1 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
3 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
4 sufficiently aligned and an offset is used at the same time.
5 (expand_expr_real_1): Likewise.
7 2014-05-28 Richard Biener <rguenther@suse.de>
10 * fold-const.c (fold_comparison): When folding
11 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
12 the sign of the remaining constant operand stays the same.
14 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
16 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
17 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
18 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
20 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES
22 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
23 (m32bit-doubles) Likewise.
24 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
25 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
28 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30 * configure.ac ($gcc_cv_ld_clearcap): New test.
31 * configure: Regenerate.
32 * config.in: Regenerate.
33 * config/sol2.opt (mclear-hwcap): New option.
34 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
35 * config/sol2-clearcap.map: Moved here from
36 testsuite/gcc.target/i386/clearcap.map.
37 * config/sol2-clearcapv2.map: Move here from
38 gcc.target/i386/clearcapv2.map.
39 * config/t-sol2 (install): Depend on install-clearcap-map.
40 (install-clearcap-map): New target.
41 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
44 2014-05-28 Richard Biener <rguenther@suse.de>
46 * hwint.h (*_HALF_WIDE_INT*): Move to ...
47 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
48 ... here and remove the rest.
49 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
51 2014-05-28 Richard Biener <rguenther@suse.de>
53 PR tree-optimization/61335
54 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
55 new range fails, drop to varying.
57 2014-05-28 Tom de Vries <tom@codesourcery.com>
59 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
60 * lra.c (initialize_lra_reg_info_element): Add init of
61 actual_call_used_reg_set field.
62 (lra): Call lra_create_live_ranges before lra_inheritance for
64 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
66 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
67 instead of call_used_reg_set for -fuse-caller-save.
68 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
70 2014-05-28 Radovan Obradovic <robradovic@mips.com>
71 Tom de Vries <tom@codesourcery.com>
73 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save to
75 (@item -fuse-caller-save): New item.
77 2014-05-28 Radovan Obradovic <robradovic@mips.com>
78 Tom de Vries <tom@codesourcery.com>
80 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
83 2014-05-28 Radovan Obradovic <robradovic@mips.com>
84 Tom de Vries <tom@codesourcery.com>
86 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
87 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
88 get_call_reg_set_usage.
89 * resource.c (mark_set_resources, mark_target_live_regs): Use
90 get_call_reg_set_usage.
91 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
93 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
94 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
95 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
96 * ira-build.c (ira_create_allocno): Init
97 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
98 (create_cap_allocno, propagate_allocno_info)
99 (propagate_some_info_from_allocno)
100 (copy_info_to_removed_store_destinations): Handle
101 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
102 * ira-costs.c (ira_tune_allocno_costs): Use
103 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
105 2014-05-28 Radovan Obradovic <robradovic@mips.com>
106 Tom de Vries <tom@codesourcery.com>
108 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
109 and function_used_regs_valid fields.
110 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
111 find_all_hard_reg_sets.
112 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
113 (get_call_reg_set_usage): New function.
114 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
115 * regs.h (get_call_reg_set_usage): Declare.
117 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
120 * config/dbx.h (License): Add Runtime Library Exception.
121 * config/newlib-stdint.h (License): Same.
122 * config/rtems.h (License): Same
123 * config/initfini-array.h (License): Same
124 * config/v850/v850.h (License): Same.
125 * config/v850/v850-opts.h (License): Same
126 * config/v850/rtems.h (License): Same.
128 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
131 * doc/extend.texi (Local Labels): Note that label differences are
132 not supported for AVR.
134 2014-05-27 Andrew Pinski <apinski@cavium.com>
136 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
137 Use <w> for the register in assembly template.
138 (stack_protect_test): Use the mode of operands[0] for the
140 (stack_protect_test_<mode>): Use <w> for the register
141 in assembly template.
143 2014-05-27 DJ Delorie <dj@redhat.com>
145 * config/rx/rx.c (add_vector_labels): New.
146 (rx_output_function_prologue): Call it.
147 (rx_handle_func_attribute): Don't require empty arguments.
148 (rx_handle_vector_attribute): New.
149 (rx_attribute_table): Add "vector" attribute.
150 * doc/extend.texi (interrupt, vector): Document new/changed
151 RX-specific attributes.
153 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
155 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
157 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
158 predicate to detect a negative quotient.
160 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
162 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
163 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
164 Add X - Y CMP 0 to X CMP Y transformation.
165 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
167 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
169 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
172 2014-05-27 Steve Ellcey <sellcey@mips.com>
174 * config/mips/mips.c: Add include of cgraph.h.
176 2014-05-27 Richard Biener <rguenther@suse.de>
178 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
180 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
183 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
184 * config/arm/arm-cores.def (License): Same.
185 * config/arm/arm-opts.h (License): Same.
186 * config/arm/aout.h (License): Same.
187 * config/arm/bpabi.h (License): Same.
188 * config/arm/elf.h (License): Same.
189 * config/arm/linux-elf.h (License): Same.
190 * config/arm/linux-gas.h (License): Same.
191 * config/arm/netbsd-elf.h (License): Same.
192 * config/arm/uclinux-eabi.h (License): Same.
193 * config/arm/uclinux-elf.h (License): Same.
194 * config/arm/vxworks.h (License): Same.
196 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
198 * config/arm/neon.md (neon_bswap<mode>): New pattern.
199 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
200 (arm_init_neon_builtins): Handle NEON_BSWAP.
201 Define required type nodes.
202 (arm_expand_neon_builtin): Handle NEON_BSWAP.
203 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
204 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
205 * config/arm/iterators.md (VDQHSD): New mode iterator.
207 2014-05-27 Richard Biener <rguenther@suse.de>
209 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
210 Try using literal operands when comparing value-ranges failed.
212 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
214 * ira.c (commutative_operand): Adjust for change to recog_data.
215 [Missing from previous commit.]
217 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
219 * system.h (TEST_BIT): New macro.
220 * recog.h (alternative_mask): New type.
221 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
222 (recog_data_d): Replace alternative_enabled_p array with
223 enabled_alternatives.
224 (target_recog): New structure.
225 (default_target_recog, this_target_recog): Declare.
226 (get_enabled_alternatives, recog_init): Likewise.
227 * recog.c (default_target_recog, this_target_recog): New variables.
228 (get_enabled_alternatives): New function.
229 (extract_insn): Use it.
230 (recog_init): New function.
231 (preprocess_constraints, constrain_operands): Adjust for change to
233 * postreload.c (reload_cse_simplify_operands): Likewise.
234 * reload.c (find_reloads): Likewise.
235 * ira-costs.c (record_reg_classes): Likewise.
236 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
237 all alternatives after a disabled one would be skipped.
238 (ira_implicitly_set_insn_hard_regs): Likewise.
239 * ira.c (ira_setup_alts): Adjust for change to recog_data.
240 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
241 with enabled_alternatives.
242 * lra.c (free_insn_recog_data): Update accordingly.
243 (lra_update_insn_recog_data): Likewise.
244 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
245 * lra-constraints.c (process_alt_operands): Likewise. Handle
246 only_alternative as part of the enabled mask.
247 * target-globals.h (this_target_recog): Declare.
248 (target_globals): Add a recog field.
249 (restore_target_globals): Restore this_target_recog.
250 * target-globals.c: Include recog.h.
251 (default_target_globals): Initialize recog field.
252 (save_target_globals): Likewise.
253 * reginfo.c (reinit_regs): Call recog_init.
254 * toplev.c (backend_init_target): Likewise.
256 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
258 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
259 rather than any named insn's code.
261 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
264 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
265 * config/arm/arm-cores.def (License): Same.
267 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
269 * tree.h (decl_comdat_group): Declare.
270 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
271 * tree.c (decl_comdat_group): Here.
273 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
275 PR rtl-optimization/61222
276 * combine.c (simplify_shift_const_1): When moving a PLUS outside
277 the shift, truncate the PLUS operand to the result mode.
279 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
282 * config/i386/i386.c (ix86_rtx_costs)
283 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
286 2014-05-26 Martin Jambor <mjambor@suse.cz>
288 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
291 2014-05-26 Richard Biener <rguenther@suse.de>
293 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
294 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
295 Provide specializations.
296 (wi::int_traits <HOST_WIDE_INT>,
297 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
299 2014-05-26 Alan Modra <amodra@gmail.com>
302 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
303 params and return a bool. Remove dead code. Update comment.
304 Assert we have a const_int source. Remove bogus code from
305 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
306 handling of constants > 2G and reg_equal note, from..
307 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
308 return value. Update comment. If we can, use a new pseudo
309 for intermediate calculations.
310 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
312 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
313 call to rs6000_emit_set_const in splitter.
314 (movdi_internal64+2, +3): Likewise.
316 2014-05-26 Richard Biener <rguenther@suse.de>
318 * system.h: Define __STDC_FORMAT_MACROS before
319 including inttypes.h.
320 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
321 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
322 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
323 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
324 HOST_WIDEST_INT_C): Remove.
325 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
326 if C99 inttypes.h is not available.
327 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
328 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
329 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
330 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
331 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
332 (struct output_info): Likewise.
333 (print_statistics): Adjust.
334 (dump_bitmap_statistics): Likewise.
335 * bt-load.c (migrate_btr_defs): Print with PRId64.
336 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
337 (MAX_SAFE_MULTIPLIER): Adjust.
338 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
339 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
340 dump_cgraph_node): Likewise.
341 * final.c (dump_basic_block_info): Likewise.
342 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
343 * gcov.c (format_gcov): Likewise.
344 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
346 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
347 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
348 (inline_small_functions, dump_overall_stats, dump_inline_stats):
349 Use PRId64 for dumping.
350 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
351 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
352 (add_allocno_hard_regs): Adjust.
353 * loop-doloop.c (doloop_modify): Print using PRId64.
354 * loop-iv.c (inverse): Compute in uint64_t.
355 (determine_max_iter, iv_number_of_iterations): Likewise.
356 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
358 * lto-streamer-out.c (write_symbol): Use uint64_t.
359 * mcf.c (CAP_INFINITY): Use int64_t maximum.
360 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
361 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
362 * modulo-sched.c (const_iteration_count): Use int64_t.
363 (sms_schedule): Dump using PRId64.
364 * predict.c (dump_prediction): Likewise.
365 * pretty-print.h (pp_widest_integer): Remove.
366 * profile.c (get_working_sets, is_edge_inconsistent,
367 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
368 * tree-pretty-print.c (pp_double_int): Remove case handling
369 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
370 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
372 (pass_optimize_bswap::execute): Remove restriction on hosts.
373 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
374 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
375 * tree.c (widest_int_cst_value): Remove.
376 * tree.h (widest_int_cst_value): Likewise.
377 * value-prof.c (dump_histogram_value): Print using PRId64.
378 * gengtype.c (main): Also inject int64_t.
379 * ggc-page.c (struct max_alignment): Use int64_t.
380 * alloc-pool.c (struct allocation_object_def): Likewise.
381 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
383 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
384 * doc/tm.texi: Regenerated.
385 * gengtype-lex.l (IWORD): Handle [u]int64_t.
386 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
387 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
388 mmix_output_register_setting): Use [u]int64_t in prototypes.
389 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
390 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
391 mmix_output_octa, mmix_output_shifted_value): Adjust.
392 (mmix_intval): Adjust. Remove unreachable case.
393 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
395 2014-05-26 Richard Biener <rguenther@suse.de>
397 * configure.ac: Drop __int64 type check. Insist that we
398 found uint64_t and int64_t.
399 * hwint.h (HOST_BITS_PER___INT64): Remove.
400 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
401 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
402 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
403 (HOST_WIDEST_FAST_INT): Remove __int64 case.
404 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
405 for dst_q_src_df_rms_cdt.
406 * configure: Regenerate.
407 * config.in: Likewise.
409 2014-05-26 Michael Tautschnig <mt@debian.org>
412 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
413 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
415 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
417 PR rtl-optimization/61278
418 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
420 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
422 PR rtl-optimization/61220
423 Part of PR rtl-optimization/61225
424 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
425 insn; skip split_edge for a block with only one successor.
427 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
429 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
432 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
434 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
435 (update_vtable_references): New function.
436 (function_and_variable_visibility): Rewrite also vtable initializers.
437 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
439 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
441 * ggc.h (ggc_grow): New function.
442 * ggc-none.c (ggc_grow): New function.
443 * ggc-page.c (ggc_grow): Likewise.
445 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
447 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
448 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
449 comdat_can_be_unshared_p, cgraph_externally_visible_p,
450 varpool_externally_visible_p, can_replace_by_local_alias,
451 update_visibility_by_resolution_info, function_and_variable_visibility,
452 pass_data_ipa_function_and_variable_visibility,
453 make_pass_ipa_function_and_variable_visibility,
454 whole_program_function_and_variable_visibility,
455 pass_data_ipa_whole_program_visibility,
456 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
457 * cgraph.h (cgraph_local_node_p): Declare.
458 * ipa-visibility.c: New file.
459 * Makefile.in (OBJS): Add ipa-visiblity.o
461 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
463 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
464 that var decl is available.
466 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
468 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
470 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
471 (find_decls_types_r): Do not walk COMDAT_GROUP.
472 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
473 * varasm.c (make_decl_one_only): Use set_comdat_group;
474 create node if needed.
475 * ipa-inline-transform.c (save_inline_function_body): Update
476 way we decl->symtab mapping.
477 * symtab.c (symtab_hash, hash_node, eq_node
478 symtab_insert_node_to_hashtable): Remove.
479 (symtab_register_node): Update.
480 (symtab_unregister_node): Update.
481 (symtab_get_node): Reimplement as inline function.
482 (symtab_add_to_same_comdat_group): Update.
483 (symtab_dissolve_same_comdat_group_list): Update.
484 (dump_symtab_base): Update.
485 (verify_symtab_base): Update.
486 (symtab_make_decl_local): Update.
487 (fixup_same_cpp_alias_visibility): Update.
488 (symtab_nonoverwritable_alias): Update.
489 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
490 * ipa.c (update_visibility_by_resolution_info): UPdate.
491 * bb-reorder.c: Include cgraph.h
492 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
494 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
495 * cgraph.c (cgraph_get_create_node): Update.
496 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
498 (symtab_get_node): Make inline.
499 (symtab_insert_node_to_hashtable): Remove.
500 (symtab_can_be_discarded): Update.
501 (decl_comdat_group): New function.
502 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
504 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
506 (read_comdat_group): New function.
507 (input_node, input_varpool_node): Use it.
508 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
510 * mips.c (mips_start_unique_function): Likewise.
