232 (__uint128_type): Likewise.
7233 (altivec_categorize_keyword): Add support for vector __int128_t,
7234 vector __uint128_t, vector __int128, and vector unsigned __int128
7235 as a container type for TImode operations that need to be done in
7236 VSX/Altivec registers.
7237 (rs6000_macro_to_expand): Likewise.
7238 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7239 to support 128-bit integer instructions vaddcuq, vadduqm,
7240 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7241 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7243 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7244 for V1TImode, and set up preferences to use VSX/Altivec registers.
7245 Setup VSX reload handlers.
7246 (rs6000_debug_reg_global): Likewise.
7247 (rs6000_init_hard_regno_mode_ok): Likewise.
7248 (rs6000_preferred_simd_mode): Likewise.
7249 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7250 (easy_altivec_constant): Likewise.
7251 (output_vec_const_move): Likewise.
7252 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7254 (rs6000_expand_vector_extract): Likewise.
7255 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7257 (rs6000_const_vec): Add support for V1TImode.
7258 (rs6000_emit_le_vsx_load): Swap double words when loading or
7259 storing TImode/V1TImode.
7260 (rs6000_emit_le_vsx_store): Likewise.
7261 (rs6000_emit_le_vsx_move): Likewise.
7262 (rs6000_emit_move): Add support for V1TImode.
7263 (altivec_expand_ld_builtin): Likewise.
7264 (altivec_expand_st_builtin): Likewise.
7265 (altivec_expand_vec_init_builtin): Likewise.
7266 (altivec_expand_builtin): Likewise.
7267 (rs6000_init_builtins): Add support for V1TImode type. Add
7268 support for ISA 2.07 128-bit integer builtins. Define type names
7269 for the VSX/Altivec vector types.
7270 (altivec_init_builtins): Add support for overloaded vector
7271 functions with V1TImode type.
7272 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7273 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7275 (rs6000_split_128bit_ok_p): Likewise.
7276 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7277 __int128_t and vector __uint128_t.
7279 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7280 and mode attributes.
7287 (VS_scalar): Likewise.
7288 (VS_double): Likewise.
7289 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7291 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7292 we support the ISA 2.07 128-bit integer arithmetic instructions.
7293 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7294 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7295 and TImode types for use with the builtin functions.
7296 (V1TI_type_node): Likewise.
7297 (unsigned_V1TI_type_node): Likewise.
7298 (intTI_type_internal_node): Likewise.
7299 (uintTI_type_internal_node): Likewise.
7301 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7302 128-bit builtin functions.
7303 (UNSPEC_VADDEUQM): Likewise.
7304 (UNSPEC_VADDECUQ): Likewise.
7305 (UNSPEC_VSUBCUQ): Likewise.
7306 (UNSPEC_VSUBEUQM): Likewise.
7307 (UNSPEC_VSUBECUQ): Likewise.
7308 (VM): Add V1TImode to vector mode iterators.
7310 (VI_unit): Likewise.
7311 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7312 (altivec_vaddcuq): Likewise.
7313 (altivec_vsubuqm): Likewise.
7314 (altivec_vsubcuq): Likewise.
7315 (altivec_vaddeuqm): Likewise.
7316 (altivec_vaddecuq): Likewise.
7317 (altivec_vsubeuqm): Likewise.
7318 (altivec_vsubecuq): Likewise.
7320 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7322 (BOOL_128): Likewise.
7323 (BOOL_REGS_OUTPUT): Likewise.
7324 (BOOL_REGS_OP1): Likewise.
7325 (BOOL_REGS_OP2): Likewise.
7326 (BOOL_REGS_UNARY): Likewise.
7327 (BOOL_REGS_AND_CR0): Likewise.
7329 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7330 128-bit integer builtin support.
7331 (vec_vadduqm): Likewise.
7332 (vec_vaddecuq): Likewise.
7333 (vec_vaddeuqm): Likewise.
7334 (vec_vsubecuq): Likewise.
7335 (vec_vsubeuqm): Likewise.
7336 (vec_vsubcuq): Likewise.
7337 (vec_vsubuqm): Likewise.
7339 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7340 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7341 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7342 128-bit integer add/subtract to ISA 2.07.
7344 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7346 * config/arc/arc.c (arc_predicate_delay_insns):
7347 Fix third argument passed to conditionalize_nonjump.
7349 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7351 * config/aarch64/aarch64-builtins.c
7352 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7353 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7354 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7355 instead of __builtin_lfloor.
7356 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7358 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7360 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7361 (tree_ssa_ifcombine_bb_1): New function.
7362 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7363 is an empty forwarder block to then_bb or vice versa and then_bb
7364 and else_bb are effectively swapped.
7366 2014-03-12 Christian Bruel <christian.bruel@st.com>
7369 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7370 REG_CFA_DEF_CFA note.
7371 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7372 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7374 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7376 PR tree-optimization/60454
7377 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7379 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7381 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7382 Do not define target_cpu_default2 to generic.
7383 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7384 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7385 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7387 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7388 Marc Glisse <marc.glisse@inria.fr>
7390 PR tree-optimization/60502
7391 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7392 instead of build_low_bits_mask.
7394 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7397 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7398 if there are multiple uses, but op doesn't live on E edge.
7399 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7400 clobber stmts before __builtin_unreachable.
7402 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7404 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7405 hard_frame_pointer_rtx.
7406 * cse.c (cse_insn): Remove volatile check.
7407 * cselib.c (cselib_process_insn): Likewise.
7408 * dse.c (scan_insn): Likewise.
7410 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7412 * config/arc/arc.c (conditionalize_nonjump): New function,
7414 (arc_ifcvt): ... this.
7415 (arc_predicate_delay_insns): Use it.
7417 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7419 * config/arc/predicates.md (extend_operand): During/after reload,
7420 allow const_int_operand.
7421 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7422 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7423 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7425 (umulsi3_highpart_i): Likewise.
7427 2014-03-11 Richard Biener <rguenther@suse.de>
7429 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7430 Add asserts to guard possible wrong-code bugs.
7432 2014-03-11 Richard Biener <rguenther@suse.de>
7434 PR tree-optimization/60429
7435 PR tree-optimization/60485
7436 * tree-ssa-structalias.c (set_union_with_increment): Properly
7437 take into account all fields that overlap the shifted vars.
7438 (do_sd_constraint): Likewise.
7439 (do_ds_constraint): Likewise.
7440 (get_constraint_for_ptr_offset): Likewise.
7442 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7444 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7445 (nios2_compute_frame_layout):
7446 Add calculation of cfun->machine->fp_save_offset.
7447 (nios2_expand_prologue): Correct setting of frame pointer register
7449 (nios2_expand_epilogue): Update recovery of stack pointer from
7450 frame pointer accordingly.
7451 (nios2_initial_elimination_offset): Update calculation of offset
7452 for eliminating to HARD_FRAME_POINTER_REGNUM.
7454 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7457 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7458 cgraph_get_create_node on VAR_DECLs.
7460 2014-03-10 Richard Biener <rguenther@suse.de>
7463 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7465 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7467 * config/vms/vms.opt (vms_float_format): New variable.
7469 2014-03-08 Tobias Burnus <burnus@net-b.de>
7471 * doc/invoke.texi (-fcilkplus): Update implementation status.
7473 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7474 Richard Biener <rguenther@suse.de>
7476 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7477 consistently accross all TUs.
7478 (run_gcc): Enable -fshort-double automatically at link at link-time
7479 and disallow override.
7481 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7484 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7485 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7486 if they can't be used.
7488 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7490 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7491 for Solaris 11/x86 ld.
7492 * configure: Regenerate.
7494 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7496 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7497 (LIB_TLS_SPEC): Save as ld_tls_libs.
7498 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7499 (HAVE_AS_IX86_TLSLDM): New test.
7500 * configure, config.in: Regenerate.
7501 * config/i386/i386.c (legitimize_tls_address): Fall back to
7502 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7503 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7504 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7505 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7507 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7509 * common.opt (fira-loop-pressure): Mark as optimization.
7511 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7513 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7514 an OpenMP mappable type.
7516 2014-03-06 Matthias Klose <doko@ubuntu.com>
7518 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7519 MULTILIB_OSDIRNAMES is not defined.
7521 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7522 Meador Inge <meadori@codesourcery.com>
7525 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7526 (arm_legitimize_address): Call legitimize_tls_address for any
7527 arm_tls_referenced_p expression, handle constant addend. Call it
7528 before testing for !TARGET_ARM.
7529 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7531 2014-03-06 Richard Biener <rguenther@suse.de>
7537 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7539 * tree-streamer.c (record_common_node): Assert we don't record
7540 nodes with type double.
7541 (preload_common_node): Skip type double, complex double and double
7542 pointer since it is now frontend dependent due to fshort-double option.
7544 2014-03-06 Richard Biener <rguenther@suse.de>
7546 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7547 or -fno-lto is specified and the linker has full plugin support.
7548 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7549 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7550 * lto-wrapper.c (merge_and_complain): Merge compile-time
7551 optimization levels.
7552 (run_gcc): And pass it through to the link options.
7554 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7558 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7560 * cselib.c (remove_useless_values): Skip to avoid quadratic
7561 behavior if the condition moved from...
7562 (cselib_process_insn): ... here holds.
7564 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7567 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7568 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7571 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7572 (TM_H): Add x86-tune.def.
7574 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7576 * config/aarch64/aarch64.c (generic_tunings):
7577 Use cortexa57_extra_costs.
7579 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7582 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7583 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7584 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7587 2014-03-04 Heiher <r@hev.cc>
7589 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7590 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7592 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7594 * config/i386/predicates.md (const2356_operand): Change to ...
7595 (const2367_operand): ... this.
7596 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7598 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7599 (*avx512pf_scatterpf<mode>sf): Ditto.
7600 (avx512pf_scatterpf<mode>df): Ditto.
7601 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7602 (*avx512pf_scatterpf<mode>df): Ditto.
7603 * config/i386/i386.c (ix86_expand_builtin): Update
7604 incorrect hint operand error message.
7606 2014-03-04 Richard Biener <rguenther@suse.de>
7608 * lto-section-in.c (lto_get_section_data): Fix const cast.
7610 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7612 * tree-streamer.c (record_common_node): Assert we don't record
7613 nodes with type double.
7614 (preload_common_node): Skip type double, complex double and double
7615 pointer since it is now frontend dependent due to fshort-double option.
7617 2014-03-04 Richard Biener <rguenther@suse.de>
7620 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7621 (lto_input_toplevel_asms): Likewise.
7622 * lto-section-in.c (lto_get_section_data): Instead do it here
7625 2014-03-04 Richard Biener <rguenther@suse.de>
7627 PR tree-optimization/60382
7628 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7629 dead PHIs a reduction.
7631 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7633 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7635 (_mm_prefetch): Move out of GCC target("sse") pragma.
7636 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7637 GCC target("prfchw") pragma.
7638 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7640 * config/i386/i386.c (ix86_option_override_internal): Enable
7641 -mprfchw with -mprefetchwt1.
7643 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7645 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7648 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7650 * opts.h (CL_PCH_IGNORE): Define.
7651 * targhooks.c (option_affects_pch_p):
7652 Return false for options that have CL_PCH_IGNORE set.
7653 * opt-functions.awk: Process PchIgnore.
7654 * doc/options.texi: Document PchIgnore.
7656 * config/arc/arc.opt (misize): Add PchIgnore property.
7658 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7660 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7661 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7662 constraint on constants to permit them being loaded into
7663 GENERAL_REGS or BASE_REGS.
7665 2014-03-03 Nick Clifton <nickc@redhat.com>
7667 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7668 anti-cacnonical alternatives.
7669 (negandhi3_real): New pattern.
7670 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7672 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7674 * config/avr/avr-mcus.def: Remove atxmega16x1.
7675 * config/avr/avr-tables.opt: Regenerate.
7676 * config/avr/t-multilib: Regenerate.
7677 * doc/avr-mmcu.texi: Regenerate.
7679 2014-03-03 Tobias Grosser <tobias@grosser.es>
7680 Mircea Namolaru <mircea.namolaru@inria.fr>
7682 PR tree-optimization/58028
7683 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7686 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7688 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7689 not handled by recognizers.
7691 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7694 * function.c (expand_function_end): Don't emit
7695 clobber_return_register sequence if clobber_after is a BARRIER.
7696 * cfgexpand.c (construct_exit_block): Append instructions before
7697 return_label to prev_bb.
7699 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7701 * config/rs6000/constraints.md: Document reserved use of "wc".
7703 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7706 * ipa.c (function_and_variable_visibility): When dissolving comdat
7707 group, also set all symbols to local.
7709 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7714 2013-12-14 Jan Hubicka <jh@suse.cz>
7716 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7718 2014-03-02 Jon Beniston <jon@beniston.com>
7724 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7725 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7726 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7727 (simple_return, *simple_return): New patterns
7728 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7729 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7731 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7733 * dwarf2out.c (gen_subprogram_die): Tidy.
7735 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7738 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7739 (*mov_t_msb_neg_negc): ... this new insn.
7741 2014-02-28 Jason Merrill <jason@redhat.com>
7744 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7747 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7750 * dwarf2out.c (decltype_auto_die): New static.
7751 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7752 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7753 (is_cxx_auto): Likewise.
7755 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7757 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7758 we are not using general regs only.
7760 2014-02-28 Richard Biener <rguenther@suse.de>
7763 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7764 previous fix and only allow to remove trivial pre-headers
7765 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7766 (remove_forwarder_block): Properly update the latch of a loop.
7768 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7771 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7772 (cselib_preserved_hash_table): New.
7773 (preserve_constants_and_equivs): Move preserved vals to it.
7774 (cselib_find_slot): Look it up first.
7775 (cselib_init): Initialize it.
7776 (cselib_finish): Release it.
7777 (dump_cselib_table): Dump it.
7779 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7782 * cselib.c (remove_useless_values): Skip to avoid quadratic
7783 behavior if the condition moved from...
7784 (cselib_process_insn): ... here holds.
7786 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7789 * var-tracking.c (vt_initialize): Apply the same condition to
7790 preserve the CFA base value.
7792 2014-02-28 Joey Ye <joey.ye@arm.com>
7795 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7796 if reload in progress or completed.
7798 2014-02-28 Tobias Burnus <burnus@net-b.de>
7801 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7804 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7806 * doc/tm.texi.in (Condition Code Status): Update documention for
7807 relative locations of cc0-setter and cc0-user.
7809 2014-02-27 Jeff Law <law@redhat.com>
7811 PR rtl-optimization/52714
7812 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7813 into two independent simple sets, if I3 is a jump, ensure the
7814 pattern we place into I3 is a (set (pc) ...).
7816 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7817 Jeff Law <law@redhat.com>
7819 PR rtl-optimization/49847
7820 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7821 are in different blocks.
7822 * doc/tm.texi (Condition Code Status): Update documention for
7823 relative locations of cc0-setter and cc0-user.
7825 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7828 * lra.c (lra_emit_add): Check SUBREG too.
7830 2014-02-27 Andreas Schwab <schwab@suse.de>
7832 * config/m68k/m68k.c (m68k_option_override): Disable
7833 -flive-range-shrinkage for classic m68k.
7834 (m68k_override_options_after_change): Likewise.
7836 2014-02-27 Marek Polacek <polacek@redhat.com>
7839 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7840 -Wmaybe-uninitialized.
7842 2014-02-27 Alan Modra <amodra@gmail.com>
7845 * reload1.c (emit_input_reload_insns): When reload_override_in,
7846 set old to rl->in_reg when rl->in_reg is a subreg.
7848 2014-02-26 Richard Biener <rguenther@suse.de>
7851 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7853 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7855 * common/config/i386/predicates.md (const1256_operand): Remove.
7856 (const2356_operand): New.
7857 (const_1_to_2_operand): Remove.
7858 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7859 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7860 (*avx512pf_gatherpf<mode>sf): Ditto.
7861 (avx512pf_gatherpf<mode>df): Ditto.
7862 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7863 (*avx512pf_gatherpf<mode>df): Ditto.
7864 (avx512pf_scatterpf<mode>sf): Ditto.
7865 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7866 (*avx512pf_scatterpf<mode>sf): Ditto.
7867 (avx512pf_scatterpf<mode>df): Ditto.
7868 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7869 (*avx512pf_scatterpf<mode>df): Ditto.
7870 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7872 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7874 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7875 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7876 (_mm512_mask_testn_epi64_mask): Move to ...
7877 * config/i386/avx512cdintrin.h: Here.
7878 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7879 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7880 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7881 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7882 TARGET_AVX512F from TARGET_AVX512CD.
7884 2014-02-26 Richard Biener <rguenther@suse.de>
7887 * ipa.c (walk_polymorphic_call_targets): Properly guard
7888 call to inline_update_overall_summary.
7890 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7893 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7894 and latches only if requested. Fix latch if it is removed.
7895 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7896 LOOPS_HAVE_PREHEADERS.
7898 2014-02-25 Andrew Pinski <apinski@cavium.com>
7900 * builtins.c (expand_builtin_thread_pointer): Create a new target
7901 when the target is NULL.
7903 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7905 PR rtl-optimization/60317
7906 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7907 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7908 * lra-assigns.c: Include params.h.
7909 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7910 other reload pseudos considerations.
7912 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7914 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7915 to use canonical form for nor<mode>3.
7917 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7920 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7923 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7925 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7926 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7927 (ix86_handle_option): Handle OPT_mprefetchwt1.
7928 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7929 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7931 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7932 OPTION_MASK_ISA_PREFETCHWT1.
7933 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7934 (PTA_PREFETCHWT1): New.
7935 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7936 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7937 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7938 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7939 (*prefetch_avx512pf_<mode>_: Change into ...
7940 (*prefetch_prefetchwt1_<mode>: This.
7941 * config/i386/i386.opt (mprefetchwt1): New.
7942 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7943 (_mm_prefetch): Handle intent to write.
7944 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7946 2014-02-25 Richard Biener <rguenther@suse.de>
7949 * emit-rtl.c (mem_attrs_htab): Remove.
7950 (mem_attrs_htab_hash): Likewise.
7951 (mem_attrs_htab_eq): Likewise.
7952 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7953 (init_emit_once): Do not allocate mem_attrs_htab.
7955 2014-02-25 Richard Biener <rguenther@suse.de>
7958 * lto-opts.c (lto_write_options): Output non-explicit conservative
7959 -fwrapv, -fno-trapv and -fno-strict-overflow.
7960 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7961 (run_gcc): And pass them through.
7963 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7965 * sel-sched.c (calculate_new_fences): New parameter ptime.
7966 Calculate it as a maximum over all fence cycles.
7967 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7968 Print the final schedule timing when sched_verbose.
7970 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7972 PR rtl-optimization/60292
7973 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7974 bit fot the fence instruction.
7976 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7978 * calls.h: Fix typo in comment.
7980 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7982 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7983 adjusting offsetable addresses.
7985 2014-02-24 Guozhi Wei <carrot@google.com>
7987 * sparseset.h (sparseset_pop): Fix the wrong index.
7989 2014-02-24 Walter Lee <walt@tilera.com>
7991 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7992 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7994 * common/config/tilegx/tilegx-common.c
7995 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7996 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7998 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7999 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8000 (tilegx_gimplify_va_arg_expr): Handle big endian.
8001 (tilegx_expand_unaligned_load): Ditto.
8002 (tilegx_expand_unaligned_store): Ditto.
8003 (TARGET_RETURN_IN_MSB): New.
8004 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8005 (TARGET_ENDIAN_DEFAULT): New.
8006 (TARGET_BIG_ENDIAN): Handle big endian.
8007 (BYTES_BIG_ENDIAN): Ditto.
8008 (WORDS_BIG_ENDIAN): Ditto.
8009 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8012 * config/tilegx/tilegx.md (extv): Handle big endian.
8014 (insn_st<n>): Ditto.
8015 (insn_st<n>_add<bitsuffix>): Ditto.
8016 (insn_stnt<n>): Ditto.
8017 (insn_stnt<n>_add<bitsuffix>):Ditto.
8018 (vec_interleave_highv8qi): Handle big endian.
8019 (vec_interleave_highv8qi_be): New.
8020 (vec_interleave_highv8qi_le): New.
8021 (insn_v1int_h): Handle big endian.
8022 (vec_interleave_lowv8qi): Handle big endian.
8023 (vec_interleave_lowv8qi_be): New.
8024 (vec_interleave_lowv8qi_le): New.
8025 (insn_v1int_l): Handle big endian.
8026 (vec_interleave_highv4hi): Handle big endian.
8027 (vec_interleave_highv4hi_be): New.
8028 (vec_interleave_highv4hi_le): New.
8029 (insn_v2int_h): Handle big endian.
8030 (vec_interleave_lowv4hi): Handle big endian.
8031 (vec_interleave_lowv4hi_be): New.
8032 (vec_interleave_lowv4hi_le): New.
8033 (insn_v2int_l): Handle big endian.
8034 (vec_interleave_highv2si): Handle big endian.
8035 (vec_interleave_highv2si_be): New.
8036 (vec_interleave_highv2si_le): New.
8037 (insn_v4int_h): Handle big endian.
8038 (vec_interleave_lowv2si): Handle big endian.
8039 (vec_interleave_lowv2si_be): New.
8040 (vec_interleave_lowv2si_le): New.
8041 (insn_v4int_l): Handle big endian.
8042 * config/tilegx/tilegx.opt (mbig-endian): New option.
8043 (mlittle-endian): New option.
8044 * doc/install.texi: Document tilegxbe-linux.
8045 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8047 2014-02-24 Martin Jambor <mjambor@suse.cz>
8050 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8051 there are no parameter descriptors.
8053 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8055 PR rtl-optimization/60268
8056 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8057 initialization to ...
8058 (sched_rgn_init): ... here.
8059 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8061 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8063 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8066 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8068 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8071 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8073 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8074 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8076 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8078 * config/microblaze/predicates.md: Add cmp_op predicate.
8079 * config/microblaze/microblaze.md: Add branch_compare instruction
8080 which uses cmp_op predicate and emits cmp insn before branch.
8081 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8082 to microblaze_expand_conditional_branch and consolidate logic.
8083 (microblaze_expand_conditional_branch): emit branch_compare
8084 insn instead of handling cmp op separate from branch insn.
8086 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8088 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8091 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8093 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8094 define_insn with define_expand and new define_insn
8095 *altivec_lve<VI_char>x_internal.
8096 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8097 and new define_insn *altivec_stve<VI_char>x_internal.
8098 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8100 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8102 (altivec_expand_stvex_be): New function.
8104 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8106 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8107 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8108 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8109 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8111 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8114 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8115 instead of emit_move_insn.
8117 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8119 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8121 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8122 gen_altivec_vsumsws.
8124 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8126 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8127 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8128 (altivec_lvxl_<mode>): New define_expand incorporating
8129 -maltivec=be semantics where needed.
8130 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8131 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8132 semantics where needed.
8133 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8134 (altivec_stvx_<mode>): New define_expand incorporating
8135 -maltivec=be semantics where needed.
8136 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8137 VM2 iterator instead of V4SI.
8138 (altivec_stvxl_<mode>): New define_expand incorporating
8139 -maltivec=be semantics where needed.
8140 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8141 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8142 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8143 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8144 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8145 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8146 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8147 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8148 ALTIVEC_BUILTIN_STVXL.
8149 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8150 (altivec_expand_stvx_be): Likewise.
8151 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8152 (altivec_expand_lvx_be): Likewise.
8153 (altivec_expand_stvx_be): Likewise.
8154 (altivec_expand_builtin): Add cases for
8155 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8156 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8157 (altivec_init_builtins): Add definitions for
8158 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8159 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8161 2014-02-21 Catherine Moore <clm@codesourcery.com>
8163 * doc/invoke.texi (mvirt, mno-virt): Document.
8164 * config/mips/mips.opt (mvirt): New option.
8165 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8167 2014-02-21 Richard Biener <rguenther@suse.de>
8169 PR tree-optimization/60276
8170 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8171 (STMT_VINFO_MIN_NEG_DIST): New macro.
8172 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8173 STMT_VINFO_MIN_NEG_DIST.
8174 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8175 made for negative dependence distances still hold.
8177 2014-02-21 Richard Biener <rguenther@suse.de>
8180 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8181 DECL_INITIAL for globals not in the current function context.
8183 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8185 PR tree-optimization/56490
8186 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8187 * tree-ssa-uninit.c: Include params.h.
8188 (compute_control_dep_chain): Add num_calls argument, return false
8189 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8190 num_calls to recursive call.
8191 (find_predicates): Change dep_chain into normal array,
8192 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8193 variable and adjust compute_control_dep_chain caller.
8194 (find_def_preds): Likewise.
8196 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8198 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8199 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8201 2014-02-21 Nick Clifton <nickc@redhat.com>
8203 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8204 (pushhi1): Likewise.
8205 (popqi1): Add mode to pre_dec.
8208 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8210 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8211 mode for mask of V8SFmode permutation.
8213 2014-02-20 Richard Henderson <rth@redhat.com>
8216 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8217 a new pseudo for OLDVAL.
8219 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8222 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8223 gen_reg_rtx if d->testing_p.
8224 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8225 if d->testing_p and we will certainly return true.
8226 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8229 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8231 * emit-rtl.c (gen_reg_rtx): Assert that
8232 crtl->emit.regno_pointer_align_length is non-zero.
8234 2014-02-20 Richard Henderson <rth@redhat.com>
8237 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8238 on failure the store back into EXPECT.
8240 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8241 Sandra Loosemore <sandra@codesourcery.com>
8243 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8244 * config/nios2/nios2.c (nios2_function_profiler): Add
8245 -fPIC (flag_pic == 2) support.
8246 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8247 (nios2_large_offset_p): New function.
8248 (nios2_unspec_reloc_p): Move up position, update to use
8249 nios2_large_offset_p.
8250 (nios2_unspec_address): Remove function.
8251 (nios2_unspec_offset): New function.
8252 (nios2_large_got_address): New function.
8253 (nios2_got_address): Add large offset support.
8254 (nios2_legitimize_tls_address): Update usage of removed and new
8256 (nios2_symbol_binds_local_p): New function.
8257 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8258 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8259 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8260 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8261 processing for (const (unspec ...)).
8262 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8264 2014-02-20 Richard Biener <rguenther@suse.de>
8266 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8267 doing the substitution.
8268 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8270 2014-02-20 Martin Jambor <mjambor@suse.cz>
8273 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8274 info when checking whether lattices are bottom.
8276 2014-02-20 Richard Biener <rguenther@suse.de>
8279 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8282 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8285 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8286 parameter specifying the scaling.
8287 (inline_call): Update.
8288 (want_inline_recursively): Guard division by zero.
8289 (recursive_inlining): Update.
8290 * ipa-inline.h (clone_inlined_nodes): Update.
8292 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8295 * config/i386/i386.c (classify_argument): Pass structures of size
8296 64 bytes or less in register.
8298 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8299 Kirill Yukhin <kirill.yukhin@intel.com>
8301 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8302 (_mm_rcp28_round_ss): Ditto.
8303 (_mm_rsqrt28_round_sd): Ditto.
8304 (_mm_rsqrt28_round_ss): Ditto.
8305 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8306 (_mm_rcp14_round_ss): Ditto.
8307 (_mm_rsqrt14_round_sd): Ditto.
8308 (_mm_rsqrt14_round_ss): Ditto.
8309 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8310 the first input operand, get rid of match_dup.
8311 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8313 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8315 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8316 operand as the first input operand, set type attribute.
8317 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8319 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8320 operand as the first input operand, set type attribute.
8322 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8324 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8327 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8330 * config/i386/i386.c (construct_container): Remove TFmode check
8331 for X86_64_INTEGER_CLASS.
8333 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8336 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8337 only when -Wpsabi is enabled.
8339 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8342 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8343 passing arrays in registers are the same as for structs, so remove the
8344 special case for them.
8346 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8348 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8349 destination type, extract only the valid bits if the source type is not
8350 integral and has a different mode.
8352 2014-02-19 Richard Biener <rguenther@suse.de>
8355 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8358 2014-02-19 Richard Biener <rguenther@suse.de>
8361 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8362 (ipa_modify_call_arguments): Emit an argument load explicitely and
8363 preserve virtual SSA form there and for the replacement call.
8364 Do not update SSA form nor free dominance info.
8366 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8368 * ipa.c (function_and_variable_visibility): Also clear WEAK
8369 flag when disolving COMDAT_GROUP.
8371 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8373 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8374 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8376 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8377 do more sanity checks.
8378 (detect_type_change): Return true when giving up early.
8379 (compute_complex_assign_jump_func): Fix type parameter of
8380 ipa_set_ancestor_jf.
8381 (compute_complex_ancestor_jump_func): Likewise.
8382 (update_jump_functions_after_inlining): Fix updating of
8384 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8386 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8388 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8389 inline clones when edge disappears.
8391 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8394 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8395 Split 64-bit moves into 2 patterns. Do not allow the use of
8396 direct move for TDmode in little endian, since the decimal value
8397 has little endian bytes within a word, but the 64-bit pieces are
8398 ordered in a big endian fashion, and normal subreg's of TDmode are
8400 (mov<mode>_64bit_dm): Likewise.
8401 (movtd_64bit_nodm): Likewise.
8403 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8405 PR tree-optimization/60174
8406 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8407 statement of an SSA_NAME that occurs in an abnormal PHI node.
8409 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8412 * final.c (SEEN_BB): Remove.
8413 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8414 (final_scan_insn): Don't force_source_line on second
8415 NOTE_INSN_BASIC_BLOCK.
8417 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8420 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8421 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8422 (type_natural_mode): Warn ABI change when %zmm register is not
8423 available for AVX512F vector value passing.
8425 2014-02-18 Kai Tietz <ktietz@redhat.com>
8428 * config/i386/i386.c (ix86_expand_prologue): Use value in
8429 rax register as displacement when restoring %r10 or %rax.
8430 Fix wrong offset when restoring both registers.
8432 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8434 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8435 assertion with conditional return.
8437 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8438 Uros Bizjak <ubizjak@gmail.com>
8441 * config/i386/driver-i386.c (host_detect_local_cpu): If
8442 YMM state is not saved by the OS, also clear has_f16c. Move
8443 CPUID 0x80000001 handling before YMM state saving checking.
8445 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8447 PR rtl-optimization/58960
8448 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8449 factored out from ...
8450 (sched_init): ... here.
8451 (free_global_sched_pressure_data): New, factored out from ...
8452 (sched_finish): ... here.
8453 * sched-int.h (free_global_sched_pressure_data): Declare.
8454 * sched-rgn.c (nr_regions_initial): New static global.
8455 (haifa_find_rgns): Initialize it.