511 (ix86_code_end): Likewise.
512 (rs6000_code_end): Likweise.
513 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
515 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
517 * gengtype-state.c (fatal_reading_state): Bring offline.
518 * optabs.c (widening_optab_handler): Bring offline.
519 * optabs.h (widening_optab_handler): Likewise.
520 * final.c (get_attr_length_1): Likewise.
522 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
524 * sched-int.h (sd_iterator_cond): Manually tail recurse.
526 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
528 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
529 (ppc440-compare): Include shift with dot.
530 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
531 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
533 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
535 (e6500_sfx2): Include it.
536 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
537 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
538 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
539 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
540 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
541 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
542 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
543 *lshiftrt_internal1le, *lshiftrt_internal1be,
544 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
545 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
546 *rotldi3_internal10le, *rotldi3_internal10be,
547 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
548 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
549 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
550 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
551 define_insns): Use type "shift" in the appropriate alternatives.
553 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
555 * config/rs6000/rs6000.md (type): Add "logical". Delete
557 (dot): Adjust comment.
558 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
559 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
560 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
561 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
562 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
563 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
564 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
565 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
567 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
568 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
569 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
570 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
571 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
572 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
573 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
574 * config/rs6000/8540.md (ppc8540_su): Adjust.
575 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
576 cell-cmp-microcoded): Adjust.
577 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
578 * config/rs6000/e500mc.md (e500mc_su): Adjust.
579 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
580 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
581 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
582 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
583 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
584 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
585 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
587 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
588 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
589 Adjust. Adjust comment.
590 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
591 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
593 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
595 * config/rs6000/rs6000.md (type): Add "add".
596 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
597 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
598 define_insns): Use it.
599 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
601 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
602 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
603 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
604 * config/rs6000/601.md (ppc601-integer): Adjust.
605 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
606 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
607 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
608 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
609 * config/rs6000/8540.md (ppc8540_su): Adjust.
610 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
611 cell-cmp-microcoded): Adjust.
612 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
613 * config/rs6000/e500mc.md (e500mc_su): Adjust.
614 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
615 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
616 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
617 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
618 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
619 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
620 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
622 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
623 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
624 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
625 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
627 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
629 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
630 "delayed_compare", "var_delayed_compare".
631 (var_shift): New attribute.
632 (cell_micro): Adjust.
633 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
634 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
635 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
636 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
637 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
638 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
639 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
640 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
641 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
642 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
643 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
644 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
645 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
646 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
647 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
648 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
649 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
650 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
651 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
652 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
653 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
654 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
655 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
656 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
657 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
659 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
660 * config/rs6000/440.md (ppc440-integer): Adjust.
661 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
662 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
663 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
664 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
665 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
666 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
667 * config/rs6000/8540.md (ppc8540_su): Adjust.
668 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
669 cell-cmp-microcoded): Adjust.
670 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
671 * config/rs6000/e500mc.md (e500mc_su): Adjust.
672 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
673 e500mc64_delayed): Adjust.
674 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
675 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
676 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
677 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
678 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
679 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
680 power6-delayed-compare, power6-var-delayed-compare): Adjust.
681 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
682 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
684 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
685 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
687 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
689 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
690 (bits): New mode_attr.
691 (idiv_ldiv): Delete mode_attr.
692 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
693 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
694 rs6000_adjust_priority, is_nonpipeline_insn,
695 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
697 * config/rs6000/40x.md (ppc403-idiv): Adjust.
698 * config/rs6000/440.md (ppc440-idiv): Adjust.
699 * config/rs6000/476.md (ppc476-idiv): Adjust.
700 * config/rs6000/601.md (ppc601-idiv): Adjust.
701 * config/rs6000/603.md (ppc603-idiv): Adjust.
702 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
703 ppc620-ldiv): Adjust.
704 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
705 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
706 * config/rs6000/8540.md (ppc8540_divide): Adjust.
707 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
708 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
709 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
710 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
711 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
712 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
713 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
714 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
715 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
716 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
717 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
718 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
719 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
720 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
721 * config/rs6000/titan.md (titan_fxu_div): Adjust.
723 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
725 * config/rs6000/rs6000.md (type): Delete "insert_word",
726 "insert_dword". Add "insert".
727 (size): Update comment.
728 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
729 insn_must_be_first_in_group): Adjust.
730 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
731 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
732 *insvsi_internal6, insvdi_internal): Adjust.
734 * config/rs6000/40x.md (ppc403-integer): Adjust.
735 * config/rs6000/440.md (ppc440-integer): Adjust.
736 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
737 * config/rs6000/601.md (ppc601-integer): Adjust.
738 * config/rs6000/603.md (ppc603-integer): Adjust.
739 * config/rs6000/6xx.md (ppc604-integer): Adjust.
740 * config/rs6000/7450.md (ppc7450-integer): Adjust.
741 * config/rs6000/7xx.md (ppc750-integer): Adjust.
742 * config/rs6000/8540.md (ppc8540_su): Adjust.
743 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
744 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
745 * config/rs6000/e500mc.md (e500mc_su): Adjust.
746 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
747 * config/rs6000/e5500.md (e5500_sfx): Adjust.
748 * config/rs6000/e6500.md (e6500_sfx): Adjust.
749 * config/rs6000/mpc.md (mpccore-integer): Adjust.
750 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
751 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
752 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
753 * config/rs6000/power7.md (power7-integer): Adjust.
754 * config/rs6000/power8.md (power8-1cyc): Adjust.
755 * config/rs6000/rs64.md (rs64a-integer): Adjust.
756 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
758 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
760 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
761 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
762 (size): New attribute.
763 (dot): New attribute.
764 (cell_micro): Adjust.
765 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
766 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
767 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
768 umuldi3_highpart): Adjust.
769 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
770 rs6000_adjust_priority, is_nonpipeline_insn,
771 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
773 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
774 ppc405-imul3): Adjust.
775 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
776 * config/rs6000/476.md (ppc476-imul): Adjust.
777 * config/rs6000/601.md (ppc601-imul): Adjust.
778 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
779 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
780 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
781 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
782 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
784 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
785 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
786 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
788 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
789 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
790 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
791 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
792 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
793 * config/rs6000/mpc.md (mpccore-imul): Adjust.
794 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
795 power4-lmul, power4-imul, power4-imul3): Adjust.
796 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
797 power5-lmul, power5-imul, power5-imul3): Adjust.
798 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
799 power6-lmul, power6-imul, power6-imul3): Adjust.
800 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
801 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
803 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
805 * config/rs6000/titan.md (titan_imul): Adjust.
807 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
809 * config/rs6000/rs6000.md (type): Add new value "halfmul".
810 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
811 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
812 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
813 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
814 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
815 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
816 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
817 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
818 * config/rs6000/titan.md: Delete nonsensical comment.
819 (titan_imul): Add type imul3.
820 (titan_mulhw): Remove type imul3; add type halfmul.
822 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
824 * config/rs6000/rs6000.md (type): Reorder, reformat.
826 2014-05-23 Martin Jambor <mjambor@suse.cz>
828 PR tree-optimization/53787
829 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
830 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
831 analysis_done, update all uses.
832 * ipa-prop.c: Include domwalk.h
833 (param_analysis_info): Removed.
834 (param_aa_status): New type.
835 (ipa_bb_info): Likewise.
836 (func_body_info): Likewise.
837 (ipa_get_bb_info): New function.
838 (aa_overwalked): Likewise.
839 (find_dominating_aa_status): Likewise.
840 (parm_bb_aa_status_for_bb): Likewise.
841 (parm_preserved_before_stmt_p): Changed to use new param AA info.
842 (load_from_unmodified_param): Accept func_body_info as a parameter
843 instead of parms_ainfo.
844 (parm_ref_data_preserved_p): Changed to use new param AA info.
845 (parm_ref_data_pass_through_p): Likewise.
846 (ipa_load_from_parm_agg_1): Likewise. Update callers.
847 (compute_complex_assign_jump_func): Changed to use new param AA info.
848 (compute_complex_ancestor_jump_func): Likewise.
849 (ipa_compute_jump_functions_for_edge): Likewise.
850 (ipa_compute_jump_functions): Removed.
851 (ipa_compute_jump_functions_for_bb): New function.
852 (ipa_analyze_indirect_call_uses): Likewise, moved variable
854 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
855 and info, moved variable declarations down.
856 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
858 (ipa_analyze_stmt_uses): Likewise.
859 (ipa_analyze_params_uses): Removed.
860 (ipa_analyze_params_uses_in_bb): New function.
861 (ipa_analyze_controlled_uses): Likewise.
862 (free_ipa_bb_info): Likewise.
863 (analysis_dom_walker): New class.
864 (ipa_analyze_node): Handle node-specific forbidden analysis,
865 initialize and free func_body_info, use dominator walker.
866 (ipcp_modif_dom_walker): New class.
867 (ipcp_transform_function): Create and free func_body_info, use
868 ipcp_modif_dom_walker, moved a lot of functionality there.
870 2014-05-23 Marek Polacek <polacek@redhat.com>
871 Jakub Jelinek <jakub@redhat.com>
873 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
874 * gcc.c (sanitize_spec_function): Likewise.
875 * convert.c (convert_to_integer): Include "ubsan.h". Add
876 floating-point to integer instrumentation.
877 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
878 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
880 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
881 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
882 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
883 * ubsan.c: Include "realmpfr.h" and "dfp.h".
884 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
885 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
886 float/double/long double.
887 (ubsan_instrument_float_cast): New function.
888 * ubsan.h (ubsan_instrument_float_cast): Declare.
890 2014-05-23 Jiong Wang <jiong.wang@arm.com>
892 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
894 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
895 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
896 Adjust for tailcalling through registers.
897 * config/aarch64/aarch64.h (enum reg_class): New caller save
899 (REG_CLASS_NAMES): Likewise.
900 (REG_CLASS_CONTENTS): Likewise.
901 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
902 Allow tailcalling without decls.
904 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
906 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
907 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
909 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
910 gsi, and variables v_* to v*.
912 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
914 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
916 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
918 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
919 * omp-low.c: Update accordingly.
921 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
922 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
923 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
924 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
925 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
926 GF_OMP_TARGET_KIND_UPDATE.
928 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
929 Explicitly enumerate the expected region types.
931 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
934 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
935 documentation; the old documentation didn't clearly state the
936 constraints on the contents of the pointed-to storage.
938 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
940 Fix bootstrap error on ia64
941 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
942 Return default value.
944 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
946 PR tree-optimization/54733
947 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
949 (find_bswap_or_nop_load): New.
950 (find_bswap_1): Renamed to ...
951 (find_bswap_or_nop_1): This. Also add support for memory source.
952 (find_bswap): Renamed to ...
953 (find_bswap_or_nop): This. Also add support for memory source and
954 detection of bitwise operations equivalent to load in host endianness.
955 (execute_optimize_bswap): Likewise. Also move its leading comment back
956 in place and split statement transformation into ...
957 (bswap_replace): This.
959 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
961 PR rtl-optimization/61215
962 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
963 simplify_gen_subreg until final substitution.
965 2014-05-23 Alan Modra <amodra@gmail.com>
968 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
969 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
970 Use "Y" constraint rather than "m".
972 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
974 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
976 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
977 New function declaration.
978 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
979 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
980 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
981 (aarch64_init_builtins) : Initialize builtins
982 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
983 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
984 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
985 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
986 and __builtins_aarch64_set_fpsr.
987 (aarch64_atomic_assign_expand_fenv): New function.
988 * config/aarch64/aarch64.md (set_fpcr): New pattern.
989 (get_fpcr) : Likewise.
990 (set_fpsr) : Likewise.
991 (get_fpsr) : Likewise.
992 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
993 and UNSPECV_SET_FPSR.
994 * doc/extend.texi (AARCH64 Built-in Functions) : Document
995 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
996 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
998 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1000 PR rtl-optimization/60969
1001 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1002 constraints. Set up mem cost for NO_REGS case.
1004 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1006 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1008 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1010 * config/darwin.c: Include "lto-section-names.h".
1011 (LTO_SEGMENT_NAME): Don't define.
1012 * config/i386/winnt.c: Include "lto-section-names.h".
1013 * lto-streamer.c: Include "lto-section-names.h".
1014 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1015 * lto-wrapper.c: Include "lto-section-names.h".
1016 (LTO_SECTION_NAME_PREFIX): Don't define.
1017 * lto-section-names.h: New file.
1018 * cgraphunit.c: Include "lto-section-names.h".
1020 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1022 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1024 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1027 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1029 2014-05-22 Nick Clifton <nickc@redhat.com>
1031 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1033 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1035 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1036 -> (T)A transformation to integer types.
1038 2014-05-22 Teresa Johnson <tejohnson@google.com>
1040 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1041 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1042 (gcov_rewrite): Use gcov_nonruntime_assert.
1044 (gcov_write_words): Ditto.
1045 (gcov_write_length): Ditto.
1046 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1047 gcc_assert from IN_LIBGCOV code.
1048 (gcov_read_summary): Use gcov_error to flag profile corruption.
1049 (gcov_sync): Use gcov_nonruntime_assert.
1050 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1051 (gcov_histo_index): Use gcov_nonruntime_assert.
1052 (static void gcov_histogram_merge): Ditto.
1053 (compute_working_sets): Ditto.
1054 * gcov-io.h (gcov_nonruntime_assert): Define.
1055 (gcov_error): Define for !IN_LIBGCOV
1057 2014-05-22 Richard Biener <rguenther@suse.de>
1059 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1060 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1061 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1062 and deallocation site.
1063 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1064 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1065 passing through the incoming points-to set.
1066 (handle_lhs_call): Use flags argument instead of recomputing it.
1067 (find_func_aliases_for_call): Call handle_lhs_call with proper
1070 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1072 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1073 all padding bits in REAL_VALUE_TYPE are cleared.
1075 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1077 Cleanup and improve multipass_dfa_lookahead_guard
1078 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1079 (core2i7_first_cycle_multipass_begin,)
1080 (core2i7_first_cycle_multipass_issue,)
1081 (core2i7_first_cycle_multipass_backtrack): Update signature.
1082 * config/ia64/ia64.c
1083 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1084 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1085 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1087 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1088 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1090 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1092 * doc/tm.texi: Regenerate.
1094 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1095 * haifa-sched.c (ready_try): Make signed to allow negative values.
1096 (rebug_ready_list_1): Update.
1097 (choose_ready): Simplify.
1098 (sched_extend_ready_list): Update.