8456 (schedule_region): Disable sched-pressure for the newly
8459 2014-02-17 Richard Biener <rguenther@suse.de>
8461 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8462 release SSA defs of pattern stmts.
8464 2014-02-17 Richard Biener <rguenther@suse.de>
8466 * tree-inline.c (expand_call_inline): Release the virtual
8467 operand defined by the call we are about to inline.
8469 2014-02-17 Richard Biener <rguenther@suse.de>
8471 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8473 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8474 Ilya Tocar <ilya.tocar@intel.com>
8476 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8477 arguments order in builtin.
8478 (_mm512_permutexvar_epi64): Ditto.
8479 (_mm512_mask_permutexvar_epi64): Ditto
8480 (_mm512_maskz_permutexvar_epi32): Ditto
8481 (_mm512_permutexvar_epi32): Ditto
8482 (_mm512_mask_permutexvar_epi32): Ditto
8484 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8486 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8487 (p8_vmrgow): Likewise.
8489 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8491 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8494 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8497 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8498 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8499 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8500 using direct move instructions on ISA 2.07. Also adjust
8501 instruction length for 64-bit.
8502 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8503 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8505 2014-02-15 Alan Modra <amodra@gmail.com>
8509 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8510 find_replacement on parts of insn rtl that might be reloaded.
8512 2014-02-15 Richard Biener <rguenther@suse.de>
8514 PR tree-optimization/60183
8515 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8516 (tree_ssa_phiprop): Calculate and free post-dominators.
8518 2014-02-14 Jeff Law <law@redhat.com>
8520 PR rtl-optimization/60131
8521 * ree.c (get_extended_src_reg): New function.
8522 (combine_reaching_defs): Use it rather than assuming location of REG.
8523 (find_and_remove_re): Verify first operand of extension is
8524 a REG before adding the insns to the copy list.
8526 2014-02-14 Roland McGrath <mcgrathr@google.com>
8528 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8529 * configure: Regenerated.
8530 * config.in: Regenerated.
8531 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8532 instead of ASM_SHORT.
8534 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8535 Richard Earnshaw <rearnsha@arm.com>
8537 PR rtl-optimization/59535
8538 * lra-constraints.c (process_alt_operands): Encourage alternative
8539 when unassigned pseudo class is superset of the alternative class.
8540 (inherit_reload_reg): Don't inherit when optimizing for code size.
8541 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8542 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8543 modes not less than 4 for Thumb1.
8545 2014-02-14 Kyle McMartin <kyle@redhat.com>
8548 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8550 2014-02-14 Richard Biener <rguenther@suse.de>
8552 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8553 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8554 Do not strip INDIRECT_REFs.
8556 2014-02-14 Richard Biener <rguenther@suse.de>
8559 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8560 DECL_FUNCTION_SPECIFIC_TARGET.
8561 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8562 * tree-streamer-out.c (pack_ts_target_option): Remove.
8563 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8564 (write_ts_function_decl_tree_pointers): Do not stream
8565 DECL_FUNCTION_SPECIFIC_TARGET.
8566 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8567 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8568 (lto_input_ts_function_decl_tree_pointers): Do not stream
8569 DECL_FUNCTION_SPECIFIC_TARGET.
8571 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8573 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8574 (get_initial_def_for_induction, vectorizable_induction): Ignore
8575 debug stmts when looking for exit_phi.
8576 (vectorizable_live_operation): Fix up condition.
8578 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8580 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8581 nreverse() because it changes the content of original tree list.
8583 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8585 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8586 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8588 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8590 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8591 GNU coding standards.
8593 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8596 * dwarf2out.c (gen_subprogram_die): Don't call
8597 add_calling_convention_attribute if subr_die is old_die.
8599 2014-02-13 Sharad Singhai <singhai@google.com>
8601 * doc/optinfo.texi: Fix order of nodes.
8603 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8605 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8606 operands[2], not operands[3].
8608 2014-02-13 Richard Biener <rguenther@suse.de>
8611 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8612 mention the possibility of an in-tree build.
8613 (CLooG): Update recommended version to 0.18.1, mention the
8614 possibility of an in-tree build and clarify that the ISL
8615 bundled with CLooG does not work.
8617 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8620 * expr.c (compress_float_constant): If x is a hard register,
8621 extend into a pseudo and then move to x.
8623 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8625 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8626 caused by bad second argument to warning_at() with -mhotpatch and
8627 nested functions (e.g. with gfortran).
8629 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8631 * opts.c (option_name): Remove "enabled by default" rider.
8633 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8635 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8637 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8638 Uros Bizjak <ubizjak@gmail.com>
8641 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8642 * configure: Regenerated.
8644 2014-02-12 Richard Biener <rguenther@suse.de>
8646 * vec.c (vec_prefix::calculate_allocation): Move as
8647 inline variant to vec.h.
8648 (vec_prefix::calculate_allocation_1): New out-of-line version.
8649 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8650 (vec_prefix::m_has_auto_buf): Rename to ...
8651 (vec_prefix::m_using_auto_storage): ... this.
8652 (vec_prefix::calculate_allocation): Inline the easy cases
8653 and dispatch to calculate_allocation_1 which doesn't need the
8655 (va_heap::reserve): Use gcc_checking_assert.
8656 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8657 m_using_auto_storage.
8658 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8660 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8661 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8662 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8664 2014-02-12 Richard Biener <rguenther@suse.de>
8666 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8667 when we found a dependence.
8669 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8671 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8673 (maybe_fold_stmt): ... into this new function.
8674 * omp-low.c (lower_omp): Update comment.
8676 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8679 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8682 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8684 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8685 identifiers in comments.
8686 (cortexa53_extra_costs): Likewise.
8687 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8688 (cortexa7_extra_costs): Likewise.
8689 (cortexa12_extra_costs): Likewise.
8690 (cortexa15_extra_costs): Likewise.
8691 (v7m_extra_costs): Likewise.
8693 2014-02-12 Richard Biener <rguenther@suse.de>
8696 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8697 of posix_memalign being successful.
8698 (lower_stmt): Restrict lowering of posix_memalign to when
8699 -ftree-bit-ccp is enabled.
8701 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8703 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8705 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8707 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8709 PR rtl-optimization/60116
8710 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8711 other_insn once the combination has been validated.
8713 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8716 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8718 * ipa-devirt.c: Include demangle.h
8719 (odr_violation_reported): New static variable.
8720 (add_type_duplicate): Update odr_violations.
8721 (maybe_record_node): Add completep parameter; update it.
8722 (record_target_from_binfo): Add COMPLETEP parameter;
8723 update it as needed.
8724 (possible_polymorphic_call_targets_1): Likewise.
8725 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8726 rename FINAL to COMPLETE.
8727 (record_targets_from_bases): Sanity check we found the binfo;
8728 fix COMPLETEP updating.
8729 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8730 parameter, fix computing of COMPLETEP.
8731 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8732 at LTO time do demangling.
8733 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8734 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8736 (gimple_get_virt_method_for_binfo): Likewise.
8737 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8738 gimple_get_virt_method_for_vtable): Update prototypes.
8740 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8743 * genautomata.c (add_presence_absence): Fix typo with
8744 {final_}presence_list.
8746 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8749 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8750 for VSX/Altivec vectors that land in GPR registers.
8752 2014-02-11 Richard Henderson <rth@redhat.com>
8753 Jakub Jelinek <jakub@redhat.com>
8756 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8757 around drhs if type conversion to lacc->type is not useless.
8759 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8761 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8763 (cortex-a57.cortex-a53): Likewise.
8764 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8766 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8768 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8771 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8773 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8774 add_options_for_arm_vfp3.
8776 2014-02-11 Jeff Law <law@redhat.com>
8779 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8780 object with an undesirable mode.
8782 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8785 * config/i386/sol2-9.h: New file.
8786 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8787 *-*-solaris2.9*): Use it.
8789 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8791 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8792 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8794 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8796 * config/microblaze/microblaze.c: Extend mcpu version format
8798 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8800 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8802 2014-02-10 Richard Henderson <rth@redhat.com>
8805 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8806 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8807 ms-abi vs -mno-accumulate-outgoing-args.
8808 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8809 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8812 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8815 * cfgexpand.c (expand_asm_operands): Attach source location to
8816 ASM_INPUT rtx objects.
8817 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8819 2014-02-10 Nick Clifton <nickc@redhat.com>
8821 * config/mn10300/mn10300.c (popcount): New function.
8822 (mn10300_expand_prologue): Include saved registers in stack usage
8825 2014-02-10 Jeff Law <law@redhat.com>
8828 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8829 when changing the SET_DEST of a prior insn to avoid an input reload.
8831 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8833 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8834 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8835 -mcall-openbsd, or -mcall-linux.
8836 (CC1_ENDIAN_BIG_SPEC): Remove.
8837 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8838 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8839 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8840 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8841 and %cc1_endian_default.
8842 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8844 2014-02-10 Richard Biener <rguenther@suse.de>
8846 PR tree-optimization/60115
8847 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8848 MEM_REF handling. Properly verify that the accesses are not
8849 out of the objects bound.
8851 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8853 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8856 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8858 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8859 proper constants and fix formatting.
8860 (possible_polymorphic_call_targets): Fix formatting.
8862 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8863 Ilya Tocar <ilya.tocar@intel.com>
8865 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8866 (_mm512_loadu_epi32): Renamed into...
8867 (_mm512_loadu_si512): This.
8868 (_mm512_storeu_epi32): Renamed into...
8869 (_mm512_storeu_si512): This.
8870 (_mm512_maskz_ceil_ps): Removed.
8871 (_mm512_maskz_ceil_pd): Ditto.
8872 (_mm512_maskz_floor_ps): Ditto.
8873 (_mm512_maskz_floor_pd): Ditto.
8874 (_mm512_floor_round_ps): Ditto.
8875 (_mm512_floor_round_pd): Ditto.
8876 (_mm512_ceil_round_ps): Ditto.
8877 (_mm512_ceil_round_pd): Ditto.
8878 (_mm512_mask_floor_round_ps): Ditto.
8879 (_mm512_mask_floor_round_pd): Ditto.
8880 (_mm512_mask_ceil_round_ps): Ditto.
8881 (_mm512_mask_ceil_round_pd): Ditto.
8882 (_mm512_maskz_floor_round_ps): Ditto.
8883 (_mm512_maskz_floor_round_pd): Ditto.
8884 (_mm512_maskz_ceil_round_ps): Ditto.
8885 (_mm512_maskz_ceil_round_pd): Ditto.
8886 (_mm512_expand_pd): Ditto.
8887 (_mm512_expand_ps): Ditto.
8888 * config/i386/i386.c (ix86_builtins): Remove
8889 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8890 (bdesc_args): Ditto.
8891 * config/i386/predicates.md (const1256_operand): New.
8892 (const_1_to_2_operand): Ditto.
8893 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8894 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8895 (*avx512pf_gatherpf<mode>sf): Ditto.
8896 (avx512pf_gatherpf<mode>df): Ditto.
8897 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8898 (*avx512pf_gatherpf<mode>df): Ditto.
8899 (avx512pf_scatterpf<mode>sf): Ditto.
8900 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8901 (*avx512pf_scatterpf<mode>sf): Ditto.
8902 (avx512pf_scatterpf<mode>df): Ditto.
8903 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8904 (*avx512pf_scatterpf<mode>df): Ditto.
8905 (avx512f_expand<mode>): Removed.
8906 (<shift_insn><mode>3<mask_name>): Change predicate type.
8908 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8910 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8911 not at the end of datarefs vector use ordered_remove to avoid
8912 reordering datarefs vector.
8915 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8916 mark local addressable non-static vars as GOVD_PRIVATE
8917 instead of GOVD_LOCAL.
8918 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8919 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8923 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8924 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8925 assume_aligned or alloc_align attributes.
8926 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8927 arguments. Handle also assume_aligned and alloc_align attributes.
8928 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8929 calls to functions with assume_aligned or alloc_align attributes.
8930 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8932 2014-02-08 Terry Guo <terry.guo@arm.com>
8934 * doc/invoke.texi: Document ARM -march=armv7e-m.
8936 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8938 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8939 flag on __cilkrts_rethrow builtin.
8942 * ipa-cp.c (determine_versionability): Fail at -O0
8943 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8944 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8947 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8950 * tree-inline.c (copy_forbidden): Fail for
8951 __attribute__((optimize (0))) functions.
8953 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8955 * varpool.c: Include pointer-set.h.
8956 (varpool_remove_unreferenced_decls): Variables in other partitions
8957 will not be output; be however careful to not lose information
8960 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8962 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8963 lookup in the vtable constructor.
8965 2014-02-07 Jeff Law <law@redhat.com>
8968 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8969 define_insn_and_split.
8971 * ipa-inline.c (inline_small_functions): Fix typos.
8973 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8975 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8976 (s390_can_use_return_insn): Declare.
8977 * config/s390/s390.h (EPILOGUE_USES): Define.
8978 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8980 (s390_chunkify_start): Handle return JUMP_LABELs.
8981 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8982 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8983 (s390_can_use_return_insn): New functions.
8984 (s390_fix_long_loop_prediction): Handle conditional returns.
8985 (TARGET_SET_UP_BY_PROLOGUE): Define.
8986 * config/s390/s390.md (ANY_RETURN): New code iterator.
8987 (*creturn, *csimple_return, return, simple_return): New patterns.
8989 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8991 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8992 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8993 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8994 REG_CFA_RESTORE list when deciding not to restore a register.
8996 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8998 * config/s390/s390.c: Include tree-pass.h and context.h.
8999 (s390_early_mach): New function, split out from...
9000 (s390_emit_prologue): ...here.
9001 (pass_data_s390_early_mach): New pass structure.
9002 (pass_s390_early_mach): New class.
9003 (s390_option_override): Create and register early_mach pass.
9004 Move to end of file.
9006 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9008 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9009 to match for the exit block.
9011 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9013 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9014 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9015 Reject misaligned operands.
9017 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9019 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9021 2014-02-07 Richard Biener <rguenther@suse.de>
9024 * gimple-low.c (lower_builtin_posix_memalign): New function.
9025 (lower_stmt): Call it to lower posix_memalign in a way
9026 to make alignment info accessible.
9028 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9031 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9032 __builtin_setjmp_receiver.
9034 2014-02-07 Richard Biener <rguenther@suse.de>
9037 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9038 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9039 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9040 Handle BUILT_IN_POSIX_MEMALIGN.
9041 (find_func_clobbers): Likewise.
9042 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9043 (call_may_clobber_ref_p_1): Likewise.
9045 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9048 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9051 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9054 * lto-cgraph.c (lto_output_node): Use
9055 symtab_get_symbol_partitioning_class.
9056 (lto_output_varpool_node): likewise.
9057 (symtab_get_symbol_partitioning_class): Move here from
9059 * cgraph.h (symbol_partitioning_class): Likewise.
9060 (symtab_get_symbol_partitioning_class): Declare.
9062 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9064 * ggc.h (ggc_internal_cleared_alloc): New macro.
9065 * vec.h (vec_safe_copy): Handle memory stats.
9066 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9067 * target-globals.c (save_target_globals): Likewise.
9069 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9072 * expr.c (emit_move_resolve_push): Export; be bit more selective
9073 on when to clear alias set.
9074 * expr.h (emit_move_resolve_push): Declare.
9075 * function.h (struct function): Add tail_call_marked.
9076 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9077 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9078 * config/i386/i386.md (TImode move expander): De not call
9080 (FP push expanders): Preserve memory attributes.
9081 * config/i386/sse.md (push<mode>1): Remove.
9082 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9083 (ix86_expand_push): Remove.
9084 * config/i386/mmx.md (push<mode>1): Remove.
9086 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9088 PR rtl-optimization/60030
9089 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9090 lopart with paradoxical subreg before shifting it up by hprec.
9092 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9094 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9095 Remove extra newline at end of file.
9096 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9097 (arm_issue_rate): Handle cortexa57.
9098 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9099 (cortex-a57.cortex-a53): Likewise.
9101 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9104 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9105 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9107 (arm_expand_prologue): Adjust all callers.
9108 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9109 info, registers also at the lowest numbered registers side. Use
9110 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9114 * var-tracking.c (adjust_mems): Before adding a SET to
9115 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9117 2014-02-06 Alan Modra <amodra@gmail.com>
9120 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9121 change SDmode to DDmode when lra_in_progress.
9123 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9126 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9127 free_data_ref on the dr first, and before goto again also set dr
9128 to the next dr. For simd_lane_access, free old datarefs[i] before
9129 overwriting it. For get_vectype_for_scalar_type failure, don't
9130 free_data_ref if simd_lane_access.
9132 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9135 * tree.h (opts_for_fn): New inline function.
9136 (opt_for_fn): Define.
9137 * config/i386/i386.c (ix86_function_regparm): Use
9138 opt_for_fn (decl, optimize) instead of optimize.
9140 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9142 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9143 for SYMBOL_REF in large memory model.
9145 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9147 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9149 (cortex-a57): Likewise.
9150 (cortex-a57.cortex-a53): Likewise.
9152 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9153 Kugan Vivekanandarajah <kuganv@linaro.org>
9155 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9157 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9159 2014-02-06 Richard Biener <rguenther@suse.de>
9161 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9162 set_loop_copy and initialize_original_copy_tables.
9164 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9166 * config/aarch64/aarch64-simd.md
9167 (aarch64_ashr_simddi): Change QI to SI.
9169 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9170 Jakub Jelinek <jakub@redhat.com>
9173 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9176 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9178 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9179 CODE_FOR_altivec_vpku[hw]um to
9180 CODE_FOR_altivec_vpku[hw]um_direct.
9181 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9182 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9183 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9184 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9186 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9188 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9189 generation for -maltivec=be.
9190 (altivec_vsumsws): Simplify redundant test.
9192 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9194 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9195 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9196 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9197 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9198 gen_altivec_vpkuwum.
9199 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9201 (altivec_vpks<VI_char>ss): Likewise.
9202 (altivec_vpks<VI_char>us): Likewise.
9203 (altivec_vpku<VI_char>us): Likewise.
9204 (altivec_vpku<VI_char>um): Likewise.
9205 (altivec_vpku<VI_char>um_direct): New (copy of
9206 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9208 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9209 target is little endian and -maltivec=be is not specified.
9210 (*altivec_vupkhs<VU_char>_direct): New (copy of
9211 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9212 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9213 target is little endian and -maltivec=be is not specified.
9214 (*altivec_vupkls<VU_char>_direct): New (copy of
9215 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9216 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9217 little endian and -maltivec=be is not specified.
9218 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9219 little endian and -maltivec=be is not specified.
9221 2014-02-05 Richard Henderson <rth@redhat.com>
9224 * combine-stack-adj.c: Revert r206943.
9225 * sched-int.h (struct deps_desc): Add last_args_size.
9226 * sched-deps.c (init_deps): Initialize it.
9227 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9228 contain REG_ARGS_SIZE notes.
9230 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9232 * lto-cgraph.c (asm_nodes_output): Make global.
9233 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9234 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9235 (driver_handle_option): Handle OPT_fwpa.
9237 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9240 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9241 a comment typo and formatting issue. If odr_hash hasn't been
9242 created, return vNULL and set *completep to false.
9245 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9246 bb with no successors.
9248 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9251 * doc/invoke.texi (-march): Clarify documentation for ARM.
9255 2014-02-05 Richard Biener <rguenther@suse.de>
9257 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9258 when not vectorizing because of too many alias checks.
9259 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9260 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9262 2014-02-05 Nick Clifton <nickc@redhat.com>
9264 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9265 accept extended registers in any mode when compiling for the MN10300.
9267 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9269 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9270 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9271 sanitization attributes.
9272 (can_inline_edge_p): Likewise.
9273 (sanitize_attrs_match_for_inline_p): New function.
9275 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9277 * ipa-prop.c (detect_type_change): Shor circuit testing of
9278 type changes on THIS pointer.
9280 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9283 * config/pa/pa.c (legitimize_tls_address): Return original address
9284 if not passed a SYMBOL_REF rtx.
9285 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9287 (pa_emit_move_sequence): Simplify TLS source operands.
9288 (pa_legitimate_constant_p): Reject all TLS constants.
9289 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9290 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9292 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9294 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9295 groups when we know they are controlled by LTO.
9296 * varasm.c (default_binds_local_p_1): If object is in other partition,
9297 it will be resolved locally.
9299 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9301 * config/host-linux.c (linux_gt_pch_use_address): Don't
9302 use SSIZE_MAX because it is not always defined.
9304 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9307 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9308 threshold for pseudo splitting.
9309 (update_ebb_live_info): Process call argument hard registers and
9310 hard registers from insn definition too.
9311 (max_small_class_regs_num): New constant.
9312 (inherit_in_ebb): Update live hard regs through EBBs. Update
9313 reloads_num only for small register classes. Don't split for
9316 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9319 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9322 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9324 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9327 2014-02-04 Marek Polacek <polacek@redhat.com>
9329 * gdbinit.in (pel): Define.
9331 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9333 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9336 2014-02-04 Richard Biener <rguenther@suse.de>
9339 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9340 in function context local.
9341 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9342 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9343 similar to LTO_imported_decl_ref.
9345 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9347 PR tree-optimization/60002
9348 * cgraphclones.c (build_function_decl_skip_args): Clear
9351 PR tree-optimization/60023
9352 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9353 false to gsi_replace.
9354 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9355 has been in some EH region and vec_stmt could throw, add
9356 vec_stmt into the same EH region.
9357 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9358 has no lhs, ignore it.
9359 * internal-fn.c (expand_MASK_LOAD): Likewise.
9362 * tree-inline.c (copy_forbidden): Fail for
9363 __attribute__((optimize (0))) functions.
9366 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9367 is set, copy one less argument.
9368 (expand_simd_clones): Don't subtract clone_info->inbranch
9369 from simd_clone_struct_alloc argument.
9371 PR rtl-optimization/57915
9372 * recog.c (simplify_while_replacing): If all unary/binary/relational
9373 operation arguments are constant, attempt to simplify those.
9376 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9377 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9379 2014-02-04 Richard Biener <rguenther@suse.de>
9381 PR tree-optimization/60012
9382 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9383 TBAA disambiguation to all DDRs.
9385 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9388 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9389 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9391 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9394 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9396 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9398 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9399 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9401 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9404 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9405 to figure out targets of polymorphic calls with known decl.
9406 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9407 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9408 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9409 (get_polymorphic_call_info): ... here.
9410 (get_polymorphic_call_info_from_invariant): New function.
9412 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9414 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9415 lookup via vtable pointer; check for type consistency
9416 and turn inconsitent facts into UNREACHABLE.
9417 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9418 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9419 type inconsistent querries; return UNREACHABLE instead.
9421 2014-02-03 Richard Henderson <rth@twiddle.net>
9424 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9425 already processed this node.
9426 (normalize_one_pred_1): Pass along mark_set.
9427 (normalize_one_pred): Create and destroy a pointer_set_t.
9428 (normalize_one_pred_chain): Likewise.
9430 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9432 PR gcov-profile/58602
9433 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9435 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9438 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9439 -fno-devirtualize; try to devirtualize by the knowledge of
9440 virtual table pointer given by aggregate propagation.
9441 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9442 (ipa_print_node_jump_functions): Dump also offset that
9443 is relevant for polymorphic calls.
9444 (determine_known_aggregate_parts): Add arg_type parameter; use it
9445 instead of determining the type from pointer type.
9446 (ipa_compute_jump_functions_for_edge): Update call of
9447 determine_known_aggregate_parts.
9448 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9449 (gimple_get_virt_method_for_binfo): ... here; simplify using
9450 vtable_pointer_value_to_vtable.
9451 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9452 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9453 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9454 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9456 (vtable_pointer_value_to_binfo): ... here.
9457 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9459 2014-02-03 Teresa Johnson <tejohnson@google.com>
9461 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9462 redef of outer loop index variable.
9464 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9468 * doc/extend.texi (Function Attributes): Typo.
9470 2014-02-03 Cong Hou <congh@google.com>
9472 PR tree-optimization/60000
9473 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9474 if the vectorized statement is a store. A store statement can only
9475 appear at the end of pattern statements.
9477 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9479 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9480 (ix86_option_override_internal): Default long double to 64-bit for
9481 32-bit Bionic and to 128-bit for 64-bit Bionic.
9483 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9484 TARGET_LONG_DOUBLE_128 is true.
9485 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9487 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9488 (mlong-double-64): Negate -mlong-double-128.
9489 (mlong-double-128): New option.
9491 * config/i386/i386-c.c (ix86_target_macros): Define
9492 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9494 * doc/invoke.texi: Document -mlong-double-128.
9496 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9498 PR rtl-optimization/60024
9499 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9501 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9503 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9505 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9507 PR rtl-optimization/57662
9508 * sel-sched.c (code_motion_path_driver): Do not mark already not
9509 existing blocks in the visiting bitmap.
9511 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9513 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9514 on the insn being emitted.
9516 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9517 Will Deacon <will.deacon@arm.com>
9519 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9521 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9523 * config/arm/arm-tables.opt: Regenerate.
9525 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9527 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9528 for vector types other than V16QImode.
9529 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9530 define_expand, and call altivec_expand_vec_perm_le when producing
9531 code with little endian element order.
9532 (*altivec_vperm_<mode>_internal): New insn having previous
9533 behavior of altivec_vperm_<mode>.
9534 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9535 altivec_expand_vec_perm_le when producing code with little endian
9537 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9538 behavior of altivec_vperm_<mode>_uns.
9540 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9542 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9543 (altivec_vsumsws): Add handling for -maltivec=be with a little
9545 (altivec_vsumsws_direct): New.
9546 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9547 gen_altivec_vsumsws.
9549 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9551 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9552 vtable_pointer_value_to_binfo): New functions.
9553 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9554 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9556 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9558 * config/nios2/nios2.md (load_got_register): Initialize GOT
9559 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9560 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9562 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9564 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9565 preserverd by passthrough, do not propagate the type.
9567 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9569 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9570 (mips_atomic_assign_expand_fenv): New function.
9571 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9573 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9575 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9576 (__builtin_mips_set_fcsr): Likewise.
9577 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9578 MIPS_USI_FTYPE_VOID.
9579 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9580 (mips16_expand_set_fcsr): Likewise.
9581 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9582 (mips16_set_fcsr_stub): Likewise.
9583 (mips16_get_fcsr_one_only_stub): New class.
9584 (mips16_set_fcsr_one_only_stub): Likewise.
9585 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9586 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9587 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9588 (hard_float): New availability predicate.
9589 (mips_builtins): Add get_fcsr and set_fcsr.
9590 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9591 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9592 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9593 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9594 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9597 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9599 * config/mips/mips.c (mips_one_only_stub): New class.
9600 (mips_need_mips16_rdhwr_p): Replace with...
9601 (mips16_rdhwr_stub): ...this new variable.
9602 (mips16_stub_call_address): New function.
9603 (mips16_rdhwr_one_only_stub): New class.
9604 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9605 (mips_output_mips16_rdhwr): Delete.
9606 (mips_finish_stub): New function.
9607 (mips_code_end): Use it to handle rdhwr stubs.
9609 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9612 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9613 when calculating size of integer atomic types.
9615 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9617 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9619 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9621 PR tree-optimization/60003
9622 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9623 * profile.c (branch_prob): Use gimple_call_builtin_p
9624 to check for BUILT_IN_SETJMP_RECEIVER.
9625 * tree-inline.c (copy_bb): Call notice_special_calls.
9627 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9630 * lra-constraints.c (process_alt_operands): Update reload_sum only
9633 2014-01-31 Richard Henderson <rth@redhat.com>
9636 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9637 until after else_eh is processed.
9639 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9641 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9642 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9643 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9644 in smmintrin.h, remove them.
9645 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9646 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9647 * config/i386/i386.md (ROUND_SAE): Fix value.
9648 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9649 (const48_operand): New.
9650 * config/i386/subst.md (round), (round_expand): Use
9651 const_4_or_8_to_11_operand.
9652 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9654 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9656 * config/i386/constraints.md (Yk): Swap meaning with k.
9657 * config/i386/i386.md (movhi_internal): Change Yk to k.
9658 (movqi_internal): Ditto.
9659 (*k<logic><mode>): Ditto.
9662 (kandn<mode>): Ditto.
9663 (*<code>hi_1): Ditto.
9664 (*<code>qi_1): Ditto.
9665 (kxnor<mode>): Ditto.
9666 (kortestzhi): Ditto.
9667 (kortestchi): Ditto.
9669 (*one_cmplhi2_1): Ditto.
9670 (*one_cmplqi2_1): Ditto.
9671 * config/i386/sse.md (): Change k to Yk.
9672 (avx512f_load<mode>_mask): Ditto.
9673 (avx512f_blendm<mode>): Ditto.
9674 (avx512f_store<mode>_mask): Ditto.
9675 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9676 (avx512f_storedqu<mode>_mask): Ditto.
9677 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9679 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9680 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9681 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9682 (avx512f_maskcmp<mode>3): Ditto.
9683 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9684 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9685 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9686 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9687 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9688 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9689 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9690 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9691 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9692 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9693 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9694 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9695 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9696 (vec_extract_lo_<mode>_maskm): Ditto.
9697 (vec_extract_hi_<mode>_maskm): Ditto.
9698 (avx512f_vternlog<mode>_mask): Ditto.
9699 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9700 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9701 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9702 (avx512f_<code>v8div16qi2_mask): Ditto.
9703 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9704 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9705 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9706 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9707 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9708 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9709 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9710 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9711 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9712 (avx512cd_maskb_vec_dupv8di): Ditto.
9713 (avx512cd_maskw_vec_dupv16si): Ditto.
9714 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9715 (avx512f_vpermi2var<mode>3_mask): Ditto.
9716 (avx512f_vpermi2var<mode>3_mask): Ditto.
9717 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9718 (*avx512f_gathersi<mode>): Ditto.
9719 (*avx512f_gathersi<mode>_2): Ditto.
9720 (*avx512f_gatherdi<mode>): Ditto.
9721 (*avx512f_gatherdi<mode>_2): Ditto.
9722 (*avx512f_scattersi<mode>): Ditto.
9723 (*avx512f_scatterdi<mode>): Ditto.
9724 (avx512f_compress<mode>_mask): Ditto.
9725 (avx512f_compressstore<mode>_mask): Ditto.
9726 (avx512f_expand<mode>_mask): Ditto.
9727 * config/i386/subst.md (mask): Change k to Yk.
9728 (mask_scalar_merge): Ditto.