1100 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1102 Remove IA64 speculation tweaking flags
1103 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1104 speculation tuning flags.
1105 (msched-prefer-non-data-spec-insns,)
1106 (msched-prefer-non-control-spec-insns): Obsolete options.
1107 * haifa-sched.c (choose_ready): Remove handling of
1108 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1109 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1110 and PREFER_NON_DATA_SPEC.
1111 * sel-sched.c (process_spec_exprs): Remove handling of
1112 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1114 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1116 Improve scheduling debug output
1117 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1118 (advance_one_cycle): Update.
1119 (schedule_insn, queue_to_ready): Add debug printouts.
1120 (debug_ready_list_1): New static function.
1121 (debug_ready_list): Update.
1122 (max_issue): Add debug printouts.
1123 (dump_insn_stream): New static function.
1124 (schedule_block): Use it. Also better indent printouts.
1126 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1128 Fix sched_insn debug counter
1129 * haifa-sched.c (schedule_insn): Update.
1130 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1131 (save_backtrack_point, restore_backtrack_point): Update.
1132 (first_nonscheduled_insn): New static function.
1133 (queue_to_ready, choose_ready): Use it.
1134 (schedule_block): Init nonscheduled_insns_begin.
1135 (sched_emit_insn): Update.
1138 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1140 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1142 (aarch64_secondary_reload) : LikeWise.
1143 (aarch64_class_max_nregs) : Remove CORE_REGS.
1144 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1145 (REG_CLASS_NAMES) : Likewise.
1146 (REG_CLASS_CONTENTS) : LikeWise.
1147 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1149 2014-05-21 Guozhi Wei <carrot@google.com>
1152 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1154 (vqdmulhq_n_s16): Likewise.
1156 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1158 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1160 2014-05-21 Marek Polacek <polacek@redhat.com>
1163 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1165 2014-05-21 Martin Jambor <mjambor@suse.cz>
1167 * doc/invoke.texi (Optimize Options): Document parameters
1168 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1169 ipa-cp-array-index-hint-bonus.
1171 2014-05-21 Mark Wielaard <mjw@redhat.com>
1174 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1175 version >= 3 or not strict DWARF.
1176 * langhooks.h (struct lang_hooks_for_types): Add
1177 enum_underlying_base_type.
1178 * langhooks.c (lhd_enum_underlying_base_type): New function.
1179 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1180 enum_underlying_base_type.
1181 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1182 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1183 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1185 2014-05-21 Richard Biener <rguenther@suse.de>
1187 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1189 2014-05-21 John Marino <gnugcc@marino.st>
1191 * config.gcc (*-*-dragonfly*): New target.
1192 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1193 * configure: Regenerate.
1194 * config/dragonfly-stdint.h: New.
1195 * config/dragonfly.h: New.
1196 * config/dragonfly.opt: New.
1197 * config/i386/dragonfly.h: New.
1198 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1200 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1202 * tree.def (VOID_CST): New.
1203 * tree-core.h (TI_VOID): New.
1204 * tree.h (void_node): New.
1205 * tree.c (tree_node_structure_for_code, tree_code_size)
1206 (iterative_hash_expr): Handle VOID_CST.
1207 (build_common_tree_nodes): Initialize void_node.
1209 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1211 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1213 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1215 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1218 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1219 flag_reorder_blocks_and_partition.
1220 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1222 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1225 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1227 (*addc_r_t): Add new insn_and_split.
1229 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1232 * omp-low.c (handle_simd_reference): New function.
1233 (lower_rec_input_clauses): Use it. Defer adding reference
1234 initialization even for reduction without placeholder if in simd,
1235 handle it properly later on.
1237 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1239 PR tree-optimization/60899
1240 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1241 assume all static symbols will have definition wile parsing and
1242 check the do have definition later in compilation; check that
1243 variable referring symbol will be output before concluding that
1244 reference is safe; be conservative for referring local statics;
1245 be more precise about when comdat is output in other partition.
1247 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1250 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1252 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1253 (ipa_inline): Loop inline_to_all_callers until no more aliases
1256 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1258 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1259 set writeonly flag only for vars actually written to.
1261 2014-05-20 Dehao Chen <dehao@google.com>
1263 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1264 and callee count to get clone count.
1265 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1268 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1270 PR rtl-optimization/61243
1271 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1273 2014-05-20 Xinliang David Li <davidxl@google.com>
1275 * cgraphunit.c (walk_polymorphic_call_targets): Add
1276 dbgcnt and fopt-info support.
1277 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1278 * ipa-devirt.c (ipa_devirt): Ditto.
1279 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1280 * ipa.c (walk_polymorphic_call_targets): Ditto.
1281 * gimple-fold.c (fold_gimple_assign): Ditto.
1282 (gimple_fold_call): Ditto.
1283 * dbgcnt.def: New counter.
1285 2014-05-20 DJ Delorie <dj@redhat.com>
1287 * config/msp430/msp430.md (split): Don't allow subregs when
1288 splitting SImode adds.
1289 (andneghi): Fix subtraction logic.
1290 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1292 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1294 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1296 * except.c (switch_to_exception_section, resolve_unique_section,
1297 get_named_text_section, default_function_rodata_section,
1298 align_variable, get_block_for_decl, default_section_type_flags):
1299 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1300 * symtab.c (symtab_add_to_same_comdat_group,
1301 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1302 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1304 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1305 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1306 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1307 (c6x_function_in_section_p): Likewise.
1308 * config/darwin.c (machopic_select_section): Likewise.
1309 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1310 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1311 * config/mep/mep.c (mep_select_section): LIkewise.
1312 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1314 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1316 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1317 EH region of calls to pure functions that can throw an exception.
1318 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1319 (copy_reference_ops_from_call): Also copy the EH region of the call if
1320 it can throw an exception.
1322 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1324 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1325 nested VEC_SELECTs that are inverses of each other.
1327 2014-05-20 Richard Biener <rguenther@suse.de>
1329 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1330 (extract_and_process_scc_for_name): not here.
1331 (cond_dom_walker::before_dom_children): Only process
1332 stmts that end the BB in interesting ways.
1333 (run_scc_vn): Mark param uses as visited.
1335 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1337 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1340 2014-05-20 Nick Clifton <nickc@redhat.com>
1342 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1343 (msp430_gimplify_va_arg_expr): New function.
1344 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1346 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1347 operand 0 in order to prevent confusion about the number of
1350 2014-05-20 Richard Biener <rguenther@suse.de>
1352 PR tree-optimization/61221
1353 * tree-ssa-pre.c (el_to_update): Remove.
1354 (eliminate_dom_walker::before_dom_children): Handle released
1355 VDEFs by value-numbering them to the associated VUSE. Update
1356 stmt immediately for substituted call address.
1357 (eliminate): Remove delayed stmt updating code.
1358 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1359 possibly late re-numbered vuses.
1360 (vn_reference_lookup_2): Adjust.
1361 (vn_reference_lookup_pieces): Likewise.
1362 (vn_reference_lookup): Likewise.
1364 2014-05-20 Richard Biener <rguenther@suse.de>
1366 * config.gcc: Remove need_64bit_hwint.
1367 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1368 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1370 * config.in: Regenerate.
1371 * configure: Likewise.
1373 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1375 * doc/extend.texi: Create Label Attributes section,
1376 move all label attributes into it and reference it.
1378 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1380 * arm.c (thumb1_reorg): When scanning backwards skip anything
1381 that's not a proper insn.
1383 2014-05-19 Richard Biener <rguenther@suse.de>
1385 PR tree-optimization/61221
1386 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1387 Do nothing for unreachable blocks.
1388 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1389 Improve unreachability detection.
1391 2014-05-19 Richard Biener <rguenther@suse.de>
1393 PR tree-optimization/61209
1394 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1396 2014-05-19 Nick Clifton <nickc@redhat.com>
1398 * except.c (init_eh): Fix computation of builtin setjmp buffer
1399 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1401 2014-05-19 Richard Biener <rguenther@suse.de>
1403 PR tree-optimization/61184
1404 * tree-vrp.c (is_negative_overflow_infinity): Use
1405 TREE_OVERFLOW_P and do that check first.
1406 (is_positive_overflow_infinity): Likewise.
1407 (is_overflow_infinity): Likewise.
1408 (vrp_operand_equal_p): Properly treat operands with
1409 differing overflow as not equal.
1411 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1413 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1414 shift simplification where it was intended.
1416 2014-05-19 Christian Bruel <christian.bruel@st.com>
1419 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1421 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1424 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1427 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1429 * reg-notes.def (CROSSING_JUMP): Likewise.
1430 * rtl.h (rtx_def): Update comment for jump flag.
1431 (CROSSING_JUMP_P): Define.
1432 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1433 of a REG_CROSSING_JUMP note.
1434 * cfghooks.c (tidy_fallthru_edges): Likewise.
1435 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1436 * emit-rtl.c (try_split): Likewise.
1437 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1438 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1439 * jump.c (redirect_jump_2): Likewise.
1440 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1441 (relax_delay_slots): Likewise.
1442 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1443 (bbit_di): Likewise.
1444 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1445 * config/sh/sh.md (jump_compact): Likewise.
1446 * bb-reorder.c (rotate_loop): Likewise.
1447 (pass_duplicate_computed_gotos::execute): Likewise.
1448 (add_reg_crossing_jump_notes): Rename to...
1449 (update_crossing_jump_flags): ...this.
1450 (pass_partition_blocks::execute): Update accordingly.
1452 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1454 * tree.h: Remove extraneous template <>.
1456 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1458 * ipa.c (symtab_remove_unreachable_nodes): Remove
1459 symbol from comdat group if its body was eliminated.
1460 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1461 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1462 (symtab_unregister_node): ... this one.
1463 (verify_symtab_base): More strict checking of comdats.
1464 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1466 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1468 * tree-pass.h (make_pass_ipa_comdats): New pass.
1469 * timevar.def (TV_IPA_COMDATS): New timevar.
1470 * passes.def (pass_ipa_comdats): Add.
1471 * Makefile.in (OBJS): Add ipa-comdats.o
1472 * ipa-comdats.c: New file.
1474 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1476 * ipa.c (update_visibility_by_resolution_info): New function.
1477 (function_and_variable_visibility): Use it.
1479 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1481 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1483 (FOR_EACH_DEFINED_SYMBOL): New macro.
1484 (varpool_first_static_initializer, varpool_next_static_initializer,
1485 varpool_first_defined_variable, varpool_next_defined_variable):
1487 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1489 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1491 * ggc-page.c (ggc_handle_finalizers): Add comment.
1493 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1495 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1496 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1497 (ggc_internal_cleared_alloc): Likewise.
1498 * ggc-page.c (finalizer): New class.
1499 (vec_finalizer): Likewise.
1500 (globals::finalizers): New member.
1501 (globals::vec_finalizers): Likewise.
1502 (ggc_internal_alloc): Record the finalizer if any for the block being
1504 (ggc_handle_finalizers): New function.
1505 (ggc_collect): Call ggc_handle_finalizers.
1506 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1508 (ggc_internal_cleared_alloc): Likewise.
1509 (finalize): New function.
1510 (need_finalization_p): Likewise.
1511 (ggc_alloc): Install the type's destructor as the finalizer if it
1513 (ggc_cleared_alloc): Likewise.
1514 (ggc_vec_alloc): Likewise.
1515 (ggc_cleared_vec_alloc): Likewise.
1517 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1519 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1521 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1523 * alias.c (record_alias_subset): Adjust.
1524 * bitmap.c (bitmap_element_allocate): Likewise.
1525 (bitmap_gc_alloc_stat): Likewise.
1526 * cfg.c (init_flow): Likewise.
1527 (alloc_block): Likewise.
1528 (unchecked_make_edge): Likewise.
1529 * cfgloop.c (alloc_loop): Likewise.
1530 (flow_loops_find): Likewise.
1531 (rescan_loop_exit): Likewise.
1532 * cfgrtl.c (init_rtl_bb_info): Likewise.
1533 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1534 (cgraph_allocate_node): Likewise.
1535 (cgraph_create_edge_1): Likewise.
1536 (cgraph_allocate_init_indirect_info): Likewise.
1537 * cgraphclones.c (cgraph_clone_edge): Likewise.
1538 * cgraphunit.c (add_asm_node): Likewise.
1539 (init_lowered_empty_function): Likewise.
1540 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1541 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1542 (alpha_use_linkage): Likewise.
1543 * config/arc/arc.c (arc_init_machine_status): Likewise.
1544 * config/arm/arm.c (arm_init_machine_status): Likewise.
1545 * config/avr/avr.c (avr_init_machine_status): Likewise.
1546 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1547 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1548 * config/cris/cris.c (cris_init_machine_status): Likewise.
1549 * config/darwin.c (machopic_indirection_name): Likewise.
1550 (darwin_build_constant_cfstring): Likewise.
1551 (darwin_enter_string_into_cfstring_table): Likewise.
1552 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1553 * config/frv/frv.c (frv_init_machine_status): Likewise.
1554 * config/i386/i386.c (get_dllimport_decl): Likewise.
1555 (ix86_init_machine_status): Likewise.
1556 (assign_386_stack_local): Likewise.
1557 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1558 (i386_pe_maybe_record_exported_symbol): Likewise.
1559 (i386_pe_record_stub): Likewise.
1560 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1561 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1562 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1563 (m32c_note_pragma_address): Likewise.
1564 * config/mep/mep.c (mep_init_machine_status): Likewise.
1565 (mep_note_pragma_flag): Likewise.
1566 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1567 (mips16_local_alias): Likewise.
1568 (mips_init_machine_status): Likewise.
1569 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1570 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1571 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1572 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1573 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1574 * config/pa/pa.c (pa_init_machine_status): Likewise.
1575 (pa_get_deferred_plabel): Likewise.
1576 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1577 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1578 (rs6000_init_machine_status): Likewise.
1579 (output_toc): Likewise.
1580 * config/s390/s390.c (s390_init_machine_status): Likewise.
1581 * config/score/score.c (score_output_external): Likewise.
1582 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1583 * config/spu/spu.c (spu_init_machine_status): Likewise.
1584 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1585 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1586 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1587 * coverage.c (coverage_end_function): Likewise.
1588 * dbxout.c (dbxout_init): Likewise.
1589 * doc/gty.texi: Don't mention variable_size attribute.
1590 * dwarf2cfi.c (new_cfi): Adjust.
1591 (new_cfi_row): Likewise.
1592 (copy_cfi_row): Likewise.
1593 (create_cie_data): Likewise.
1594 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1595 (new_loc_descr): Likewise.
1596 (find_AT_string_in_table): Likewise.
1597 (add_addr_table_entry): Likewise.
1598 (new_die): Likewise.
1599 (add_var_loc_to_decl): Likewise.
1600 (clone_die): Likewise.
1601 (clone_as_declaration): Likewise.
1602 (break_out_comdat_types): Likewise.
1603 (new_loc_list): Likewise.
1604 (add_loc_descr_to_each): Likewise.
1605 (add_location_or_const_value_attribute): Likewise.
1606 (add_linkage_name): Likewise.
1607 (lookup_filename): Likewise.
1608 (dwarf2out_var_location): Likewise.
1609 (new_line_info_table): Likewise.
1610 (dwarf2out_init): Likewise.
1611 (mem_loc_descriptor): Likewise.
1612 (loc_descriptor): Likewise.
1613 (add_const_value_attribute): Likewise.
1614 (tree_add_const_value_attribute): Likewise.
1615 (comp_dir_string): Likewise.
1616 (dwarf2out_vms_debug_main_pointer): Likewise.
1617 (string_cst_pool_decl): Likewise.
1618 * emit-rtl.c (set_mem_attrs): Likewise.
1619 (get_reg_attrs): Likewise.
1620 (start_sequence): Likewise.
1621 (init_emit): Likewise.
1622 (init_emit_regs): Likewise.
1623 * except.c (init_eh_for_function): Likewise.
1624 (gen_eh_region): Likewise.
1625 (gen_eh_region_catch): Likewise.
1626 (gen_eh_landing_pad): Likewise.
1627 (add_call_site): Likewise.
1628 * function.c (add_frame_space): Likewise.
1629 (insert_temp_slot_address): Likewise.
1630 (assign_stack_temp_for_type): Likewise.
1631 (get_hard_reg_initial_val): Likewise.
1632 (allocate_struct_function): Likewise.
1633 (prepare_function_start): Likewise.
1634 (types_used_by_var_decl_insert): Likewise.
1635 * gengtype.c (variable_size_p): Remove function.
1636 (enum alloc_quantity): Remove enum.
1637 (write_typed_alloc_def): Remove function.
1638 (write_typed_struct_alloc_def): Likewise.
1639 (write_typed_typedef_alloc_def): Likewise.
1640 (write_typed_alloc_defns): Likewise.
1642 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1643 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1644 * ggc.h (ggc_alloc): new function.
1645 (ggc_cleared_alloc): Likewise.
1646 (ggc_vec_alloc): Template on type of vector element, and remove
1647 element size argument.
1648 (ggc_cleared_vec_alloc): Likewise.
1649 * gimple.c (gimple_build_omp_for): Adjust.
1650 (gimple_copy): Likewise.
1651 * ipa-cp.c (get_replacement_map): Likewise.
1652 (find_aggregate_values_for_callers_subset): Likewise.
1653 (known_aggs_to_agg_replacement_list): Likewise.
1654 * ipa-devirt.c (get_odr_type): Likewise.
1655 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1656 (read_agg_replacement_chain): Likewise.
1657 * loop-iv.c (get_simple_loop_desc): Likewise.
1658 * lto-cgraph.c (input_node_opt_summary): Likewise.
1659 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1660 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1661 (input_eh_region): Likewise.
1662 (input_eh_lp): Likewise.
1663 (input_cfg): Likewise.
1664 * optabs.c (set_optab_libfunc): Likewise.
1665 (init_tree_optimization_optabs): Likewise.
1666 (set_conv_libfunc): Likewise.
1667 * passes.c (do_per_function_toporder): Likewise.
1668 * rtl.h: Don't use variable_size gty attribute.
1669 * sese.c (if_region_set_false_region): Adjust.
1670 * stringpool.c (gt_pch_save_stringpool): Likewise.
1671 * target-globals.c (save_target_globals): Likewise.
1672 * toplev.c (general_init): Likewise.
1673 * trans-mem.c (record_tm_replacement): Likewise.
1674 (split_bb_make_tm_edge): Likewise.
1675 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1676 * tree-data-ref.h (lambda_vector_new): Likewise.
1677 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1678 * tree-iterator.c (tsi_link_before): Likewise.
1679 (tsi_link_after): Likewise.
1680 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1681 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1682 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1683 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1684 * tree-ssa.c (init_tree_ssa): Adjust.
1685 * tree-ssanames.c (set_range_info): Likewise.
1686 (get_ptr_info): Likewise.
1687 (duplicate_ssa_name_ptr_info): Likewise.
1688 (duplicate_ssa_name_range_info): Likewise.
1689 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1690 (unpack_ts_fixed_cst_value_fields): Likewise.
1691 * tree.c (build_fixed): Likewise.
1692 (build_real): Likewise.
1693 (build_string): Likewise.
1694 (decl_priority_info): Likewise.
1695 (decl_debug_expr_insert): Likewise.
1696 (decl_value_expr_insert): Likewise.
1697 (decl_debug_args_insert): Likewise.
1698 (type_hash_add): Likewise.
1699 (build_omp_clause): Likewise.
1700 * ubsan.c (decl_for_type_insert): Likewise.
1701 * varasm.c (get_unnamed_section): Likewise.
1702 (get_noswitch_section): Likewise.
1703 (get_section): Likewise.
1704 (get_block_for_section): Likewise.
1705 (create_block_symbol): Likewise.
1706 (build_constant_desc): Likewise.
1707 (create_constant_pool): Likewise.
1708 (force_const_mem): Likewise.
1709 (record_tm_clone_pair): Likewise.
1710 * varpool.c (varpool_create_empty_node): Likewise.
1712 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1714 * dwarf2out.c (tree_add_const_value_attribute): Call
1715 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1716 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1717 instead of ggc_internal_<x>alloc_stat.
1718 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1719 (ggc_realloc): Likewise.
1720 * ggc-none.c (ggc_internal_alloc): Likewise.
1721 (ggc_internal_cleared_alloc): Likewise.
1722 * ggc-page.c: Likewise.
1723 * ggc.h (ggc_internal_alloc_stat): Likewise.
1724 (ggc_internal_alloc): Remove macro.
1725 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1726 (ggc_internal_cleared_alloc): Remove macro.
1727 (GGC_RESIZEVEC): Adjust.
1728 (ggc_resizevar): Remove macro.
1729 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1730 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1731 (ggc_internal_vec_cleared_alloc): Remove macro.
1732 (ggc_alloc_atomic_stat): Drop _stat suffix.
1733 (ggc_alloc_atomic): Remove macro.
1734 (ggc_alloc_cleared_atomic): Remove macro.
1735 (ggc_alloc_string_stat): Drop _stat suffix.
1736 (ggc_alloc_string): Remove macro.
1737 (ggc_alloc_rtx_def_stat): Adjust.
1738 (ggc_alloc_tree_node_stat): Likewise.
1739 (ggc_alloc_cleared_tree_node_stat): Likewise.
1740 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1741 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1742 * gimple.c (gimple_build_omp_for): Likewise.
1743 (gimple_copy): Likewise.
1744 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1745 * toplev.c (realloc_for_line_map): Adjust.
1746 * tree-data-ref.h (lambda_vector_new): Likewise.
1747 * tree-phinodes.c (allocate_phi_node): Likewise.
1748 * tree.c (grow_tree_vec_stat): Likewise.
1749 * vec.h (va_gc::reserve): Adjust.
1751 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1753 * config/microblaze/microblaze.c (break_handler): New Declaration.
1754 (microblaze_break_function_p,microblaze_is_break_handler): New.
1755 (compute_frame_size): Use microblaze_break_function_p.
1756 Add the test of break_handler.
1757 (microblaze_function_prologue) : Add the test of variable
1758 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1759 (microblaze_function_epilogue) : Add the test of break_handler.
1760 (microblaze_globalize_label) : Add the test of break_handler.
1761 Check the name by BREAK_HANDLER_NAME.
1763 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1765 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1766 microblaze_is_break_handler test.
1767 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1768 Use SYMBOL_REF_DECL.
1770 * config/microblaze/microblaze-protos.h
1771 (microblaze_break_function_p,microblaze_is_break_handler):
1774 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1775 new MicroBlaze break_handler functions.
1777 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1779 * doc/extend.texi (Size of an asm): Move node text according
1780 to its @menu entry position.
1782 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1784 PR tree-optimization/61140
1785 PR tree-optimization/61150
1786 PR tree-optimization/61197
1787 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1789 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1791 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1793 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1795 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1796 __SIZEOF_INT128__ is defined.
1798 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1800 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1801 (rs6000_delegitimize_address): Use it.
1803 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1805 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1806 inplace argument. Store the new address in the original MEM when true.
1807 * emit-rtl.c (change_address_1): Likewise.
1808 (adjust_address_1, adjust_automodify_address_1, offset_address):
1810 * rtl.h (plus_constant): Add an inplace argument.
1811 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1812 when true. Avoid generating (plus X (const_int 0)).
1813 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1814 in-place. Pass true to plus_constant.
1815 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1817 2014-05-16 Dehao Chen <dehao@google.com>
1819 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1821 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1824 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1826 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1828 2014-05-16 Dehao Chen <dehao@google.com>
1830 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1831 optimize_function_for_size_p.
1832 * regs.h (REG_FREQ_FROM_BB): Likewise.
1834 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1837 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1838 negt_reg_operand cases.
1839 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1841 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1843 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1845 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1848 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1851 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1853 * tree-cfg.c (dump_function_to_file): Dump the return type of
1854 functions, in a line to itself before the function body, mimicking
1855 the layout of a C function.
1857 2014-05-16 Dehao Chen <dehao@google.com>
1859 * cfghooks.c (make_forwarder_block): Use direct computation to
1860 get fall-through edge's count and frequency.
1862 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1864 * config/arc/arc.c (arc_init): Fix typo in error message.
1865 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1866 (split_stack_prologue_scratch_regno): Likewise.
1867 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1868 word from error message.
1870 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1872 * ira-costs.c: Fix typo in comment.
1874 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1876 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1878 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1880 * varpool.c (dump_varpool_node): Dump write-only flag.
1881 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1883 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1884 write-only variables.
1885 * ipa.c (process_references): New function.
1886 (set_readonly_bit): New function.
1887 (set_writeonly_bit): New function.
1888 (clear_addressable_bit): New function.
1889 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1890 fix handling of aliases.
1891 * cgraph.h (struct varpool_node): Add writeonly flag.
1893 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1895 PR rtl-optimization/60969
1896 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1897 Calculate costs for this case.
1899 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1901 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1902 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1904 2014-05-16 Richard Biener <rguenther@suse.de>
1906 PR tree-optimization/61194
1907 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1908 bool patterns ending in a COND_EXPR.
1910 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1912 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1914 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1916 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1917 where we were unable to cost an RTX.
1919 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1921 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1924 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1925 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1927 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1929 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1930 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1932 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1933 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1935 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1936 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1938 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1941 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1942 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1944 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1947 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1948 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1950 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1951 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1953 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1954 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1956 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1959 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1960 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1962 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1963 ZERO_EXTEND and SIGN_EXTEND better.
1965 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1966 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1968 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1971 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1972 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1974 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1975 costs when costing loads and stores to memory.
1977 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1978 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1980 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1983 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1985 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1987 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1988 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1990 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1992 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1994 (aarch64_rtx_mult_cost): New.
1995 (aarch64_rtx_costs): Use it, refactor as appropriate.
1997 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1998 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2000 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2001 emit instructions, return number of instructions which would
2003 (aarch64_add_constant): Update call to aarch64_build_constant.
2004 (aarch64_output_mi_thunk): Likewise.
2005 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2008 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2010 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2011 (TARGET_RTX_COSTS): Call it.
2013 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2015 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2016 (cortexa57_vector_cost): Likewise.
2017 (cortexa57_tunings): Use them.
2019 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2021 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2022 (cpu_addrcost_table): Use it.
2023 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2024 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2027 2014-05-16 Richard Biener <rguenther@suse.de>
2029 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2030 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2031 (visit_phi): Ignore edges marked as not executable.
2032 (class cond_dom_walker): New.
2033 (cond_dom_walker::before_dom_children): Value-number
2034 control statements and mark successor edges as not
2035 executable if possible.
2036 (run_scc_vn): First walk all control statements in
2037 dominator order, marking edges as not executable.
2038 * tree-inline.c (copy_edges_for_bb): Be not confused
2039 about random edge flags.
2041 2014-05-16 Richard Biener <rguenther@suse.de>
2043 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2045 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2048 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2049 (__TM_simple_begin): Use it.
2050 (__TM_begin): Likewise.
2052 2014-05-15 Martin Jambor <mjambor@suse.cz>
2055 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2056 type_preserved flag when the indirect edge is polymorphic.
2058 2014-05-15 Martin Jambor <mjambor@suse.cz>
2060 PR tree-optimization/61090
2061 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2062 build_ref_for_model.
2064 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2066 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2067 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2069 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2071 PR tree-optimization/61158
2072 * fold-const.c (fold_binary_loc): If X is zero-extended and
2073 shiftc >= prec, make sure zerobits is all ones instead of
2074 invoking undefined behavior.
2076 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2078 * regcprop.h: New file.
2079 * regcprop.c (skip_debug_insn_p): New decl.
2080 (replace_oldest_value_reg): Check skip_debug_insn_p.
2081 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2082 * shrink-wrap.c: Include regcprop.h.
2083 (prepare_shrink_wrap): Call
2084 copyprop_hardreg_forward_bb_without_debug_insn.
2086 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2088 * shrink-wrap.h: Update comment.
2089 * shrink-wrap.c: Update comment.
2090 (next_block_for_reg): Rename to live_edge_for_reg.
2091 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2092 (move_insn_for_shrink_wrap): Split live_edge.
2093 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2095 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2097 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2099 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2100 * config/sparc/sparc.md (fptype_ut699): New attribute.
2101 (in_branch_delay): Return false if -mfix-ut699 is specified and
2102 fptype_ut699 is set to single.
2103 (truncdfsf2): Add fptype_ut699 attribute.
2104 (fix_truncdfsi2): Likewise.
2105 (floatsisf2): Change fptype attribute.
2106 (fix_truncsfsi2): Likewise.
2107 (negtf2_notv9): Delete.
2108 (negtf2_v9): Likewise.
2109 (negtf2_hq): New instruction.
2110 (negtf2): New instruction and splitter.
2111 (negdf2_notv9): Rewrite.
2112 (abstf2_notv9): Delete.
2113 (abstf2_hq_v9): Likewise.
2114 (abstf2_v9): Likewise.