9731 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9733 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9735 2014-01-31 Richard Biener <rguenther@suse.de>
9738 * builtins.c (fold_builtin_memory_op): Make sure to not
9739 use a floating-point mode or a boolean or enumeral type for
9742 2014-01-30 DJ Delorie <dj@redhat.com>
9744 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9745 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9746 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9747 whenever main() has an epilogue.
9749 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9751 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9752 unused variable "field".
9753 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9754 (vsx_mergeh_<mode>): Likewise.
9755 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9756 (altivec_vmrghh): Likewise.
9757 (altivec_vmrghw): Likewise.
9758 (altivec_vmrglb): Likewise.
9759 (altivec_vmrglh): Likewise.
9760 (altivec_vmrglw): Likewise.
9761 (altivec_vspltb): Add missing uses.
9762 (altivec_vsplth): Likewise.
9763 (altivec_vspltw): Likewise.
9764 (altivec_vspltsf): Likewise.
9766 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9769 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9770 frame related instructions.
9772 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9774 PR rtl-optimization/59959
9775 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9776 any reload of register whose subreg is invalid.
9778 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9780 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9781 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9782 Add missing return type - void.
9784 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9786 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9787 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9788 remove element index adjustment for endian (now handled in vsx.md
9790 (altivec_expand_vec_perm_const): Use
9791 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9792 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9793 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9794 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9795 define_expand and a new define_insn *altivec_vspltb_internal;
9796 adjust for -maltivec=be on a little endian target.
9797 (altivec_vspltb_direct): New.
9798 (altivec_vsplth): Divide into a define_expand and a new
9799 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9800 little endian target.
9801 (altivec_vsplth_direct): New.
9802 (altivec_vspltw): Divide into a define_expand and a new
9803 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9804 little endian target.
9805 (altivec_vspltw_direct): New.
9806 (altivec_vspltsf): Divide into a define_expand and a new
9807 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9808 a little endian target.
9810 2014-01-30 Richard Biener <rguenther@suse.de>
9812 PR tree-optimization/59993
9813 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9814 can propagate form the earlier stmt and avoid the transform
9815 when the intermediate result is needed.
9817 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9819 * README.Portability: Fix typo.
9821 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9823 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9824 comparison_operator with ordered_comparison_operator.
9826 2014-01-30 Nick Clifton <nickc@redhat.com>
9828 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9829 Rename to mn10300_store_multiple_regs.
9830 * config/mn10300/mn10300.c: Likewise.
9831 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9832 store_multiple_regs.
9833 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9834 Call mn10300_store_multiple_regs.
9836 2014-01-30 Nick Clifton <nickc@redhat.com>
9837 DJ Delorie <dj@redhat.com>
9839 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9840 %fp 2 to keep registers after it properly word-aligned.
9841 (rl78_alloc_physical_registers_umul): Handle the case where both
9842 input operands are the same.
9844 2014-01-30 Richard Biener <rguenther@suse.de>
9846 PR tree-optimization/59903
9847 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9850 2014-01-30 Jason Merrill <jason@redhat.com>
9853 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9856 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9858 2014-01-30 Richard Biener <rguenther@suse.de>
9860 PR tree-optimization/59951
9861 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9863 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9866 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9867 SFmode to DFmode case.
9869 2014-01-29 DJ Delorie <dj@redhat.com>
9871 * config/msp430/msp430.opt (-minrt): New.
9872 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9874 (ENDFILE_SPEC): Likewise.
9876 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9878 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9879 (estimate_function_body_sizes): Use it.
9881 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9884 * dwarf2out.c (is_cxx_auto): New.
9885 (is_base_type): Use it.
9886 (gen_type_die_with_usage): Likewise.
9888 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9890 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9891 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9892 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9893 -maltivec=be with LE targets.
9894 (vsx_mergeh_<mode>): Likewise.
9895 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9896 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9897 (altivec_vmrghb): Replace with define_expand and new
9898 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9899 (altivec_vmrghb_direct): New define_insn.
9900 (altivec_vmrghh): Replace with define_expand and new
9901 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9902 (altivec_vmrghh_direct): New define_insn.
9903 (altivec_vmrghw): Replace with define_expand and new
9904 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9905 (altivec_vmrghw_direct): New define_insn.
9906 (*altivec_vmrghsf): Adjust for endianness.
9907 (altivec_vmrglb): Replace with define_expand and new
9908 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9909 (altivec_vmrglb_direct): New define_insn.
9910 (altivec_vmrglh): Replace with define_expand and new
9911 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9912 (altivec_vmrglh_direct): New define_insn.
9913 (altivec_vmrglw): Replace with define_expand and new
9914 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9915 (altivec_vmrglw_direct): New define_insn.
9916 (*altivec_vmrglsf): Adjust for endianness.
9917 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9918 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9919 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9920 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9921 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9922 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9923 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9924 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9926 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9928 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9929 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9932 2014-01-29 Richard Biener <rguenther@suse.de>
9934 PR tree-optimization/58742
9935 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9936 associate_pointerplus_align.
9937 (associate_pointerplus_diff): New function.
9938 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9939 and associate_pointerplus_diff.
9941 2014-01-29 Richard Biener <rguenther@suse.de>
9943 * lto-streamer.h (LTO_major_version): Bump to 3.
9944 (LTO_minor_version): Reset to 0.
9946 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9948 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9949 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9950 (arm_file_start): Generate correct asm header for armv7ve.
9951 * config/arm/bpabi.h: Add multilib support for armv7ve.
9952 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9953 and cortex-a15 to armv7ve.
9954 * config/arm/t-aprofile: Add multilib support for armv7ve.
9955 * doc/invoke.texi: Document -march=armv7ve.
9957 2014-01-29 Richard Biener <rguenther@suse.de>
9959 PR tree-optimization/58742
9960 * tree-ssa-forwprop.c (associate_plusminus): Return true
9961 if we changed sth, defer EH cleanup to ...
9962 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9963 (simplify_mult): New function.
9965 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9968 PR tree-optimization/59920
9969 * tree.c (build_common_builtin_nodes): Remove
9970 __builtin_setjmp_dispatcher initialization.
9971 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9972 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9973 instead of gsi_after_labels + manually skipping debug stmts.
9974 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9975 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9976 * tree-inline.c (copy_edges_for_bb): Remove
9977 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9978 argument. Ignore computed_goto_p stmts. Don't call
9979 make_abnormal_goto_edges. If a call might need abnormal edges
9980 for non-local gotos, see if it already has an edge to
9981 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9982 with true argument, don't do anything then, otherwise add
9983 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9984 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9986 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9987 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9988 (lower_stmt): Don't set data->calls_builtin_setjmp.
9989 (lower_builtin_setjmp): Adjust comment.
9990 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9991 * tree-cfg.c (found_computed_goto): Remove.
9992 (factor_computed_gotos): Remove.
9993 (make_goto_expr_edges): Return bool, true for computed gotos.
9994 Don't call make_abnormal_goto_edges.
9995 (build_gimple_cfg): Don't set found_computed_goto, don't call
9996 factor_computed_gotos.
9997 (computed_goto_p): No longer static.
9998 (make_blocks): Don't set found_computed_goto.
9999 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10000 (make_edges): If make_goto_expr_edges returns true, push bb
10001 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10002 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10003 vector. Record mapping between bbs and OpenMP regions if there
10004 are any, adjust make_gimple_omp_edges caller. Call
10005 handle_abnormal_edges.
10006 (make_abnormal_goto_edges): Remove.
10007 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10008 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10009 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10010 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10011 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10012 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10013 filling *region also set *region_idx to (*region)->entry->index.
10016 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10017 For REGs set ORIGINAL_REGNO.
10019 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10021 * doc/md.texi: Mention that a target shouldn't implement
10022 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10025 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10027 PR tree-optimization/59594
10028 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10029 a copy of the datarefs vector rather than the vector itself.
10031 2014-01-28 Jason Merrill <jason@redhat.com>
10034 * dwarf2out.c (auto_die): New static.
10035 (gen_type_die_with_usage): Handle C++1y 'auto'.
10036 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10039 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10042 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10043 (SPEC_X32): Likewise.
10044 (SPEC_64): Likewise.
10045 * config/i386/i386.c (ix86_option_override_internal): Turn off
10046 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10048 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10049 * config/i386/i386.h (TARGET_16BIT): New macro.
10050 (TARGET_16BIT_P): Likewise.
10051 * config/i386/i386.opt: Add m16.
10052 * doc/invoke.texi: Document -m16.
10054 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10056 PR preprocessor/59935
10057 * input.c (location_get_source_line): Bail out on when line number
10058 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10060 2014-01-28 Richard Biener <rguenther@suse.de>
10062 PR tree-optimization/58742
10063 * tree-ssa-forwprop.c (associate_plusminus): Handle
10064 pointer subtraction of the form (T)(P + A) - (T)P.
10066 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10068 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10071 2014-01-28 Richard Biener <rguenther@suse.de>
10074 2014-01-28 Richard Biener <rguenther@suse.de>
10076 PR rtl-optimization/45364
10077 PR rtl-optimization/59890
10078 * var-tracking.c (local_get_addr_clear_given_value): Handle
10079 already cleared slot.
10080 (val_reset): Handle not allocated local_get_addr_cache.
10081 (vt_find_locations): Use post-order on the inverted CFG.
10083 2014-01-28 Richard Biener <rguenther@suse.de>
10085 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10087 2014-01-28 Richard Biener <rguenther@suse.de>
10089 PR rtl-optimization/45364
10090 PR rtl-optimization/59890
10091 * var-tracking.c (local_get_addr_clear_given_value): Handle
10092 already cleared slot.
10093 (val_reset): Handle not allocated local_get_addr_cache.
10094 (vt_find_locations): Use post-order on the inverted CFG.
10096 2014-01-28 Alan Modra <amodra@gmail.com>
10098 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10099 * configure.ac <recursive call for build != host>: Define
10100 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10101 and LD_FOR_BUILD too.
10102 * configure: Regenerate.
10104 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10106 * config/i386/i386.c (get_builtin_code_for_version): Separate
10107 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10108 Broadwell from Haswell.
10110 2014-01-27 Steve Ellcey <sellcey@mips.com>
10112 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10113 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10114 * config/mips/mips.c (mips_option_override): Change setting
10116 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10117 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10118 Change from Mask to Var.
10120 2014-01-27 Jeff Law <law@redhat.com>
10122 * ipa-inline.c (inline_small_functions): Fix typo.
10124 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10126 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10127 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10128 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10129 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10130 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10131 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10132 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10133 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10134 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10135 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10136 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10137 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10138 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10139 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10140 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10141 (_mm512_storeu_epi64): Ditto.
10142 (_mm512_cmpge_epi32_mask): Ditto.
10143 (_mm512_cmpge_epu32_mask): Ditto.
10144 (_mm512_cmpge_epi64_mask): Ditto.
10145 (_mm512_cmpge_epu64_mask): Ditto.
10146 (_mm512_cmple_epi32_mask): Ditto.
10147 (_mm512_cmple_epu32_mask): Ditto.
10148 (_mm512_cmple_epi64_mask): Ditto.
10149 (_mm512_cmple_epu64_mask): Ditto.
10150 (_mm512_cmplt_epi32_mask): Ditto.
10151 (_mm512_cmplt_epu32_mask): Ditto.
10152 (_mm512_cmplt_epi64_mask): Ditto.
10153 (_mm512_cmplt_epu64_mask): Ditto.
10154 (_mm512_cmpneq_epi32_mask): Ditto.
10155 (_mm512_cmpneq_epu32_mask): Ditto.
10156 (_mm512_cmpneq_epi64_mask): Ditto.
10157 (_mm512_cmpneq_epu64_mask): Ditto.
10158 (_mm512_expand_pd): Ditto.
10159 (_mm512_expand_ps): Ditto.
10160 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10161 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10162 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10163 * config/i386/i386.c (ix86_builtins): Add
10164 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10165 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10166 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10167 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10168 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10169 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10170 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10171 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10172 IX86_BUILTIN_PMOVUSQW512_MEM.
10173 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10174 __builtin_ia32_pmovsqd512mem_mask,
10175 __builtin_ia32_pmovqd512mem_mask,
10176 __builtin_ia32_pmovusqw512mem_mask,
10177 __builtin_ia32_pmovsqw512mem_mask,
10178 __builtin_ia32_pmovqw512mem_mask,
10179 __builtin_ia32_pmovusdw512mem_mask,
10180 __builtin_ia32_pmovsdw512mem_mask,
10181 __builtin_ia32_pmovdw512mem_mask,
10182 __builtin_ia32_pmovqb512mem_mask,
10183 __builtin_ia32_pmovusqb512mem_mask,
10184 __builtin_ia32_pmovsqb512mem_mask,
10185 __builtin_ia32_pmovusdb512mem_mask,
10186 __builtin_ia32_pmovsdb512mem_mask,
10187 __builtin_ia32_pmovdb512mem_mask.
10188 (bdesc_args): Add __builtin_ia32_expanddf512,
10189 __builtin_ia32_expandsf512.
10190 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10191 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10192 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10193 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10194 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10195 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10196 (avx512f_<code>v8div16qi2_mask_store): This.
10197 (avx512f_expand<mode>): New.
10199 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10201 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10203 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10204 (_mm512_prefetch_i32scatter_pd): Ditto.
10205 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10206 (_mm512_prefetch_i64scatter_pd): Ditto.
10207 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10208 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10209 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10210 (_mm512_prefetch_i32scatter_ps): Ditto.
10211 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10212 (_mm512_prefetch_i64scatter_ps): Ditto.
10213 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10214 * config/i386/i386-builtin-types.def: Define
10215 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10216 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10217 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10218 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10219 IX86_BUILTIN_SCATTERPFQPD.
10220 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10221 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10222 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10223 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10224 __builtin_ia32_scatterpfqps.
10225 (ix86_expand_builtin): Expand new built-ins.
10226 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10227 fix memory access data type.
10228 (*avx512pf_gatherpf<mode>_mask): Ditto.
10229 (*avx512pf_gatherpf<mode>): Ditto.
10230 (avx512pf_scatterpf<mode>): Ditto.
10231 (*avx512pf_scatterpf<mode>_mask): Ditto.
10232 (*avx512pf_scatterpf<mode>): Ditto.
10233 (GATHER_SCATTER_SF_MEM_MODE): New.
10234 (avx512pf_gatherpf<mode>df): Ditto.
10235 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10236 (*avx512pf_scatterpf<mode>df): Ditto.
10238 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10241 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10242 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10245 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10247 * common/config/arm/arm-common.c
10248 (arm_rewrite_mcpu): Handle multiple names.
10250 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10252 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10254 * gimple-builder.h (create_gimple_tmp): Delete.
10256 2014-01-27 Christian Bruel <christian.bruel@st.com>
10258 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10261 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10263 * config/pa/pa.md (call): Generate indirect long calls to non-local
10264 functions when outputing 32-bit code.
10265 (call_value): Likewise except for special call to buggy powf function.
10267 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10268 portable runtime and PIC indirect calls.
10269 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10270 and PIC call sequences. Use ldo instead of blr to set return register
10271 in PIC call sequence.
10273 2014-01-25 Walter Lee <walt@tilera.com>
10275 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10276 avoid clobbering a live register.
10278 2014-01-25 Walter Lee <walt@tilera.com>
10280 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10281 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10282 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10283 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10285 2014-01-25 Walter Lee <walt@tilera.com>
10287 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10288 arguments on even registers.
10289 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10291 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10292 (BIGGEST_ALIGNMENT): Ditto.
10293 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10295 2014-01-25 Walter Lee <walt@tilera.com>
10297 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10298 insns before bundling.
10299 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10301 2014-01-25 Walter Lee <walt@tilera.com>
10303 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10304 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10305 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10307 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10309 * config/mips/constraints.md (kl): Delete.
10310 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10311 define expands, using...
10312 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10313 instructions for MIPS16.
10314 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10315 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10317 2014-01-25 Walter Lee <walt@tilera.com>
10319 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10323 2014-01-25 Walter Lee <walt@tilera.com>
10325 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10326 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10328 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10330 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10333 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10335 * print-rtl.c (in_call_function_usage): New var.
10336 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10337 EXPR_LIST mode as mode and not as reg note name.
10339 PR middle-end/59561
10340 * cfgloopmanip.c (copy_loop_info): If
10341 loop->warned_aggressive_loop_optimizations, make sure
10342 the flag is set in target loop too.
10344 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10346 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10348 * builtins.def: Likewise.
10349 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10350 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10351 * ira.c (ira_setup_eliminable_regset): Likewise.
10352 * omp-low.c (gate_expand_omp): Likewise.
10353 (execute_lower_omp): Likewise.
10354 (diagnose_sb_0): Likewise.
10355 (gate_diagnose_omp_blocks): Likewise.
10356 (simd_clone_clauses_extract): Likewise.
10359 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10361 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10362 correction for little endian...
10363 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10366 2014-01-24 Jeff Law <law@redhat.com>
10368 PR tree-optimization/59919
10369 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10370 for non-returning calls.
10372 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10374 * common/config/aarch64/aarch64-common.c
10375 (aarch64_rewrite_mcpu): Handle multiple names.
10376 * config/aarch64/aarch64.h
10377 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10379 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10381 * input.c (add_file_to_cache_tab): Handle the case where fopen
10384 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10387 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10388 from push operand if code of push isn't PRE_DEC.
10390 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10393 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10394 -mquad-memory-atomic. Update -mquad-memory documentation to say
10395 it is only used for non-atomic loads/stores.
10397 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10398 -mquad-memory or -mquad-memory-atomic switches.
10400 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10401 -mquad-memory-atomic to ISA 2.07 support.
10403 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10404 to separate support of normal quad word memory operations (ldq, stq)
10405 from the atomic quad word memory operations.
10407 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10408 support to separate non-atomic quad word operations from atomic
10409 quad word operations. Disable non-atomic quad word operations in
10410 little endian mode so that we don't have to swap words after the
10411 load and before the store.
10412 (quad_load_store_p): Add comment about atomic quad word support.
10413 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10414 options printed with -mdebug=reg.
10416 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10417 -mquad-memory-atomic as the test for whether we have quad word
10418 atomic instructions.
10419 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10420 or -mp8-vector are used, allow byte/half-word atomic operations.
10422 * config/rs6000/sync.md (load_lockedti): Insure that the address
10423 is a proper indexed or indirect address for the lqarx instruction.
10424 On little endian systems, swap the hi/lo registers after the lqarx
10426 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10427 insure the address is valid for the lqarx instruction.
10428 (store_conditionalti): Insure that the address is a proper indexed
10429 or indirect address for the stqcrx. instruction. On little endian
10430 systems, swap the hi/lo registers before doing the stqcrx.
10432 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10433 insure the address is valid for the stqcrx. instruction.
10435 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10436 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10437 type of quad memory support is available.
10439 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10441 PR regression/59915
10442 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10443 there is a danger of looping.
10445 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10447 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10448 force flag_ira_loop_pressure if set via command line.
10450 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10452 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10453 (ashr_simd): New builtin handling DI mode.
10454 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10455 (aarch64_sshr_simddi): New match pattern.
10456 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10457 (vshrd_n_s64): Likewise.
10458 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10460 2014-01-23 Nick Clifton <nickc@redhat.com>
10462 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10463 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10464 favour of mcu specific scripts.
10465 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10468 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10469 Alex Velenko <Alex.Velenko@arm.com>
10471 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10472 (vaddv_s16): Likewise.
10473 (vaddv_s32): Likewise.
10474 (vaddv_u8): Likewise.
10475 (vaddv_u16): Likewise.
10476 (vaddv_u32): Likewise.
10477 (vaddvq_s8): Likewise.
10478 (vaddvq_s16): Likewise.
10479 (vaddvq_s32): Likewise.
10480 (vaddvq_s64): Likewise.
10481 (vaddvq_u8): Likewise.
10482 (vaddvq_u16): Likewise.
10483 (vaddvq_u32): Likewise.
10484 (vaddvq_u64): Likewise.
10485 (vaddv_f32): Likewise.
10486 (vaddvq_f32): Likewise.
10487 (vaddvq_f64): Likewise.
10488 (vmaxv_f32): Likewise.
10489 (vmaxv_s8): Likewise.
10490 (vmaxv_s16): Likewise.
10491 (vmaxv_s32): Likewise.
10492 (vmaxv_u8): Likewise.
10493 (vmaxv_u16): Likewise.
10494 (vmaxv_u32): Likewise.
10495 (vmaxvq_f32): Likewise.
10496 (vmaxvq_f64): Likewise.
10497 (vmaxvq_s8): Likewise.
10498 (vmaxvq_s16): Likewise.
10499 (vmaxvq_s32): Likewise.
10500 (vmaxvq_u8): Likewise.
10501 (vmaxvq_u16): Likewise.
10502 (vmaxvq_u32): Likewise.
10503 (vmaxnmv_f32): Likewise.
10504 (vmaxnmvq_f32): Likewise.
10505 (vmaxnmvq_f64): Likewise.
10506 (vminv_f32): Likewise.
10507 (vminv_s8): Likewise.
10508 (vminv_s16): Likewise.
10509 (vminv_s32): Likewise.
10510 (vminv_u8): Likewise.
10511 (vminv_u16): Likewise.
10512 (vminv_u32): Likewise.
10513 (vminvq_f32): Likewise.
10514 (vminvq_f64): Likewise.
10515 (vminvq_s8): Likewise.
10516 (vminvq_s16): Likewise.
10517 (vminvq_s32): Likewise.
10518 (vminvq_u8): Likewise.
10519 (vminvq_u16): Likewise.
10520 (vminvq_u32): Likewise.
10521 (vminnmv_f32): Likewise.
10522 (vminnmvq_f32): Likewise.
10523 (vminnmvq_f64): Likewise.
10525 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10527 * config/aarch64/aarch64-simd.md
10528 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10529 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10530 (*aarch64_mul3_elt<mode>): Likewise.
10531 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10532 (*aarch64_mul3_elt_to_64v2df): Likewise.
10533 (*aarch64_mla_elt<mode>): Likewise.
10534 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10535 (*aarch64_mls_elt<mode>): Likewise.
10536 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10537 (*aarch64_fma4_elt<mode>): Likewise.
10538 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10539 (*aarch64_fma4_elt_to_64v2df): Likewise.
10540 (*aarch64_fnma4_elt<mode>): Likewise.
10541 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10542 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10543 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10544 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10545 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10546 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10547 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10548 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10549 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10551 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10553 * config/aarch64/aarch64-simd.md
10554 (aarch64_be_checked_get_lane<mode>): New define_expand.
10555 * config/aarch64/aarch64-simd-builtins.def
10556 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10557 New builtin definition.
10558 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10559 Use new safe be builtin.
10561 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10563 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10565 (aarch64_be_st1<mode>): Likewise.
10566 (aarch_ld1<VALL:mode>): Define_expand modified.
10567 (aarch_st1<VALL:mode>): Likewise.
10568 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10569 (UNSPEC_ST1): Likewise.
10571 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10573 * config/microblaze/microblaze.md: Add trap insn and attribute
10575 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10577 PR preprocessor/58580
10578 * input.h (location_get_source_line): Take an additional line_size
10580 (void diagnostics_file_cache_fini): Declare new function.
10581 * input.c (struct fcache): New type.
10582 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10583 New static constants.
10584 (diagnostic_file_cache_init, total_lines_num)
10585 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10586 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10587 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10588 (get_next_line, read_next_line, goto_next_line, read_line_num):
10589 New static function definitions.
10590 (diagnostic_file_cache_fini): New function.
10591 (location_get_source_line): Take an additional output line_len
10592 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10594 * diagnostic.c (diagnostic_finish): Call
10595 diagnostic_file_cache_fini.
10596 (adjust_line): Take an additional input parameter for the length
10597 of the line, rather than calculating it with strlen.
10598 (diagnostic_show_locus): Adjust the use of
10599 location_get_source_line and adjust_line with respect to their new
10600 signature. While displaying a line now, do not stop at the first
10601 null byte. Rather, display the zero byte as a space and keep
10602 going until we reach the size of the line.
10603 * Makefile.in: Add vec.o to OBJS-libcommon
10605 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10606 Ilya Tocar <ilya.tocar@intel.com>
10608 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10609 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10610 (__builtin_ia32_kmov16): Ditto.
10611 * config/i386/i386.md (UNSPEC_KMOV): New.
10614 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10616 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10617 (_mm512_storeu_si512): Ditto.
10619 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10622 * rtl.h (get_referenced_operands): Declare.
10623 * recog.c (get_referenced_operands): New function.
10624 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10625 operands have been referenced when recording LO_SUM references.
10627 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10629 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10631 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10633 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10634 Enable for generic and recent AMD targets.
10636 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10638 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10639 ARG_SIZE note when adjustment was eliminated.
10641 2014-01-22 Jeff Law <law@redhat.com>
10643 PR tree-optimization/59597
10644 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10645 in file. Accept new argument REGISTERING and use it to modify
10646 dump output appropriately.
10647 (register_jump_thread): Corresponding changes.
10648 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10649 thread paths involving joiner blocks. Add code to dump cancelled
10650 jump threading paths.
10652 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10654 PR rtl-optimization/59477
10655 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10656 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10658 2014-01-22 Tom Tromey <tromey@redhat.com>
10660 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10662 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10664 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10666 PR rtl-optimization/59896
10667 * lra-constraints.c (process_alt_operands): Check unused note for
10668 matched operands of insn with no output reloads.
10670 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10672 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10673 (mips_move_from_gpr_cost): Likewise.
10675 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10677 PR rtl-optimization/59858
10678 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10679 ira_class_hard_regs_num.
10680 (process_alt_operands): Increase reject for dying matched operand.
10682 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10685 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10686 smaller than size, perform several stores or loads and stores
10687 at dst + count - size to store or copy all of size bytes, rather
10688 than just last modesize bytes.
10690 2014-01-20 DJ Delorie <dj@redhat.com>
10692 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10693 that CLOBBERs are REGs before propogating their values.
10695 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10697 PR middle-end/59789
10698 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10699 (cgraph_inline_failed_type): New function.
10700 * cgraph.h (DEFCIFCODE): Add type.
10701 (cgraph_inline_failed_type_t): New enum.
10702 (cgraph_inline_failed_type): New prototype.
10703 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10704 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10705 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10706 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10707 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10708 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10709 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10710 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10711 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10712 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10713 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10714 OPTIMIZATION_MISMATCH.
10715 * tree-inline.c (expand_call_inline): Emit errors during
10716 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10718 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10721 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10722 mode attribute in insn output.
10724 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10726 * output.h (output_constant): Delete.
10727 * varasm.c (output_constant): Make private.
10729 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10731 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10733 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10735 PR middle-end/59860
10736 * tree.h (fold_builtin_strcat): New prototype.
10737 * builtins.c (fold_builtin_strcat): No longer static. Add len
10738 argument, if non-NULL, don't call c_strlen. Optimize
10739 directly into __builtin_memcpy instead of __builtin_strcpy.
10740 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10741 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10743 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10745 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10746 for SImode_address_operand operands, having only a REG argument.
10748 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10750 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10751 loader name using mbig-endian.
10752 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10754 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10756 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10757 (-mtune): Likewise.
10760 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10762 * config/aarch64/aarch64-protos.h
10763 (aarch64_cannot_change_mode_class_ptr): Declare.
10764 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10765 aarch64_cannot_change_mode_class_ptr): New.
10766 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10767 backend hook aarch64_cannot_change_mode_class.
10769 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10771 * common/config/aarch64/aarch64-common.c
10772 (aarch64_handle_option): Don't handle any option order logic here.
10773 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10774 selected_cpu, warn on architecture version mismatch.
10775 (aarch64_override_options): Fix parsing order for option strings.
10777 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10778 Iain Sandoe <iain@codesourcery.com>
10781 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10782 warning. Amend comment to reflect current functionality.
10784 2014-01-20 Richard Biener <rguenther@suse.de>
10786 PR middle-end/59860
10787 * builtins.c (fold_builtin_strcat): Remove case better handled
10788 by tree-ssa-strlen.c.
10790 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10792 * config/aarch64/aarch64.opt
10793 (mcpu, march, mtune): Make case-insensitive.
10795 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10798 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10799 if operands[1] is a REG or ZERO_EXTEND of a REG.
10801 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10803 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10805 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10807 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10808 long non-pic millicode calls.
10810 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10812 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10814 2014-01-19 Kito Cheng <kito@0xlab.org>
10816 * builtins.c (expand_movstr): Check movstr expand done or fail.
10818 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10819 H.J. Lu <hongjiu.lu@intel.com>
10822 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10823 to DImode for zero-extended addresses.
10825 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10827 PR rtl-optimization/57763
10828 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10829 on the new indirect jump_insn and increment LABEL_NUSES (label).
10831 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10835 * config.gcc (x86_archs): New variable.
10836 (x86_64_archs): Likewise.
10837 (x86_cpus): Likewise.
10838 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10839 --with-arch/--with-cpu= options.
10840 Support --with-arch=/--with-cpu={nehalem,westmere,
10841 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10843 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10845 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10846 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10848 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10850 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10852 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10855 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10856 clear cpp_get_options (parse_in)->warn_unused_macros for
10857 ix86_target_macros_internal with cpp_define.
10859 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10861 * jump.c (delete_related_insns): Keep (use (insn))s.
10862 * reorg.c (redundant_insn): Check for barriers too.
10864 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10866 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10868 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10870 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10871 call to $$dyncall when TARGET_LONG_CALLS is true.
10873 2014-01-17 Jeff Law <law@redhat.com>
10875 * ree.c (combine_set_extension): Temporarily disable test for
10876 changing number of hard registers.
10878 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10880 PR middle-end/58125
10881 * ipa-inline-analysis.c (inline_free_summary):
10882 Do not free summary of aliases.
10884 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10886 PR middle-end/59706
10887 * gimplify.c (gimplify_expr): Use create_tmp_var
10888 instead of create_tmp_var_raw. If cond doesn't have
10889 integral type, don't add the IFN_ANNOTATE builtin at all.
10891 2014-01-17 Martin Jambor <mjambor@suse.cz>
10894 * ipa-cp.c (prev_edge_clone): New variable.
10895 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10896 Also resize prev_edge_clone vector.
10897 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10898 (ipcp_edge_removal_hook): New function.
10899 (ipcp_driver): Register ipcp_edge_removal_hook.
10901 2014-01-17 Andrew Pinski <apinski@cavium.com>
10902 Steve Ellcey <sellcey@mips.com>
10905 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10908 2014-01-17 Jeff Law <law@redhat.com>
10910 PR middle-end/57904
10911 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10912 so that pass_ccp runs first.
10914 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10916 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10917 (ix86_adjust_cost): Use !TARGET_XXX.
10918 (do_reorder_for_imul): Likewise.