2115 (abstf2_hq): New instruction.
2116 (abstf2): New instruction and splitter.
2117 (absdf2_notv9): Rewrite.
2119 2014-05-14 Cary Coutant <ccoutant@google.com>
2122 * opts.c (common_handle_option): Don't special-case "-g".
2123 (set_debug_level): Default to at least level 2 with "-g".
2125 2014-05-14 DJ Delorie <dj@redhat.com>
2127 * config/msp430/msp430.c (msp430_builtin): Add
2128 MSP430_BUILTIN_DELAY_CYCLES.
2129 (msp430_init_builtins): Register void __delay_cycles(long long).
2130 (msp430_builtin_decl): Add it.
2131 (cg_magic_constant): New.
2132 (msp430_expand_delay_cycles): New.
2133 (msp430_expand_builtin): Call it.
2134 (msp430_print_operand_raw): Change integer printing from "int" to
2136 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2137 (delay_cycles_start): New.
2138 (delay_cycles_end): New.
2139 (delay_cycles_32): New.
2140 (delay_cycles_32x): New.
2141 (delay_cycles_16): New.
2142 (delay_cycles_16x): New.
2143 (delay_cycles_2): New.
2144 (delay_cycles_1): New.
2145 * doc/extend.texi: Document __delay_cycles().
2147 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2149 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2150 length attribute computation.
2152 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2155 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2157 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2160 * config/sparc/sparc.md: Fix types of low and high in DI constant
2161 splitter. Use gen_int_mode in some other splitters.
2163 2014-05-14 Martin Jambor <mjambor@suse.cz>
2166 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2168 2014-05-14 James Norris <jnorris@codesourcery.com>
2170 * omp-low.c (expand_parallel_call): Remove shadow variable.
2171 (expand_omp_taskreg): Likewise.
2173 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2175 * common/config/i386/i386-common.c
2176 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2177 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2178 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2179 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2180 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2181 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2182 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2183 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2184 xsavecintrin.h, xsavesintrin.h.
2185 (x86_64-*-*): Ditto.
2186 * config/i386/clflushoptintrin.h: New.
2187 * config/i386/xsavecintrin.h: Ditto.
2188 * config/i386/xsavesintrin.h: Ditto.
2189 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2190 (bit_XSAVES): Ditto.
2191 (bit_XSAVES): Ditto.
2192 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2193 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2195 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2196 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2197 OPTION_MASK_ISA_XSAVES.
2198 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2200 (PTA_CLFLUSHOPT) Define.
2201 (PTA_XSAVEC): Ditto.
2202 (PTA_XSAVES): Ditto.
2203 (ix86_option_override_internal): Handle new options.
2204 (ix86_valid_target_attribute_inner_p): Ditto.
2205 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2206 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2207 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2208 (bdesc_special_args): Add __builtin_ia32_xsaves,
2209 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2210 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2211 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2212 (ix86_expand_builtin): Handle new builtins.
2213 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2214 (TARGET_CLFLUSHOPT_P): Ditto.
2215 (TARGET_XSAVEC): Ditto.
2216 (TARGET_XSAVEC_P): Ditto.
2217 (TARGET_XSAVES): Ditto.
2218 (TARGET_XSAVES_P): Ditto.
2219 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2220 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2221 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2223 (ANY_XRSTOR64): Ditto.
2225 (xrstor): Change into <xrstor>.
2226 (xrstor_rex64): Change into <xrstor>_rex64.
2227 (xrstor64): Change into <xrstor>64
2229 * config/i386/i386.opt (mclflushopt): New.
2232 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2234 * doc/invoke.texi: Document new options.
2236 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2238 PR rtl-optimization/60866
2239 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2240 Default it to -1. Pass it down to init_simplejump_data.
2241 (init_simplejump_data): New parameter old_seqno. Pass it down
2242 to get_seqno_for_a_jump.
2243 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2244 initializing new jump seqno as a last resort. Add comment.
2245 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2246 jump and pass it down to sel_init_new_insn.
2247 (sel_redirect_edge_and_branch_force): Likewise.
2249 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2251 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2252 shifted values to avoid build warning.
2254 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2256 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2257 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2258 (cfg_layout_merge_blocks): Likewise.
2259 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2261 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2263 PR rtl-optimization/60901
2264 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2265 bb predecessor belongs to the same scheduling region. Adjust comment.
2267 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2269 * doc/sourcebuild.texi: (dfp_hw): Document.
2270 (p8vector_hw): Likewise.
2271 (powerpc_eabi_ok): Likewise.
2272 (powerpc_elfv2): Likewise.
2273 (powerpc_htm_ok): Likewise.
2274 (ppc_recip_hw): Likewise.
2277 2014-05-13 Cary Coutant <ccoutant@google.com>
2279 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2281 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2283 * gengtype-parse.c (require3): Eliminate in favor of...
2285 (require_template_declaration): Update to support optional single *
2288 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2289 (create_user_defined_type): Handle a single level of explicit
2290 pointerness within template arguments.
2291 (struct write_types_data): Add field "kind".
2292 (filter_type_name): Handle "*" character.
2293 (write_user_func_for_structure_ptr): Require a write_types_data
2294 rather than just a prefix string, so that we can look up the kind
2295 of the wtd and use it as an index into wrote_user_func_for_ptr,
2296 ensuring that such functions are written at most once. Support
2297 subclasses by invoking the marking function of the ultimate base class.
2298 (write_user_func_for_structure_body): Require a write_types_data
2299 rather than just a prefix string, so that we can pass this to
2300 write_user_func_for_structure_ptr.
2301 (write_func_for_structure): Likewise.
2302 (ggc_wtd): Add initializer of new "kind" field.
2303 (pch_wtd): Likewise.
2305 * gengtype.h (enum write_types_kinds): New.
2306 (struct type): Add field wrote_user_func_for_ptr to the "s"
2309 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2311 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2312 instead of const_binop.
2313 (fold_binary_loc): Likewise.
2315 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2317 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2318 calculation to match get_ref_base_and_extent.
2320 2014-05-13 Catherine Moore <clm@codesourcery.com>
2321 Sandra Loosemore <sandra@codesourcery.com>
2323 * configure.ac: Fix assembly for explicit JALR relocation check.
2324 * configure: Regenerate.
2326 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2328 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2329 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2330 Remove associated type declarations and initialisations.
2331 (arm_expand_neon_builtin): Likewise.
2332 (neon_emit_pair_result_insn): Delete.
2333 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2334 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2335 (neon_vzip<mode>): Likewise.
2336 (neon_vuzp<mode>): Likewise.
2338 2014-05-13 Richard Biener <rguenther@suse.de>
2341 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2342 it needs revisiting whether the call still may be tail-called.
2344 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2346 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2347 * rtl.h (block_symbol): Reduce number of fields to 2.
2348 (rtx_def): Add u2.symbol_ref_flags.
2349 (SYMBOL_REF_FLAGS): Use it.
2350 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2351 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2352 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2353 Lower index of SYMBOL_REF_DATA.
2354 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2355 Print SYMBOL_REF_FLAGS at the same time.
2356 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2358 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2360 * rtl.def (VAR_LOCATION): Remove "i" field.
2361 * rtl.h (rtx_def): Add u2.var_location_status.
2362 (PAT_VAR_LOCATION_STATUS): Use it.
2363 (gen_rtx_VAR_LOCATION): Declare.
2364 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2365 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2366 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2368 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2370 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2371 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2373 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2375 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2376 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2377 * rtl.h (rtx_def): Add insn_uid to u2 field.
2378 (RTX_FLAG_CHECK8): Delete in favor of...
2379 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2380 (INSN_DELETED_P): Update accordingly.
2381 (INSN_UID): Use u2.insn_uid.
2382 (INSN_CHAIN_CODE_P): Define.
2383 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2384 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2385 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2386 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2387 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2388 indices accordingly.
2389 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2390 Update indices for insn-chain rtxes.
2391 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2392 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2393 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2394 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2395 * combine.c (try_combine): Likewise.
2396 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2398 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2400 * rtl.def (REG): Remove middle field.
2401 * rtl.h (rtx_def): Add orignal_regno to u2.
2402 (ORIGINAL_REGNO): Use it instead of field 1.
2403 (REG_ATTRS): Lower field index accordingly.
2404 * gengtype.c (adjust_field_rtx_def): Remove handling of
2405 ORIGINAL_REGNO. Move REG_ATTRS index down.
2406 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2407 code that prints the REGNO.
2409 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2411 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2414 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2416 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2418 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2420 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2421 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2423 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2425 * config/aarch64/aarch64-protos.h
2426 (aarch64_hard_regno_caller_save_mode): New prototype.
2427 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2429 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2431 2014-05-13 Christian Bruel <christian.bruel@st.com>
2433 * target.def (mode_switching): New hook vector.
2434 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2435 (mode_exit, modepriority_to_mode): Likewise.
2436 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2437 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2438 * target.h: Include tm.h and hard-reg-set.h.
2439 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2440 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2441 * doc/tm.texi Regenerate.
2442 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2443 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2444 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2445 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2446 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2447 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2448 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2449 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2450 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2451 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2452 (ix86_emit_mode_set): Hookify.
2453 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2455 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2456 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2457 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2458 (epiphany_mode_priority_to_mode): Remove declaration.
2459 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2460 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2461 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2463 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2464 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2465 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2467 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2470 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2471 is const0_rtx, return immediately. Don't test count == 0 when
2474 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2476 * Makefile.in: add shrink-wrap.o.
2477 * config/i386/i386.c: include "shrink-wrap.h"
2478 * function.c: Likewise.
2479 (requires_stack_frame_p, next_block_for_reg,
2480 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2481 dup_block_and_redirect): Move to shrink-wrap.c
2482 (thread_prologue_and_epilogue_insns): Extract three code segments
2483 as functions in shrink-wrap.c
2484 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2486 * shrink-wrap.c: New file.
2487 * shrink-wrap.h: New file.
2489 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2491 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2492 reference to Solaris.
2494 2014-05-12 Mike Stump <mikestump@comcast.net>
2497 * genattrtab.c (filename): Add.
2498 (convert_set_attr_alternative): Improve error message.
2499 (check_defs): Restore read_md_filename for error messages.
2500 (gen_insn): Save filename.
2502 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2504 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2505 -fno-local-ivars and -fivar-visibility.
2506 * c-family/c.opt: Make -Wshadow also implicitly enable
2509 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2511 * doc/tm.texi: Remove reference to deleted macro.
2512 * doc/tm.texi.in: Likewise.
2514 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2517 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2520 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2523 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2524 * config/arm/aout.h (License): Same.
2525 * config/arm/bpabi.h (License): Same.
2526 * config/arm/elf.h (License): Same.
2527 * config/arm/linux-elf.h (License): Same.
2528 * config/arm/linux-gas.h (License): Same.
2529 * config/arm/netbsd-elf.h (License): Same.
2530 * config/arm/uclinux-eabi.h (License): Same.
2531 * config/arm/uclinux-elf.h (License): Same.
2532 * config/arm/vxworks.h (License): Same.
2534 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2536 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2537 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2538 number of operands to 3.
2539 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2540 * tree-nested.c (convert_nonlocal_omp_clauses,
2541 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2542 * gimplify.c (gimplify_scan_omp_clauses): Handle
2543 OMP_CLAUSE_LINEAR_STMT.
2544 * omp-low.c (lower_rec_input_clauses): Fix typo.
2545 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2546 cast between Fortran boolean_type_node and C _Bool if
2549 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2551 PR tree-optimization/61136
2552 * wide-int.h (multiple_of_p): Define a version that doesn't return
2554 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2555 integer_zerop/const_binop pair.
2556 (multiple_of_p): Likewise, converting both operands to widest_int
2559 2014-05-09 Teresa Johnson <tejohnson@google.com>
2561 * cgraphunit.c (analyze_functions): Use correct dump file.
2563 2014-05-09 Florian Weimer <fweimer@redhat.com>
2565 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2567 (stack_protect_return_slot_p): New function.
2568 (expand_used_vars): Call stack_protect_decl_p and
2569 stack_protect_return_slot_p for -fstack-protector-strong.
2571 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2574 * fold-const.c (fold_binary_loc): Changed width of mask.
2576 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2578 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2579 unsigned int initializers for regno_in, regno_out.
2581 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2584 * config/avr/avr.md (cc): Add new attribute set_vzn.
2585 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2586 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2587 with INC, DEC or NEG.
2588 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2589 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2590 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2592 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2595 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2597 * wide-int.cc (UTItype): Define.
2598 (UDWtype): Define for appropriate W_TYPE_SIZE.
2600 2014-05-09 Richard Biener <rguenther@suse.de>
2602 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2603 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2604 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2605 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2606 ssa_propagate): Adjust.
2608 2014-05-08 Jeff Law <law@redhat.com>
2610 PR tree-optimization/61009
2611 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2612 tri-state rather than a boolean. When a block is too big to
2613 thread through, inform caller via negative return value.
2614 (thread_across_edge): If a block was too big for normal threading,
2615 then it's too big for a joiner too, so remove temporary equivalences
2616 and return immediately.
2618 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2619 Matthias Klose <doko@ubuntu.com>
2622 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2624 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2627 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2629 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2632 * config/alpha/alpha.c: Include gimple-iterator.h.
2633 (alpha_gimple_fold_builtin): New function. Move
2634 ALPHA_BUILTIN_UMULH folding from ...
2635 (alpha_fold_builtin): ... here.
2636 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2638 2014-05-08 Wei Mi <wmi@google.com>
2641 * config/i386/i386.c (ix86_compute_frame_layout): Update
2642 preferred_stack_boundary for call, expanded from tls descriptor.
2643 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2644 to depend on SP register.
2645 (*tls_local_dynamic_base_32_gnu): Ditto.
2646 (*tls_local_dynamic_32_once): Ditto.
2647 (tls_global_dynamic_64_<mode>): Set
2648 ix86_tls_descriptor_calls_expanded_in_cfun.
2649 (tls_local_dynamic_base_64_<mode>): Ditto.
2650 (tls_global_dynamic_32): Set
2651 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2652 to depend on SP register.
2653 (tls_local_dynamic_base_32): Ditto.
2655 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2657 * config/arm/arm_neon.h: Update comment.
2658 * config/arm/neon-docgen.ml: Delete.
2659 * config/arm/neon-gen.ml: Delete.
2660 * doc/arm-neon-intrinsics.texi: Update comment.
2662 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2664 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2666 (vand, vorr, veor, vorn, vbic): Remove.