10919 (swap_top_of_ready_list): Likewise.
10920 (ix86_sched_reorder): Likewise.
10922 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10924 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10925 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10926 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10927 (intel_memset): New. Duplicate slm_memset.
10928 (intel_cost): New. Duplicate slm_cost.
10929 (m_INTEL): New macro.
10930 (processor_target_table): Add "intel".
10931 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10932 with PROCESSOR_INTEL for "intel".
10933 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10934 PROCESSOR_SILVERMONT.
10935 (ix86_issue_rate): Likewise.
10936 (ix86_adjust_cost): Likewise.
10937 (ia32_multipass_dfa_lookahead): Likewise.
10938 (swap_top_of_ready_list): Likewise.
10939 (ix86_sched_reorder): Likewise.
10940 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10941 instead of TARGET_OPT_AGU.
10942 * config/i386/i386.h (TARGET_INTEL): New.
10943 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10944 (processor_type): Add PROCESSOR_INTEL.
10945 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10946 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10948 2014-01-17 Marek Polacek <polacek@redhat.com>
10951 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10954 2014-01-17 Richard Biener <rguenther@suse.de>
10956 PR tree-optimization/46590
10957 * opts.c (default_options_table): Add entries for
10958 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10959 all enabled at -O1 but not for -Og.
10960 * common.opt (fbranch-count-reg): Remove Init(1).
10961 (fmove-loop-invariants): Likewise.
10962 (ftree-pta): Likewise.
10964 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10966 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10967 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10968 decls to at least the GCC 4.8 used alignments.
10971 * tree-nested.c (convert_nonlocal_reference_stmt,
10972 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10973 of GIMPLE_BIND stmts, adjust associated decls.
10975 2014-01-17 Richard Biener <rguenther@suse.de>
10977 PR tree-optimization/46590
10978 * vec.h (vec<>::bseach): New member function implementing
10979 binary search according to C89 bsearch.
10980 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10981 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10982 bitmap pointer again. Make accesses_in_loop a flat array.
10983 (mem_ref_obstack): New global.
10984 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10985 (mark_ref_stored): Likewise.
10986 (ref_indep_loop_p_2): Likewise.
10987 (set_ref_stored_in_loop): New helper function.
10988 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10989 (memref_free): Adjust.
10990 (record_mem_ref_loc): Simplify.
10991 (gather_mem_refs_stmt): Adjust.
10992 (sort_locs_in_loop_postorder_cmp): New function.
10993 (analyze_memory_references): Sort accesses_in_loop after
10994 loop postorder number.
10995 (find_ref_loc_in_loop_cmp): New function.
10996 (for_all_locs_in_loop): Find relevant cluster of locs in
10997 accesses_in_loop and iterate without recursion.
10998 (execute_sm): Avoid uninit warning.
10999 (struct ref_always_accessed): Simplify.
11000 (ref_always_accessed::operator ()): Likewise.
11001 (ref_always_accessed_p): Likewise.
11002 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11003 loop postorder numbers here.
11004 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11007 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11010 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11011 on decls for which assemble_alias has been called.
11013 2014-01-17 Nick Clifton <nickc@redhat.com>
11015 * config/msp430/msp430.opt: (mcpu): New option.
11016 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11017 (msp430_option_override): Parse target_cpu. If the MCU name
11018 matches a generic string, clear target_mcu.
11019 (msp430_attr): Allow numeric interrupt values up to 63.
11020 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11021 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11023 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11025 * config/msp430/msp430.md (popm): Use %J rather than %I.
11026 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11027 (addhi_cy_i): Use immediate_operand for operand 2.
11028 * doc/invoke.texi: Document -mcpu option.
11030 2014-01-17 Richard Biener <rguenther@suse.de>
11032 PR rtl-optimization/38518
11033 * df.h (df_analyze_loop): Declare.
11034 * df-core.c: Include cfgloop.h.
11035 (df_analyze_1): Split out main part of df_analyze.
11036 (df_analyze): Adjust.
11037 (loop_inverted_post_order_compute): New function.
11038 (loop_post_order_compute): Likewise.
11039 (df_analyze_loop): New function avoiding whole-function
11040 postorder computes.
11041 * loop-invariant.c (find_defs): Use df_analyze_loop.
11042 (find_invariants): Adjust.
11043 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11045 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11047 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11048 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11050 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11052 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11053 traversal when removing references.
11055 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11058 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11060 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11062 PR middle-end/56791
11063 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11064 pushing a reload for an autoinc when we had previously reloaded an
11065 inner part of the address.
11067 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11069 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11071 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11072 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11073 when not giving up or versioning for alias only because of
11075 (vect_analyze_data_ref_dependences): Set to true.
11076 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11078 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11079 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11082 PR middle-end/58344
11083 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11086 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11087 operand 0 predicate for gathers, use a new pseudo as subtarget.
11089 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11091 PR middle-end/59609
11092 * lra-constraints.c (process_alt_operands): Add printing debug info.
11093 Check absence of input/output reloads for matched operands too.
11095 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11097 PR rtl-optimization/59835
11098 * ira.c (ira_init_register_move_cost): Increase cost for
11101 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11103 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11105 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11108 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11109 non-register objects. Use gen_(high/low)part more consistently.
11112 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11115 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11116 endian support, remove tests for WORDS_BIG_ENDIAN.
11117 (p8_mfvsrd_3_<mode>): Likewise.
11118 (reload_gpr_from_vsx<mode>): Likewise.
11119 (reload_gpr_from_vsxsf): Likewise.
11120 (p8_mfvsrd_4_disf): Likewise.
11122 2014-01-16 Richard Biener <rguenther@suse.de>
11124 PR rtl-optimization/46590
11125 * lcm.c (compute_antinout_edge): Use postorder iteration.
11126 (compute_laterin): Use inverted postorder iteration.
11128 2014-01-16 Nick Clifton <nickc@redhat.com>
11130 PR middle-end/28865
11131 * varasm.c (output_constant): Return the number of bytes actually
11133 (output_constructor_array_range): Update the field size with the
11134 number of bytes emitted by output_constant.
11135 (output_constructor_regular_field): Likewise. Also do not
11136 complain if the total number of bytes emitted is now greater
11137 than the expected fieldpos.
11138 * output.h (output_constant): Update prototype and descriptive comment.
11140 2014-01-16 Marek Polacek <polacek@redhat.com>
11142 PR middle-end/59827
11143 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11144 it is error_mark_node.
11146 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11148 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11149 VALID_AVX256_REG_OR_OI_MODE.
11151 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11153 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11154 current procedure should be profiled.
11156 2014-01-15 Andrew Pinski <apinski@cavium.com>
11158 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11159 of moving from/to the STACK_REG register class.
11161 2014-01-15 Richard Henderson <rth@redhat.com>
11164 * reginfo.c (global_regs_decl): Globalize.
11165 * rtl.h (global_regs_decl): Declare.
11166 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11167 reserved via global_regs.
11169 2014-01-15 Teresa Johnson <tejohnson@google.com>
11171 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11173 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11175 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11176 and vmulosh rather than call gen_vec_widen_smult_*.
11177 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11178 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11179 (vec_widen_smult_even_v16qi): Likewise.
11180 (vec_widen_umult_even_v8hi): Likewise.
11181 (vec_widen_smult_even_v8hi): Likewise.
11182 (vec_widen_umult_odd_v16qi): Likewise.
11183 (vec_widen_smult_odd_v16qi): Likewise.
11184 (vec_widen_umult_odd_v8hi): Likewise.
11185 (vec_widen_smult_odd_v8hi): Likewise.
11186 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11187 vmuloub rather than call gen_vec_widen_umult_*.
11188 (vec_widen_umult_lo_v16qi): Likewise.
11189 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11190 vmulosb rather than call gen_vec_widen_smult_*.
11191 (vec_widen_smult_lo_v16qi): Likewise.
11192 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11193 rather than call gen_vec_widen_umult_*.
11194 (vec_widen_umult_lo_v8hi): Likewise.
11195 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11196 rather than call gen_vec_widen_smult_*.
11197 (vec_widen_smult_lo_v8hi): Likewise.
11199 2014-01-15 Jeff Law <law@redhat.com>
11201 PR tree-optimization/59747
11202 * ree.c (find_and_remove_re): Properly handle case where a second
11203 eliminated extension requires widening a copy created for elimination
11204 of a prior extension.
11205 (combine_set_extension): Ensure that the number of hard regs needed
11206 for a destination register does not change when we widen it.
11208 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11210 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11211 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11212 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11213 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11214 (avr-*-rtems*): Likewise.
11215 (bfin*-rtems*): Likewise.
11216 (moxie-*-rtems*): Likewise.
11217 (h8300-*-rtems*): Likewise.
11218 (i[34567]86-*-rtems*): Likewise.
11219 (lm32-*-rtems*): Likewise.
11220 (m32r-*-rtems*): Likewise.
11221 (m68k-*-rtems*): Likewise.
11222 (microblaze*-*-rtems*): Likewise.
11223 (mips*-*-rtems*): Likewise.
11224 (powerpc-*-rtems*): Likewise.
11225 (sh-*-rtems*): Likewise.
11226 (sparc-*-rtems*): Likewise.
11227 (sparc64-*-rtems*): Likewise.
11228 (v850-*-rtems*): Likewise.
11229 (m32c-*-rtems*): Likewise.
11231 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11233 PR rtl-optimization/59511
11234 * ira.c (ira_init_register_move_cost): Use memory costs for some
11235 cases of register move cost calculations.
11236 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11237 instead of BB frequency.
11238 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11239 * lra-assigns.c (find_hard_regno_for): Ditto.
11241 2014-01-15 Richard Biener <rguenther@suse.de>
11243 PR tree-optimization/59822
11244 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11245 (vectorizable_load): Use it to hoist defs of uses of invariant
11246 loads out of the loop.
11248 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11249 Kugan Vivekanandarajah <kuganv@linaro.org>
11252 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11255 2014-01-15 Richard Biener <rguenther@suse.de>
11257 PR rtl-optimization/59802
11258 * lcm.c (compute_available): Use inverted postorder to seed
11259 the initial worklist.
11261 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11264 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11265 ADDR_REGS for invalid symrefs in non-PIC code.
11267 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11270 * builtins.c (determine_block_size): Initialize *probable_max_size
11271 even if len_rtx is CONST_INT.
11273 2014-01-14 Andrew Pinski <apinski@cavium.com>
11275 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11276 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11277 (cortexa53_tunings): Likewise.
11278 (aarch64_sched_issue_rate): New function.
11279 (TARGET_SCHED_ISSUE_RATE): Define.
11281 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11283 * ira-costs.c (find_costs_and_classes): Add missed
11284 ira_init_register_move_cost_if_necessary.
11286 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11289 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11291 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11294 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11295 to indicate if type is used for function return value. Warn ABI
11296 change if the vector mode isn't available for function return value.
11297 (ix86_function_arg_advance): Pass false to type_natural_mode.
11298 (ix86_function_arg): Likewise.
11299 (ix86_gimplify_va_arg): Likewise.
11300 (function_arg_32): Don't warn ABI change.
11301 (ix86_function_value): Pass true to type_natural_mode.
11302 (ix86_return_in_memory): Likewise.
11303 (ix86_struct_value_rtx): Removed.
11304 (TARGET_STRUCT_VALUE_RTX): Likewise.
11306 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11308 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11309 converting a conditional jump into a conditional return.
11311 2014-01-14 Richard Biener <rguenther@suse.de>
11313 PR tree-optimization/58921
11314 PR tree-optimization/59006
11315 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11316 hoisting invariant stmts.
11317 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11318 invariant loads on the preheader edge if possible.
11320 2014-01-14 Joey Ye <joey.ye@arm.com>
11322 * doc/plugin.texi (Building GCC plugins): Update to C++.
11324 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11326 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11327 (_mm_rcp28_round_ss): Ditto.
11328 (_mm_rsqrt28_round_sd): Ditto.
11329 (_mm_rsqrt28_round_ss): Ditto.
11330 (_mm_rcp28_sd): Ditto.
11331 (_mm_rcp28_ss): Ditto.
11332 (_mm_rsqrt28_sd): Ditto.
11333 (_mm_rsqrt28_ss): Ditto.
11334 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11335 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11336 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11337 (IX86_BUILTIN_RCP28SD): Ditto.
11338 (IX86_BUILTIN_RCP28SS): Ditto.
11339 (IX86_BUILTIN_RSQRT28SD): Ditto.
11340 (IX86_BUILTIN_RSQRT28SS): Ditto.
11341 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11342 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11343 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11344 (ix86_expand_special_args_builtin): Expand new FTYPE.
11345 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11346 (srcp14<mode>): Make insn unary.
11347 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11348 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11349 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11350 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11351 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11352 Fix rounding: make it SAE only.
11353 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11355 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11357 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11358 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11359 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11360 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11361 (round_saeonly_mask_scalar_operand4): Ditto.
11362 (round_saeonly_mask_scalar_op3): Ditto.
11363 (round_saeonly_mask_scalar_op4): Ditto.
11365 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11367 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11368 Implement -maltivec=be for vec_insert and vec_extract.
11370 2014-01-10 DJ Delorie <dj@redhat.com>
11372 * config/msp430/msp430.md (call_internal): Don't allow memory
11373 references with SP as the base register.
11374 (call_value_internal): Likewise.
11375 * config/msp430/constraints.md (Yc): New. For memory references
11376 that don't use SP as a base register.
11378 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11379 "an integer without a # prefix"
11380 * config/msp430/msp430.md (epilogue_helper): Use it.
11382 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11385 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11386 AVX512F gather builtins.
11387 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11388 on gather decls with INTEGER_TYPE masktype.
11389 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11390 directly into the builtin rather than hoisting it before loop.
11392 PR tree-optimization/59387
11393 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11394 (scev_const_prop): If folded_casts and type has undefined overflow,
11395 use force_gimple_operand instead of force_gimple_operand_gsi and
11396 for each added stmt if it is assign with
11397 arith_code_with_undefined_signed_overflow, call
11398 rewrite_to_defined_overflow.
11399 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11400 gimple-fold.h instead.
11401 (arith_code_with_undefined_signed_overflow,
11402 rewrite_to_defined_overflow): Moved to ...
11403 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11404 rewrite_to_defined_overflow): ... here. No longer static.
11405 Include gimplify-me.h.
11406 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11407 rewrite_to_defined_overflow): New prototypes.
11409 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11411 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11413 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11415 * builtins.c (get_object_alignment_2): Minor tweak.
11416 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11418 2014-01-13 Christian Bruel <christian.bruel@st.com>
11420 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11421 optimized non constant lengths.
11423 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11426 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11427 load as __atomic_load_N if possible.
11429 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11431 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11433 (rs6000_expand_builtin): Adjust call.
11435 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11438 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11439 * config/rs6000/rs6000.c: Include target-globals.h.
11440 (rs6000_set_current_function): Instead of doing target_reinit
11441 unconditionally, use save_target_globals_default_opts and
11442 restore_target_globals.
11444 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11446 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11447 (rs6000_expand_builtin): Handle mffs and mtfsf.
11448 (rs6000_init_builtins): Define mffs and mtfsf.
11449 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11450 (rs6000_mffs): New pattern.
11451 (rs6000_mtfsf): New pattern.
11453 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11455 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11456 Start narrowing with START. Apply candidate-use pair
11457 and check overall cost in narrowing.
11458 (iv_ca_prune): Pass new argument.