2667 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2669 (neon_vsub_unspec): Likewise.
2670 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2672 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2674 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2675 (vadd_s16): Likewise.
2676 (vadd_s32): Likewise.
2677 (vadd_f32): Likewise.
2678 (vadd_u8): Likewise.
2679 (vadd_u16): Likewise.
2680 (vadd_u32): Likewise.
2681 (vadd_s64): Likewise.
2682 (vadd_u64): Likewise.
2683 (vaddq_s8): Likewise.
2684 (vaddq_s16): Likewise.
2685 (vaddq_s32): Likewise.
2686 (vaddq_s64): Likewise.
2687 (vaddq_f32): Likewise.
2688 (vaddq_u8): Likewise.
2689 (vaddq_u16): Likewise.
2690 (vaddq_u32): Likewise.
2691 (vaddq_u64): Likewise.
2692 (vmul_s8): Likewise.
2693 (vmul_s16): Likewise.
2694 (vmul_s32): Likewise.
2695 (vmul_f32): Likewise.
2696 (vmul_u8): Likewise.
2697 (vmul_u16): Likewise.
2698 (vmul_u32): Likewise.
2699 (vmul_p8): Likewise.
2700 (vmulq_s8): Likewise.
2701 (vmulq_s16): Likewise.
2702 (vmulq_s32): Likewise.
2703 (vmulq_f32): Likewise.
2704 (vmulq_u8): Likewise.
2705 (vmulq_u16): Likewise.
2706 (vmulq_u32): Likewise.
2707 (vsub_s8): Likewise.
2708 (vsub_s16): Likewise.
2709 (vsub_s32): Likewise.
2710 (vsub_f32): Likewise.
2711 (vsub_u8): Likewise.
2712 (vsub_u16): Likewise.
2713 (vsub_u32): Likewise.
2714 (vsub_s64): Likewise.
2715 (vsub_u64): Likewise.
2716 (vsubq_s8): Likewise.
2717 (vsubq_s16): Likewise.
2718 (vsubq_s32): Likewise.
2719 (vsubq_s64): Likewise.
2720 (vsubq_f32): Likewise.
2721 (vsubq_u8): Likewise.
2722 (vsubq_u16): Likewise.
2723 (vsubq_u32): Likewise.
2724 (vsubq_u64): Likewise.
2725 (vand_s8): Likewise.
2726 (vand_s16): Likewise.
2727 (vand_s32): Likewise.
2728 (vand_u8): Likewise.
2729 (vand_u16): Likewise.
2730 (vand_u32): Likewise.
2731 (vand_s64): Likewise.
2732 (vand_u64): Likewise.
2733 (vandq_s8): Likewise.
2734 (vandq_s16): Likewise.
2735 (vandq_s32): Likewise.
2736 (vandq_s64): Likewise.
2737 (vandq_u8): Likewise.
2738 (vandq_u16): Likewise.
2739 (vandq_u32): Likewise.
2740 (vandq_u64): Likewise.
2741 (vorr_s8): Likewise.
2742 (vorr_s16): Likewise.
2743 (vorr_s32): Likewise.
2744 (vorr_u8): Likewise.
2745 (vorr_u16): Likewise.
2746 (vorr_u32): Likewise.
2747 (vorr_s64): Likewise.
2748 (vorr_u64): Likewise.
2749 (vorrq_s8): Likewise.
2750 (vorrq_s16): Likewise.
2751 (vorrq_s32): Likewise.
2752 (vorrq_s64): Likewise.
2753 (vorrq_u8): Likewise.
2754 (vorrq_u16): Likewise.
2755 (vorrq_u32): Likewise.
2756 (vorrq_u64): Likewise.
2757 (veor_s8): Likewise.
2758 (veor_s16): Likewise.
2759 (veor_s32): Likewise.
2760 (veor_u8): Likewise.
2761 (veor_u16): Likewise.
2762 (veor_u32): Likewise.
2763 (veor_s64): Likewise.
2764 (veor_u64): Likewise.
2765 (veorq_s8): Likewise.
2766 (veorq_s16): Likewise.
2767 (veorq_s32): Likewise.
2768 (veorq_s64): Likewise.
2769 (veorq_u8): Likewise.
2770 (veorq_u16): Likewise.
2771 (veorq_u32): Likewise.
2772 (veorq_u64): Likewise.
2773 (vbic_s8): Likewise.
2774 (vbic_s16): Likewise.
2775 (vbic_s32): Likewise.
2776 (vbic_u8): Likewise.
2777 (vbic_u16): Likewise.
2778 (vbic_u32): Likewise.
2779 (vbic_s64): Likewise.
2780 (vbic_u64): Likewise.
2781 (vbicq_s8): Likewise.
2782 (vbicq_s16): Likewise.
2783 (vbicq_s32): Likewise.
2784 (vbicq_s64): Likewise.
2785 (vbicq_u8): Likewise.
2786 (vbicq_u16): Likewise.
2787 (vbicq_u32): Likewise.
2788 (vbicq_u64): Likewise.
2789 (vorn_s8): Likewise.
2790 (vorn_s16): Likewise.
2791 (vorn_s32): Likewise.
2792 (vorn_u8): Likewise.
2793 (vorn_u16): Likewise.
2794 (vorn_u32): Likewise.
2795 (vorn_s64): Likewise.
2796 (vorn_u64): Likewise.
2797 (vornq_s8): Likewise.
2798 (vornq_s16): Likewise.
2799 (vornq_s32): Likewise.
2800 (vornq_s64): Likewise.
2801 (vornq_u8): Likewise.
2802 (vornq_u16): Likewise.
2803 (vornq_u32): Likewise.
2804 (vornq_u64): Likewise.
2806 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2808 * wide-int.cc (UTItype): Define.
2809 (UDWtype): Define for appropriate W_TYPE_SIZE.
2811 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2813 PR tree-optimization/59100
2814 * tree-ssa-phiopt.c: Include tree-inline.h.
2815 (neutral_element_p, absorbing_element_p): New functions.
2816 (value_replacement): Handle conditional binary operations with a
2817 neutral or absorbing element.
2819 2014-05-08 Richard Biener <rguenther@suse.de>
2821 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2822 folding the expression.
2823 (valueize_expr): Remove.
2824 (visit_reference_op_load): Do not valueize the result of
2826 (simplify_binary_expression): Likewise.
2827 (simplify_unary_expression): Likewise.
2829 2014-05-08 Richard Biener <rguenther@suse.de>
2831 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2832 looking at TYPE_ARG_TYPES.
2834 2014-05-08 Richard Biener <rguenther@suse.de>
2836 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2837 pointer propagation special-case.
2839 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2841 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2842 core part of address expressions.
2844 2014-05-08 Alan Modra <amodra@gmail.com>
2847 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2848 loads and stores when -mno-strict-align at any alignment.
2849 (expand_block_clear): Similarly. Also correct calculation of
2852 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2855 * tree-complex.c (expand_complex_move): Keep line info when expanding
2857 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2858 of complex expression. Use new argument to display correct location
2859 for values coming from phi statement.
2860 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2861 (warn_uninitialized_phi): Pass location of phi argument to
2863 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2864 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2866 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2868 * config/rs6000/predicates.md (indexed_address_mem): New.
2869 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2870 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2871 fpstore_ux, fpstore_u.
2872 (sign_extend, indexed, update): New.
2873 (cell_micro): Adjust.
2874 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2875 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2876 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2877 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2878 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2879 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2880 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2881 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2882 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2883 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2884 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2885 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2886 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2887 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2888 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2890 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2891 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2892 *vsx_extract_<mode>_store): Adjust.
2893 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2894 is_cracked_insn, insn_must_be_first_in_group,
2895 insn_must_be_last_in_group): Adjust.
2897 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2899 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2900 ppc440-fpstore): Adjust.
2901 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2902 ppc476-fpstore): Adjust.
2903 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2904 ppc601-fpstore): Adjust.
2905 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2907 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2909 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2910 ppc7450-fpstore): Adjust.
2911 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2912 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2913 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2915 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2916 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2917 cell-fpstore, cell-fpstore-update): Adjust.
2918 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2919 ppce300c3_store, ppce300c3_fpstore): Adjust.
2920 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2921 e500mc_fpstore): Adjust.
2922 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2923 e500mc64_store, e500mc64_fpstore): Adjust.
2924 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2925 e5500_fpstore): Adjust.
2926 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2927 e6500_fpstore): Adjust.
2928 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2930 * config/rs6000/power4.md (power4-load, power4-load-ext,
2931 power4-load-ext-update, power4-load-ext-update-indexed,
2932 power4-load-update-indexed, power4-load-update, power4-fpload,
2933 power4-fpload-update, power4-store, power4-store-update,
2934 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2936 * config/rs6000/power5.md (power5-load, power5-load-ext,
2937 power5-load-ext-update, power5-load-ext-update-indexed,
2938 power5-load-update-indexed, power5-load-update, power5-fpload,
2939 power5-fpload-update, power5-store, power5-store-update,
2940 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2942 * config/rs6000/power6.md (power6-load, power6-load-ext,
2943 power6-load-update, power6-load-update-indexed,
2944 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2945 power6-fpload-update, power6-store, power6-store-update,
2946 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2948 * config/rs6000/power7.md (power7-load, power7-load-ext,
2949 power7-load-update, power7-load-update-indexed,
2950 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2951 power7-fpload-update, power7-store, power7-store-update,
2952 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2954 * config/rs6000/power8.md (power8-load, power8-load-update,
2955 power8-load-ext, power8-load-ext-update, power8-fpload,
2956 power8-fpload-update, power8-store, power8-store-update-indexed,
2957 power8-fpstore, power8-fpstore-update): Adjust.
2958 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2960 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2961 titan_lsu_store, titan_lsu_fpstore): Adjust.
2962 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2964 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2967 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2968 unrolled byte insns. Emit address increments after move insns.
2970 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2972 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2973 const_gimple, rather than a gimple.
2974 (gimple_call_builtin_p): Likewise, for the three variants.
2976 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2977 (gimple_call_builtin_p): Likewise, for the three variants.
2979 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2981 PR tree-optimization/61095
2982 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2984 2014-05-07 Richard Biener <rguenther@suse.de>
2986 PR tree-optimization/61034
2987 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2988 (maybe_skip_until): Use translate to take into account
2989 lattices when trying to do disambiguations.
2990 (get_continuation_for_phi_1): Likewise.
2991 (get_continuation_for_phi): Adjust for added translate arguments.
2992 (walk_non_aliased_vuses): Likewise.
2993 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2994 (walk_non_aliased_vuses): Likewise.
2995 (call_may_clobber_ref_p_1): Declare.
2996 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2997 calls. Stop early if we are only supposed to disambiguate.
2998 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3000 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3002 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3003 Emit an error when the function has arguments.
3005 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3007 * cfgloop.h (unswitch_loops): Remove.
3008 * doc/passes.texi: Remove references to loop-unswitch.c
3009 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3011 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3013 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3014 check for loads group of length 3.
3015 (vect_permute_load_chain): New permutations for loads group of
3017 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3018 of vec_perm_shuffle for the new permutations.
3020 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3022 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3023 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3024 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3025 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3026 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3027 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3028 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3029 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3031 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3033 * loop-unswitch.c: Delete.
3035 2014-05-07 Richard Biener <rguenther@suse.de>
3037 * config.gcc: Always set need_64bit_hwint to yes.
3039 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3041 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3042 of using optimize_size.
3044 2014-05-06 Mike Stump <mikestump@comcast.net>
3046 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3048 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3050 * config/i386/sse.md (*mov<mode>_internal)
3051 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3052 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3053 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3054 (*<code><mode>3, *andnot<mode>3<mask_name>)
3055 (<mask_codefor><code><mode>3<mask_name>): Only consider
3056 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3058 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3061 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3063 * lra-constraints.c (valid_address_p): Move earlier in file.
3064 Add a constraint argument to the address_info version.
3065 (satisfies_memory_constraint_p): New function.
3066 (satisfies_address_constraint_p): Likewise.
3067 (process_alt_operands, curr_insn_transform): Use them.
3068 (process_address): Pass the constraint to valid_address_p when
3069 checking address operands.
3071 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3073 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3074 to their respective blocks. Fix inadvertent use of "node".
3076 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3078 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3080 (init_emit_once): ...here.
3081 * rtl.h (init_derived_machine_modes): Declare.
3082 * toplev.c (do_compile): Call it even if no_backend.
3084 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3085 Mike Stump <mikestump@comcast.net>
3086 Richard Sandiford <rdsandiford@googlemail.com>
3087 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3089 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3090 (rtx_equal_for_memref_p): Update comment.
3091 (adjust_offset_for_component_ref): Use wide-int interfaces.
3092 * builtins.c (get_object_alignment_2): Likewise.
3093 (c_readstr): Likewise.
3094 (target_char_cast): Add comment.
3095 (determine_block_size): Use wide-int interfaces.
3096 (expand_builtin_signbit): Likewise.
3097 (fold_builtin_int_roundingfn): Likewise.
3098 (fold_builtin_bitop): Likewise.
3099 (fold_builtin_bswap): Likewise.
3100 (fold_builtin_logarithm): Use signop.
3101 (fold_builtin_pow): Likewise.
3102 (fold_builtin_memory_op): Use wide-int interfaces.
3103 (fold_builtin_object_size): Likewise.
3104 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3105 nb_iterations_estimate.
3106 (record_niter_bound): Use wide-int interfaces.
3107 (get_estimated_loop_iterations_int): Likewise.
3108 (get_estimated_loop_iterations): Likewise.
3109 (get_max_loop_iterations): Likewise.
3110 * cfgloop.h: Include wide-int.h.
3111 (struct nb_iter_bound): Change bound to widest_int.
3112 (struct loop): Change nb_iterations_upper_bound and
3113 nb_iterations_estimate to widest_int.
3114 (record_niter_bound): Switch to use widest_int.
3115 (get_estimated_loop_iterations): Likewise.
3116 (get_max_loop_iterations): Likewise.
3117 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3118 update for wide-int.
3119 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3120 * combine.c (try_combine): Likewise.
3121 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3122 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3124 (aarch64_float_const_representable_p): Likewise.
3125 * config/arc/arc.c: Include wide-int.h.
3126 (arc_can_use_doloop_p): Use wide-int interfaces.
3127 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3128 (vfp3_const_double_index): Likewise.
3129 * config/avr/avr.c (avr_out_round): Likewise.
3130 (avr_fold_builtin): Likewise.
3131 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3132 (bfin_can_use_doloop_p): Likewise.
3133 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3134 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3135 * config/i386/i386.c: Include wide-int.h.
3136 (ix86_data_alignment): Use wide-int interfaces.
3137 (ix86_local_alignment): Likewise.
3138 (ix86_emit_swsqrtsf): Update real_from_integer.
3139 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3140 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3141 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3142 (zero_constant): Likewise.
3143 (input_operand): Likewise.
3144 (splat_input_operand): Likewise.
3145 (non_logical_cint_operand): Change const_double to const_wide_int.
3146 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3147 (easy_altivec_constant): Remove comment.
3148 (paired_expand_vector_init): Use CONSTANT_P.
3149 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3150 (rs6000_emit_move): Update checks.
3151 (rs6000_aggregate_candidate): Use wide-int interfaces.
3152 (rs6000_expand_ternop_builtin): Likewise.
3153 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3154 (rs6000_assemble_integer): Likewise.
3155 (rs6000_hash_constant): Likewise.
3156 (output_toc): Likewise.
3157 (rs6000_rtx_costs): Likewise.
3158 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3159 * config/rs6000/rs6000-c.c: Include wide-int.h.
3160 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3161 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3162 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3163 Handle CONST_WIDE_INT.
3164 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3165 Use tree_fits_uhwi_p.
3166 * config/sparc/sparc.c: Include wide-int.h.
3167 (sparc_fold_builtin): Use wide-int interfaces.
3168 * config/vax/vax.c: Include wide-int.h.
3169 (vax_float_literal): Use real_from_integer.
3170 * coretypes.h (struct hwivec_def): New.
3172 (const_hwivec): New.
3173 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3174 (equiv_constant): Handle CONST_WIDE_INT.
3175 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3176 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3177 * dbxout.c (stabstr_U): Use wide-int interfaces.
3178 (dbxout_type): Update to use cst_fits_shwi_p.
3179 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3180 (TARGET_SUPPORTS_WIDE_INT): Add default.
3181 * dfp.c: Include wide-int.h.
3182 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3183 decimal_real_to_integer.
3184 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3185 decimal_real_to_integer.
3186 * doc/generic.texi (Constant expressions): Update for wide_int.
3187 * doc/rtl.texi (const_double): Likewise.
3188 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3189 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3190 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3191 (REAL_VALUE_FROM_INT): Remove.
3192 (TARGET_SUPPORTS_WIDE_INT): New.
3193 * doc/tm.texi: Regenerate.
3194 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3195 * double-int.h: Include wide-int.h.
3196 (struct wi::int_traits): New.
3197 * dwarf2out.c (get_full_len): New.
3198 (dw_val_equal_p): Add case dw_val_class_wide_int.
3199 (size_of_loc_descr): Likewise.
3200 (output_loc_operands): Likewise.
3201 (insert_double): Remove.
3202 (insert_wide_int): New.
3204 (print_die): Add case dw_val_class_wide_int.
3205 (attr_checksum): Likewise.
3206 (attr_checksum_ordered): Likewise.
3207 (same_dw_val_p): Likewise.
3208 (size_of_die): Likewise.
3209 (value_format): Likewise.
3210 (output_die): Likewise.
3211 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3213 (clz_loc_descriptor): Use wide-int interfaces.
3214 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3215 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3216 (round_up_to_align): Use wide-int interfaces.
3217 (field_byte_offset): Likewise.
3218 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3219 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3220 CONST_DOUBLE handling. Use wide-int interfaces.
3221 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3222 (gen_enumeration_type_die): Use add_AT_wide.
3223 (hash_loc_operands): Add case dw_val_class_wide_int.
3224 (compare_loc_operands): Likewise.
3225 * dwarf2out.h: Include wide-int.h.
3226 (wide_int_ptr): New.
3227 (enum dw_val_class): Add dw_val_class_wide_int.
3228 (struct dw_val_struct): Add val_wide.
3229 * emit-rtl.c (const_wide_int_htab): New.
3230 (const_wide_int_htab_hash): New.
3231 (const_wide_int_htab_eq): New.
3232 (lookup_const_wide_int): New.
3233 (const_double_htab_hash): Use wide-int interfaces.
3234 (const_double_htab_eq): Likewise.
3235 (rtx_to_double_int): Conditionally compile for wide-int.
3236 (immed_double_int_const): Rename to immed_wide_int_const and
3237 update for wide-int.
3238 (immed_double_const): Conditionally compile for wide-int.
3239 (init_emit_once): Use wide-int interfaces.
3240 * explow.c (plus_constant): Likewise.
3241 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3242 (lshift_value): Use wide-int interfaces.
3243 (expand_mult): Likewise.
3244 (choose_multiplier): Likewise.
3245 (expand_smod_pow2): Likewise.
3246 (make_tree): Likewise.
3247 * expr.c (convert_modes): Consolidate handling of constants.
3248 Use wide-int interfaces.
3249 (emit_group_load_1): Add note.
3250 (store_expr): Update comment.
3251 (get_inner_reference): Use wide-int interfaces.
3252 (expand_constructor): Update comment.
3253 (expand_expr_real_2): Use wide-int interfaces.
3254 (expand_expr_real_1): Likewise.
3255 (reduce_to_bit_field_precision): Likewise.
3256 (const_vector_from_tree): Likewise.
3257 * final.c: Include wide-int-print.h.
3258 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3259 * fixed-value.c: Include wide-int.h.
3260 (fixed_from_string): Use wide-int interfaces.
3261 (fixed_to_decimal): Likewise.
3262 (fixed_convert_from_real): Likewise.
3263 (real_convert_from_fixed): Likewise.
3264 * fold-const.h (mem_ref_offset): Return an offset_int.
3265 (div_if_zero_remainder): Remove code parameter.
3266 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3267 Use wide-int interfaces.
3268 (may_negate_without_overflow_p): Use wide-int interfaces.
3269 (negate_expr_p): Likewise.
3270 (fold_negate_expr): Likewise.
3271 (int_const_binop_1): Likewise.
3272 (const_binop): Likewise.
3273 (fold_convert_const_int_from_int): Likewise.
3274 (fold_convert_const_int_from_real): Likewise.
3275 (fold_convert_const_int_from_fixed): Likewise.
3276 (fold_convert_const_fixed_from_int): Likewise.
3277 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3278 (sign_bit_p): Use wide-int interfaces.
3279 (make_range_step): Likewise.
3280 (build_range_check): Likewise. Pass an integer of the correct type
3281 instead of using integer_one_node.
3282 (range_predecessor): Pass an integer of the correct type instead
3283 of using integer_one_node.
3284 (range_successor): Likewise.
3285 (merge_ranges): Likewise.
3286 (unextend): Use wide-int interfaces.
3287 (extract_muldiv_1): Likewise.
3288 (fold_div_compare): Likewise.
3289 (fold_single_bit_test): Likewise.
3290 (fold_sign_changed_comparison): Likewise.
3291 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3292 (fold_plusminus_mult_expr): Use wide-int interfaces.
3293 (native_encode_int): Likewise.
3294 (native_interpret_int): Likewise.
3295 (fold_unary_loc): Likewise.
3296 (pointer_may_wrap_p): Likewise.
3297 (size_low_cst): Likewise.
3298 (mask_with_tz): Likewise.
3299 (fold_binary_loc): Likewise.
3300 (fold_ternary_loc): Likewise.
3301 (multiple_of_p): Likewise.
3302 (tree_call_nonnegative_warnv_p): Update calls to
3303 tree_int_cst_min_precision and real_from_integer.
3304 (fold_negate_const): Use wide-int interfaces.
3305 (fold_abs_const): Likewise.
3306 (fold_relational_const): Use tree_int_cst_lt.
3307 (round_up_loc): Use wide-int interfaces.
3308 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3309 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3310 * gengtype.c: Remove include of double-int.h.
3311 (do_typedef): Use wide-int interfaces.
3312 (open_base_files): Add wide-int.h.
3313 (main): Add offset_int and widest_int typedefs.
3314 * gengtype-lex.l: Handle "^".
3315 (CXX_KEYWORD): Add "static".
3316 * gengtype-parse.c (require3): New.
3317 (require_template_declaration): Handle constant template arguments
3318 and nested templates.
3319 * gengtype-state.c: Don't include "double-int.h".
3320 * genpreds.c (write_one_predicate_function): Update comment.
3321 (write_tm_constrs_h): Add check for hval and lval use in
3323 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3324 (add_to_sequence): Likewise.
3325 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3326 and const_double_operand.
3327 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3329 * gimple-fold.c (get_base_constructor): Likewise.
3330 (fold_array_ctor_reference): Likewise.
3331 (fold_nonarray_ctor_reference): Likewise.
3332 (fold_const_aggregate_ref_1): Likewise.
3333 (gimple_val_nonnegative_real_p): Likewise.
3334 (gimple_fold_indirect_ref): Likewise.
3335 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3336 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3337 (struct slsr_cand_d): Change index to be widest_int.
3338 (struct incr_info_d): Change incr to be widest_int.
3339 (alloc_cand_and_find_basis): Use wide-int interfaces.
3340 (slsr_process_phi): Likewise.
3341 (backtrace_base_for_ref): Likewise. Return a widest_int.
3342 (restructure_reference): Take a widest_int instead of a double_int.
3343 (slsr_process_ref): Use wide-int interfaces.
3344 (create_mul_ssa_cand): Likewise.
3345 (create_mul_imm_cand): Likewise.
3346 (create_add_ssa_cand): Likewise.
3347 (create_add_imm_cand): Take a widest_int instead of a double_int.
3348 (slsr_process_add): Use wide-int interfaces.
3349 (slsr_process_cast): Likewise.
3350 (slsr_process_copy): Likewise.
3351 (dump_candidate): Likewise.
3352 (dump_incr_vec): Likewise.
3353 (replace_ref): Likewise.
3354 (cand_increment): Likewise. Return a widest_int.
3355 (cand_abs_increment): Likewise.
3356 (replace_mult_candidate): Take a widest_int instead of a double_int.
3357 (replace_unconditional_candidate): Use wide-int interfaces.
3358 (incr_vec_index): Take a widest_int instead of a double_int.
3359 (create_add_on_incoming_edge): Likewise.
3360 (create_phi_basis): Use wide-int interfaces.
3361 (replace_conditional_candidate): Likewise.
3362 (record_increment): Take a widest_int instead of a double_int.
3363 (record_phi_increments): Use wide-int interfaces.
3364 (phi_incr_cost): Take a widest_int instead of a double_int.
3365 (lowest_cost_path): Likewise.
3366 (total_savings): Likewise.
3367 (analyze_increments): Use wide-int interfaces.
3368 (ncd_with_phi): Take a widest_int instead of a double_int.
3369 (ncd_of_cand_and_phis): Likewise.
3370 (nearest_common_dominator_for_cands): Likewise.
3371 (insert_initializers): Use wide-int interfaces.
3372 (all_phi_incrs_profitable): Likewise.
3373 (replace_one_candidate): Likewise.
3374 (replace_profitable_candidates): Likewise.
3375 * godump.c: Include wide-int-print.h.
3376 (go_output_typedef): Use wide-int interfaces.
3377 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3378 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3379 (build_loop_iteration_domains): Likewise.
3380 * hooks.h: Include wide-int.h rather than double-int.h.
3381 (hook_bool_dint_dint_uint_bool_true): Delete.
3382 (hook_bool_wint_wint_uint_bool_true): Declare.
3383 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3384 (hook_bool_wint_wint_uint_bool_true): New.
3385 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3387 (ubsan_expand_si_overflow_mul_check): Likewise.
3388 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3389 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3390 (get_ancestor_addr_info): Likewise.
3391 (ipa_modify_call_arguments): Likewise.
3392 * loop-doloop.c (doloop_modify): Likewise.
3393 (doloop_optimize): Likewise.
3394 * loop-iv.c (iv_number_of_iterations): Likewise.
3395 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3396 (unroll_loop_constant_iterations): Likewise.
3397 (decide_unroll_runtime_iterations): Likewise.
3398 (unroll_loop_runtime_iterations): Likewise.
3399 (decide_peel_simple): Likewise.
3400 (decide_unroll_stupid): Likewise.
3401 * lto-streamer-in.c (streamer_read_wi): Add.
3402 (input_cfg): Use wide-int interfaces.
3403 (lto_input_tree_1): Likewise.
3404 * lto-streamer-out.c (streamer_write_wi): Add.
3405 (hash_tree): Use wide-int interfaces.
3406 (output_cfg): Likewise.
3407 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3408 (GTFILES): Add wide-int.h and signop.h.
3409 (TAGS): Look for .cc files too.
3410 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3411 * optabs.c (expand_subword_shift): Likewise.
3412 (expand_doubleword_shift): Likewise.
3413 (expand_absneg_bit): Likewise.
3414 (expand_copysign_absneg): Likewise.
3415 (expand_copysign_bit): Likewise.
3416 * postreload.c (reload_cse_simplify_set): Likewise.
3417 * predict.c (predict_iv_comparison): Likewise.
3418 * pretty-print.h: Include wide-int-print.h.
3420 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3421 * print-tree.c: Include wide-int-print.h.
3422 (print_node_brief): Use wide-int interfaces.
3423 (print_node): Likewise.
3424 * read-rtl.c (validate_const_wide_int): New.
3425 (read_rtx_code): Add CONST_WIDE_INT case.
3426 * real.c: Include wide-int.h.
3427 (real_to_integer2): Delete.
3428 (real_to_integer): New function, returning a wide_int.
3429 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3430 (ten_to_ptwo): Update call to real_from_integer.
3431 (real_digit): Likewise.
3432 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3433 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3434 (REAL_VALUE_TO_INT): Delete.
3435 (real_to_integer): Declare a wide-int form.
3436 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3437 * recog.c (const_int_operand): Improve comment.
3438 (const_scalar_int_operand): New.
3439 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3440 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3441 (split_double): Likewise.
3442 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3443 (rtx_size): Likewise.
3444 (rtx_alloc_stat_v): New.
3445 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3446 (cwi_output_hex): New.
3447 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3448 (cwi_check_failed_bounds): New.
3449 * rtl.def (CONST_WIDE_INT): New.
3450 * rtl.h: Include <utility> and wide-int.h.
3451 (struct hwivec_def): New.
3452 (CWI_GET_NUM_ELEM): New.
3453 (CWI_PUT_NUM_ELEM): New.
3454 (struct rtx_def): Add num_elem and hwiv.
3455 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3456 (CASE_CONST_UNIQUE): Likewise.
3457 (CASE_CONST_ANY): Likewise.