11460 2014-01-10 Jeff Law <law@redhat.com>
11462 PR middle-end/59743
11463 * ree.c (combine_reaching_defs): Ensure the defining statement
11464 occurs before the extension when optimizing extensions with
11465 different source and destination hard registers.
11467 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11470 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11471 vtables into the type inheritance graph.
11473 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11475 PR rtl-optimization/59754
11476 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11477 modes in the REGNO != REGNO case.
11479 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11481 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11483 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11485 PR tree-optimization/59745
11486 * tree-predcom.c (tree_predictive_commoning_loop): Call
11487 free_affine_expand_cache if giving up because components is NULL.
11489 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11490 GC in payload of target_globals struct instead of allocating them on
11491 the heap and the larger structs separately using GC.
11492 * target-globals.h (struct target_globals): Make regs, hard_regs,
11493 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11494 of GTY((skip)) and change type to void *.
11495 (reset_target_globals): Cast loads from those fields to corresponding
11498 2014-01-10 Steve Ellcey <sellcey@mips.com>
11501 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11502 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11503 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11505 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11508 * aarch64-modes.def (CC_Zmode): New flags mode.
11509 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11510 represents an equality.
11511 (aarch64_get_condition_code): Handle CC_Zmode.
11512 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11514 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11516 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11517 extraction in good case.
11519 2014-01-10 Richard Biener <rguenther@suse.de>
11521 PR tree-optimization/59374
11522 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11523 checking after SLP discovery. Mark stmts not participating
11524 in any SLP instance properly.
11526 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11528 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11529 when handling a SET rtx.
11531 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11533 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11534 (cortex-a57): Likewise.
11535 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11537 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11539 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11540 non-iwmmxt builtins.
11542 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11546 * ipa-devirt.c record_target_from_binfo): Take as argument
11547 stack of binfos and lookup matching one for virtual inheritance.
11548 (possible_polymorphic_call_targets_1): Update.
11550 2014-01-10 Huacai Chen <chenhc@lemote.com>
11552 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11553 kernel strings for Loongson-2E/2F/3A.
11555 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11557 PR middle-end/59670
11558 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11559 is_gimple_call before calling gimple_call_internal_p.
11561 2014-01-09 Steve Ellcey <sellcey@mips.com>
11563 * Makefile.in (TREE_FLOW_H): Remove.
11564 (TREE_SSA_H): Add file names from tree-flow.h.
11565 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11566 * tree.h: Remove tree-flow.h reference.
11567 * hash-table.h: Remove tree-flow.h reference.
11568 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11569 reference with tree-ssa-loop.h.
11571 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11573 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11574 default element-order behavior for -maltivec.
11575 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11576 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11577 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11578 when targeting big endian, at least for now.
11579 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11581 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11583 PR middle-end/47735
11584 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11585 var satisfies use_register_for_decl, just take into account type
11586 alignment, rather than decl alignment.
11588 PR tree-optimization/59622
11589 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11590 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11591 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11592 Don't devirtualize for inplace at all. For targets.length () == 1,
11593 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11595 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11597 * config/i386/i386.md (cpu): Remove the unused btver1.
11599 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11601 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11603 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11606 * tree-core.h (struct target_globals): New forward declaration.
11607 (struct tree_target_option): Add globals field.
11608 * tree.h (TREE_TARGET_GLOBALS): Define.
11609 (prepare_target_option_nodes_for_pch): New prototype.
11610 * target-globals.h (struct target_globals): Define even if
11611 !SWITCHABLE_TARGET.
11612 * tree.c (prepare_target_option_node_for_pch,
11613 prepare_target_option_nodes_for_pch): New functions.
11614 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11615 * config/i386/i386.c: Include target-globals.h.
11616 (ix86_set_current_function): Instead of doing target_reinit
11617 unconditionally, use save_target_globals_default_opts and
11618 restore_target_globals.
11620 2014-01-09 Richard Biener <rguenther@suse.de>
11622 PR tree-optimization/59715
11623 * tree-cfg.h (split_critical_edges): Declare.
11624 * tree-cfg.c (split_critical_edges): Export.
11625 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11627 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11629 * cfgexpand.c (expand_stack_vars): Optionally disable
11630 asan stack protection.
11631 (expand_used_vars): Likewise.
11632 (partition_stack_vars): Likewise.
11633 * asan.c (asan_emit_stack_protection): Optionally disable
11634 after return stack usage.
11635 (instrument_derefs): Optionally disable memory access instrumentation.
11636 (instrument_builtin_call): Likewise.
11637 (instrument_strlen_call): Likewise.
11638 (asan_protect_global): Optionally disable global variables protection.
11639 * doc/invoke.texi: Added doc for new options.
11640 * params.def: Added new options.
11641 * params.h: Likewise.
11643 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11645 PR rtl-optimization/59724
11646 * ifcvt.c (cond_exec_process_if_block): Don't call
11647 flow_find_head_matching_sequence with 0 longest_match.
11648 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11649 non-active insns if !stop_after.
11650 (try_head_merge_bb): Revert 2014-01-07 changes.
11652 2014-01-08 Jeff Law <law@redhat.com>
11654 * ree.c (get_sub_rtx): New function, extracted from...
11655 (merge_def_and_ext): Here.
11656 (combine_reaching_defs): Use get_sub_rtx.
11658 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11660 * cgraph.h (varpool_variable_node): Do not choke on null node.
11662 2014-01-08 Catherine Moore <clm@codesourcery.com>
11664 * config/mips/mips.md (simple_return): Attempt to use JRC
11666 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11668 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11670 PR rtl-optimization/59137
11671 * reorg.c (steal_delay_list_from_target): Call update_block for
11673 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11675 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11677 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11678 two duplicate entries.
11680 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11683 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11685 * config/mips/mips.c (mips_truncated_op_cost): New function.
11686 (mips_rtx_costs): Adjust test for BADDU.
11687 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11689 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11691 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11692 (*baddu_si): ...this new pattern.
11694 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11697 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11699 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11701 PR middle-end/57748
11702 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11704 (expand_expr, expand_normal): Adjust.
11705 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11706 inner_reference_p. Use inner_reference_p to expand inner references.
11707 (store_expr): Adjust.
11708 * cfgexpand.c (expand_call_stmt): Adjust.
11710 2014-01-08 Rong Xu <xur@google.com>
11712 * gcov-io.c (gcov_var): Move from gcov-io.h.
11713 (gcov_position): Ditto.
11714 (gcov_is_error): Ditto.
11715 (gcov_rewrite): Ditto.
11716 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11717 only part to libgcc/libgcov.h.
11719 2014-01-08 Marek Polacek <polacek@redhat.com>
11721 PR middle-end/59669
11722 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11724 2014-01-08 Marek Polacek <polacek@redhat.com>
11727 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11729 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11731 PR rtl-optimization/59649
11732 * stor-layout.c (get_mode_bounds): For BImode return
11733 0 and STORE_FLAG_VALUE.
11735 2014-01-08 Richard Biener <rguenther@suse.de>
11737 PR middle-end/59630
11738 * gimple.h (is_gimple_builtin_call): Remove.
11739 (gimple_builtin_call_types_compatible_p): New.
11740 (gimple_call_builtin_p): New overload.
11741 * gimple.c (is_gimple_builtin_call): Remove.
11742 (validate_call): Rename to ...
11743 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11744 check return types.
11745 (validate_type): New static function.
11746 (gimple_call_builtin_p): New overload and adjust.
11747 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11748 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11749 (gimple_fold_stmt_to_constant_1): Likewise.
11750 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11752 2014-01-08 Richard Biener <rguenther@suse.de>
11754 PR middle-end/59471
11755 * gimplify.c (gimplify_expr): Gimplify register-register type
11756 VIEW_CONVERT_EXPRs to separate stmts.
11758 2014-01-07 Jeff Law <law@redhat.com>
11760 PR middle-end/53623
11761 * ree.c (combine_set_extension): Handle case where source
11762 and destination registers in an extension insn are different.
11763 (combine_reaching_defs): Allow source and destination registers
11764 in extension to be different under limited circumstances.
11765 (add_removable_extension): Remove restriction that the
11766 source and destination registers in the extension are the same.
11767 (find_and_remove_re): Emit a copy from the extension's
11768 destination to its source after the defining insn if
11769 the source and destination registers are different.
11771 PR middle-end/59285
11772 * ifcvt.c (merge_if_block): If we are merging a block with more than
11773 one successor with a block with no successors, remove any BARRIER
11774 after the second block.
11776 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11778 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11780 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11783 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11784 for 14-bit register offsets when INT14_OK_STRICT is false.
11786 2014-01-07 Roland Stigge <stigge@antcom.de>
11787 Michael Meissner <meissner@linux.vnet.ibm.com>
11790 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11791 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11793 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11795 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11798 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11800 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11801 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11802 rtx is const0_rtx or not.
11804 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11807 * target-globals.c (save_target_globals): Remove this_fn_optab
11809 * toplev.c: Include optabs.h.
11810 (target_reinit): Temporarily restore the global options if another
11811 set of options are in force.
11813 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11815 PR rtl-optimization/58668
11816 * cfgcleanup.c (flow_find_cross_jump): Don't count
11817 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11818 to determine what is counted.
11819 (flow_find_head_matching_sequence): Use active_insn_p to determine
11821 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11823 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11824 determine what is counted.
11826 PR tree-optimization/59643
11827 * tree-predcom.c (split_data_refs_to_components): If one dr is
11828 read and one write, determine_offset fails and the write isn't
11829 in the bad component, just put the read into the bad component.
11831 2014-01-07 Mike Stump <mikestump@comcast.net>
11832 Jakub Jelinek <jakub@redhat.com>
11835 * tree-core.h (struct tree_optimization_option): Change optabs
11836 type from unsigned char * to void *.
11837 * optabs.c (init_tree_optimization_optabs): Adjust
11838 TREE_OPTIMIZATION_OPTABS initialization.
11840 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11843 * config/i386/i386.h (struct machine_function): Add
11844 no_drap_save_restore field.
11845 * config/i386/i386.c (ix86_save_reg): Use
11846 !cfun->machine->no_drap_save_restore instead of
11847 crtl->stack_realign_needed.
11848 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11849 this function clears frame_pointer_needed. Set
11850 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11851 and DRAP reg is needed.
11853 2014-01-06 Marek Polacek <polacek@redhat.com>
11856 * doc/implement-c.texi: Mention that other integer types are
11857 permitted as bit-field types in strictly conforming mode.
11859 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11861 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11862 is newly allocated.
11864 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11866 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11868 2014-01-06 Martin Jambor <mjambor@suse.cz>
11871 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11873 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11875 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11879 * var-tracking.c (add_stores): Preserve the value of the source even if
11880 we don't record the store.
11882 2014-01-06 Terry Guo <terry.guo@arm.com>
11884 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11886 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11889 * config/darwin.c (darwin_function_section): Adjust return values to
11890 correspond to optimisation changes made in r206070.
11892 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11894 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11895 from prefetch_block tune setting.
11896 (nocona_cost): Correct size of prefetch block to 64.
11898 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11900 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11901 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11902 used to save the static chain register in the computation of the offset
11903 from which the FP registers need to be restored.
11905 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11907 PR tree-optimization/59519
11908 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11909 ICE if get_current_def (current_new_name) is already non-NULL, as long
11910 as it is a phi result of some other phi in *new_exit_bb that has
11913 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11914 or vmovdqu* for misaligned_operand.
11915 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11916 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11917 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11918 aligned_mem for AVX512F masked aligned load and store builtins and for
11919 non-temporal moves.
11921 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11923 PR tree-optimization/59651
11924 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11925 Address range for negative step should be added by TYPE_SIZE_UNIT.
11927 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11929 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11930 overlapping registers also for registers other than the stack pointer.
11932 2014-01-03 Marek Polacek <polacek@redhat.com>
11935 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11938 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11941 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11944 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11945 (push splitter): Use <P:MODE_SIZE> instead of
11946 GET_MODE_SIZE (<P:MODE>mode).
11947 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11948 (mov -1, reg peephole2): Likewise.
11949 * config/i386/sse.md (*mov<mode>_internal,
11950 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11951 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11952 *<code><mode>3, *andnot<mode>3<mask_name>,
11953 <mask_codefor><code><mode>3<mask_name>): Likewise.
11954 * config/i386/subst.md (mask_mode512bit_condition,
11955 sd_mask_mode512bit_condition): Likewise.
11957 2014-01-02 Xinliang David Li <davidxl@google.com>
11959 PR tree-optimization/59303
11960 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11961 (dump_predicates): Better output format.
11962 (pred_equal_p): New function.
11963 (is_neq_relop_p): Ditto.
11964 (is_neq_zero_form_p): Ditto.
11965 (pred_expr_equal_p): Ditto.
11966 (pred_neg_p): Ditto.
11967 (simplify_pred): Ditto.
11968 (simplify_preds_2): Ditto.
11969 (simplify_preds_3): Ditto.
11970 (simplify_preds_4): Ditto.
11971 (simplify_preds): Ditto.
11972 (push_pred): Ditto.
11973 (push_to_worklist): Ditto.
11974 (get_pred_info_from_cmp): Ditto.
11975 (is_degenerated_phi): Ditto.
11976 (normalize_one_pred_1): Ditto.
11977 (normalize_one_pred): Ditto.
11978 (normalize_one_pred_chain): Ditto.
11979 (normalize_preds): Ditto.
11980 (normalize_cond_1): Remove function.
11981 (normalize_cond): Ditto.
11982 (is_gcond_subset_of): Ditto.
11983 (is_subset_of_any): Ditto.
11984 (is_or_set_subset_of): Ditto.
11985 (is_and_set_subset_of): Ditto.
11986 (is_norm_cond_subset_of): Ditto.
11987 (pred_chain_length_cmp): Ditto.
11988 (convert_control_dep_chain_into_preds): Type change.
11989 (find_predicates): Ditto.
11990 (find_def_preds): Ditto.
11991 (destroy_predicates_vecs): Ditto.
11992 (find_matching_predicates_in_rest_chains): Ditto.
11993 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11994 (is_pred_expr_subset): Ditto.
11995 (is_pred_chain_subset_of): Ditto.
11996 (is_included_in): Ditto.
11997 (is_superset_of): Ditto.
11999 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12001 Update copyright years.
12003 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12005 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12006 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12007 config/arc/arc.md, config/arc/arc.opt,
12008 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12009 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12010 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12011 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12012 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12013 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12014 vtable-verify.c, vtable-verify.h: Use the standard form for the
12017 2014-01-02 Tobias Burnus <burnus@net-b.de>
12019 * gcc.c (process_command): Update copyright notice dates.
12020 * gcov-dump.c: Ditto.
12022 * doc/cpp.texi: Bump @copying's copyright year.
12023 * doc/cppinternals.texi: Ditto.
12024 * doc/gcc.texi: Ditto.
12025 * doc/gccint.texi: Ditto.
12026 * doc/gcov.texi: Ditto.
12027 * doc/install.texi: Ditto.
12028 * doc/invoke.texi: Ditto.
12030 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12032 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12034 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12036 * config/i386/sse.md (*mov<mode>_internal): Guard
12037 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12039 PR rtl-optimization/59647
12040 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12041 new_rtx into UNSIGNED_FLOAT rtxes.
12043 Copyright (C) 2014 Free Software Foundation, Inc.
12045 Copying and distribution of this file, with or without modification,
12046 are permitted in any medium without royalty provided the copyright
12047 notice and this notice are preserved.