3458 (CONST_SCALAR_INT_P): Likewise.
3459 (CONST_WIDE_INT_P): New.
3461 (HWIVEC_CHECK): New.
3462 (cwi_check_failed_bounds): New.
3464 (HWIVEC_CHECK): New.
3465 (CONST_WIDE_INT_VEC) New.
3466 (CONST_WIDE_INT_NUNITS) New.
3467 (CONST_WIDE_INT_ELT) New.
3468 (rtx_mode_t): New type.
3469 (wi::int_traits <rtx_mode_t>): New.
3471 (wi::min_value): New.
3472 (wi::max_value): New.
3474 (const_wide_int_alloc): New.
3475 (immed_wide_int_const): New.
3476 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3477 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3478 * signop.h: New file.
3479 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3480 (simplify_const_unary_operation): Use wide-int interfaces.
3481 (simplify_binary_operation_1): Likewise.
3482 (simplify_const_binary_operation): Likewise.
3483 (simplify_const_relational_operation): Likewise.
3484 (simplify_immed_subreg): Likewise.
3485 * stmt.c (expand_case): Likewise.
3486 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3487 signop rather than a bool.
3488 * stor-layout.c (layout_type): Use wide-int interfaces.
3489 (initialize_sizetypes): Update calls to
3490 set_min_and_max_values_for_integral_type.
3491 (set_min_and_max_values_for_integral_type): Take a signop rather
3492 than a bool. Use wide-int interfaces.
3493 (fixup_signed_type): Update accordingly. Remove
3494 HOST_BITS_PER_DOUBLE_INT limit.
3495 (fixup_unsigned_type): Likewise.
3496 * system.h (STATIC_CONSTANT_P): New.
3497 (STATIC_ASSERT): New.
3498 * target.def (can_use_doloop_p): Take widest_ints rather than
3500 * target.h: Include wide-int.h rather than double-int.h.
3501 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3503 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3504 rather than INT_CST_LT_UNSIGNED.
3505 (can_use_doloop_if_innermost): Take widest_ints rather than
3507 * tree-affine.c: Include wide-int-print.h.
3508 (double_int_ext_for_comb): Delete.
3509 (wide_int_ext_for_comb): New.
3510 (aff_combination_zero): Use wide-int interfaces.
3511 (aff_combination_const): Take a widest_int instead of a double_int.
3512 (aff_combination_elt): Use wide-int interfaces.
3513 (aff_combination_scale): Take a widest_int instead of a double_int.
3514 (aff_combination_add_elt): Likewise.
3515 (aff_combination_add_cst): Likewise.
3516 (aff_combination_add): Use wide-int interfaces.
3517 (aff_combination_convert): Likewise.
3518 (tree_to_aff_combination): Likewise.
3519 (add_elt_to_tree): Take a widest_int instead of a double_int.
3520 (aff_combination_to_tree): Use wide-int interfaces.
3521 (aff_combination_remove_elt): Likewise.
3522 (aff_combination_add_product): Take a widest_int instead of
3524 (aff_combination_mult): Use wide-int interfaces.
3525 (aff_combination_expand): Likewise.
3526 (double_int_constant_multiple_p): Delete.
3527 (wide_int_constant_multiple_p): New.
3528 (aff_combination_constant_multiple_p): Take a widest_int pointer
3529 instead of a double_int pointer.
3530 (print_aff): Use wide-int interfaces.
3531 (get_inner_reference_aff): Take a widest_int pointer
3532 instead of a double_int pointer.
3533 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3534 * tree-affine.h: Include wide-int.h.
3535 (struct aff_comb_elt): Change type of coef to widest_int.
3536 (struct affine_tree_combination): Change type of offset to widest_int.
3537 (double_int_ext_for_comb): Delete.
3538 (wide_int_ext_for_comb): New.
3539 (aff_combination_const): Use widest_int instead of double_int.
3540 (aff_combination_scale): Likewise.
3541 (aff_combination_add_elt): Likewise.
3542 (aff_combination_constant_multiple_p): Likewise.
3543 (get_inner_reference_aff): Likewise.
3544 (aff_comb_cannot_overlap_p): Likewise.
3545 (aff_combination_zero_p): Use wide-int interfaces.
3546 * tree.c: Include tree.h.
3547 (init_ttree): Use make_int_cst.
3548 (tree_code_size): Removed code for INTEGER_CST case.
3549 (tree_size): Add INTEGER_CST case.
3550 (make_node_stat): Update comment.
3551 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3552 (build_int_cst_type): Use wide-int interfaces.
3553 (double_int_to_tree): Likewise.
3554 (double_int_fits_to_tree_p): Delete.
3555 (force_fit_type_double): Delete.
3556 (force_fit_type): New.
3557 (int_cst_hash_hash): Use wide-int interfaces.
3558 (int_cst_hash_eq): Likewise.
3559 (build_int_cst_wide): Delete.
3560 (wide_int_to_tree): New.
3561 (cache_integer_cst): Use wide-int interfaces.
3562 (build_low_bits_mask): Likewise.
3563 (cst_and_fits_in_hwi): Likewise.
3564 (real_value_from_int_cst): Likewise.
3565 (make_int_cst_stat): New.
3566 (integer_zerop): Use wide_int interfaces.
3567 (integer_onep): Likewise.
3568 (integer_all_onesp): Likewise.
3569 (integer_pow2p): Likewise.
3570 (integer_nonzerop): Likewise.
3571 (tree_log2): Likewise.
3572 (tree_floor_log2): Likewise.
3573 (tree_ctz): Likewise.
3574 (int_size_in_bytes): Likewise.
3575 (mem_ref_offset): Return an offset_int rather than a double_int.
3576 (build_type_attribute_qual_variant): Use wide_int interfaces.
3577 (type_hash_eq): Likewise
3578 (tree_int_cst_equal): Likewise.
3579 (tree_int_cst_lt): Delete.
3580 (tree_int_cst_compare): Likewise.
3581 (tree_fits_shwi_p): Use wide_int interfaces.
3582 (tree_fits_uhwi_p): Likewise.
3583 (tree_int_cst_sign_bit): Likewise.
3584 (tree_int_cst_sgn): Likewise.
3585 (tree_int_cst_min_precision): Take a signop rather than a bool.
3586 (simple_cst_equal): Use wide_int interfaces.
3587 (compare_tree_int): Likewise.
3588 (iterative_hash_expr): Likewise.
3589 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3591 (get_type_static_bounds): Use wide_int interfaces.
3592 (tree_int_cst_elt_check_failed): New.
3593 (build_common_tree_nodes): Reordered to set prec before filling in
3595 (int_cst_value): Check cst_and_fits_in_hwi.
3596 (widest_int_cst_value): Use wide_int interfaces.
3597 (upper_bound_in_type): Likewise.
3598 (lower_bound_in_type): Likewise.
3599 (num_ending_zeros): Likewise.
3600 (drop_tree_overflow): Likewise.
3601 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3602 (gen_conditions_for_pow_cst_base): Likewise.
3603 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3604 (group_case_labels_stmt): Use wide-int interfaces.
3605 (verify_gimple_assign_binary): Likewise.
3606 (print_loop): Likewise.
3607 * tree-chrec.c (tree_fold_binomial): Likewise.
3608 * tree-core.h (struct tree_base): Add int_length.
3609 (struct tree_int_cst): Change rep of value.
3610 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3611 (dr_may_alias_p): Likewise.
3612 (max_stmt_executions_tree): Likewise.
3613 * tree.def (INTEGER_CST): Update comment.
3614 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3615 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3616 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3617 (dequeue_and_dump): Use wide-int interfaces.
3618 * tree.h: Include wide-int.h.
3619 (NULL_TREE): Moved to earlier loc in file.
3620 (TREE_INT_CST_ELT_CHECK): New.
3621 (tree_int_cst_elt_check_failed): New.
3623 (TREE_INT_CST): Delete.
3624 (TREE_INT_CST_LOW): Use wide-int interfaces.
3625 (TREE_INT_CST_HIGH): Delete.
3626 (TREE_INT_CST_NUNITS): New.
3627 (TREE_INT_CST_EXT_NUNITS): Likewise.
3628 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3629 (TREE_INT_CST_ELT): Likewise.
3630 (INT_CST_LT): Delete.
3631 (tree_int_cst_elt_check): New (two forms).
3632 (type_code_size): Update comment.
3633 (make_int_cst_stat, make_int_cst): New.
3634 (tree_to_double_int): Delete.
3635 (double_int_fits_to_tree_p): Delete.
3636 (force_fit_type_double): Delete.
3637 (build_int_cstu): Replace with out-of-line function.
3638 (build_int_cst_wide): Delete.
3639 (tree_int_cst_lt): Define inline.
3640 (tree_int_cst_le): New.
3641 (tree_int_cst_compare): Define inline.
3642 (tree_int_cst_min_precision): Take a signop rather than a bool.
3643 (wi::int_traits <const_tree>): New.
3644 (wi::int_traits <tree>): New.
3645 (wi::extended_tree): New.
3646 (wi::int_traits <wi::extended_tree>): New.
3647 (wi::to_widest): New.
3648 (wi::to_offset): New.
3649 (wi::fits_to_tree_p): New.
3650 (wi::min_value): New.
3651 (wi::max_value): New.
3652 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3653 (copy_tree_body_r): Likewise.
3654 * tree-object-size.c (compute_object_offset): Likewise.
3655 (addr_object_size): Likewise.
3656 * tree-predcom.c: Include wide-int-print.h.
3657 (struct dref_d): Change type of offset to widest_int.
3658 (dump_dref): Call wide-int printer.
3659 (aff_combination_dr_offset): Use wide-int interfaces.
3660 (determine_offset): Take a widest_int pointer rather than a
3662 (split_data_refs_to_components): Use wide-int interfaces.
3663 (suitable_component_p): Likewise.
3664 (order_drefs): Likewise.
3665 (add_ref_to_chain): Likewise.
3666 (valid_initializer_p): Likewise.
3667 (determine_roots_comp): Likewise.
3668 * tree-pretty-print.c: Include wide-int-print.h.
3669 (dump_generic_node): Use wide-int interfaces.
3670 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3671 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3672 (move_fixed_address_to_symbol): Likewise.
3673 (move_hint_to_base): Likewise.
3674 (move_pointer_to_base): Likewise.
3675 (move_variant_to_index): Likewise.
3676 (most_expensive_mult_to_index): Likewise.
3677 (addr_to_parts): Likewise.
3678 (copy_ref_info): Likewise.
3679 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3680 (indirect_refs_may_alias_p): Likewise.
3681 (stmt_kills_ref_p_1): Likewise.
3682 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3683 * tree-ssa-ccp.c: Update comment at top of file. Include
3685 (struct prop_value_d): Change type of mask to widest_int.
3686 (extend_mask): New function.
3687 (dump_lattice_value): Use wide-int interfaces.
3688 (get_default_value): Likewise.
3689 (set_constant_value): Likewise.
3690 (set_value_varying): Likewise.
3691 (valid_lattice_transition): Likewise.
3692 (set_lattice_value): Likewise.
3693 (value_to_double_int): Delete.
3694 (value_to_wide_int): New.
3695 (get_value_from_alignment): Use wide-int interfaces.
3696 (get_value_for_expr): Likewise.
3697 (do_dbg_cnt): Likewise.
3698 (ccp_finalize): Likewise.
3699 (ccp_lattice_meet): Likewise.
3700 (bit_value_unop_1): Use widest_ints rather than double_ints.
3701 (bit_value_binop_1): Likewise.
3702 (bit_value_unop): Use wide-int interfaces.
3703 (bit_value_binop): Likewise.
3704 (bit_value_assume_aligned): Likewise.
3705 (evaluate_stmt): Likewise.
3706 (ccp_fold_stmt): Likewise.
3707 (visit_cond_stmt): Likewise.
3708 (ccp_visit_stmt): Likewise.
3709 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3710 (constant_pointer_difference): Likewise.
3711 (associate_pointerplus): Likewise.
3712 (combine_conversions): Likewise.
3713 * tree-ssa-loop.h: Include wide-int.h.
3714 (struct tree_niter_desc): Change type of max to widest_int.
3715 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3716 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3717 (remove_redundant_iv_tests): Likewise.
3718 (canonicalize_loop_induction_variables): Likewise.
3719 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3720 (constant_multiple_of): Take a widest_int pointer instead of
3721 a double_int pointer.
3722 (get_computation_aff): Use wide-int interfaces.
3723 (ptr_difference_cost): Likewise.
3724 (difference_cost): Likewise.
3725 (get_loop_invariant_expr_id): Likewise.
3726 (get_computation_cost_at): Likewise.
3727 (iv_elimination_compare_lt): Likewise.
3728 (may_eliminate_iv): Likewise.
3729 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3730 instead of double_int.
3731 (max_loop_iterations): Likewise.
3732 (max_stmt_executions): Likewise.
3733 (estimated_stmt_executions): Likewise.
3734 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3735 (split_to_var_and_offset): Use wide-int interfaces.
3736 (determine_value_range): Likewise.
3737 (bound_difference_of_offsetted_base): Likewise.
3738 (bounds_add): Take a widest_int instead of a double_int.
3739 (number_of_iterations_ne_max): Use wide-int interfaces.
3740 (number_of_iterations_ne): Likewise.
3741 (number_of_iterations_lt_to_ne): Likewise.
3742 (assert_loop_rolls_lt): Likewise.
3743 (number_of_iterations_lt): Likewise.
3744 (number_of_iterations_le): Likewise.
3745 (number_of_iterations_cond): Likewise.
3746 (number_of_iterations_exit): Likewise.
3747 (finite_loop_p): Likewise.
3748 (derive_constant_upper_bound_assign): Likewise.
3749 (derive_constant_upper_bound): Return a widest_int.
3750 (derive_constant_upper_bound_ops): Likewise.
3751 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3752 (record_estimate): Take a widest_int rather than a double_int.
3753 (record_nonwrapping_iv): Use wide-int interfaces.
3754 (double_int_cmp): Delete.
3755 (wide_int_cmp): New.
3756 (bound_index): Take a widest_int rather than a double_int.
3757 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3758 (maybe_lower_iteration_bound): Likewise.
3759 (estimate_numbers_of_iterations_loop): Likewise.
3760 (estimated_loop_iterations): Take a widest_int pointer than than
3761 a double_int pointer.
3762 (estimated_loop_iterations_int): Use wide-int interfaces.
3763 (max_loop_iterations): Take a widest_int pointer than than
3764 a double_int pointer.
3765 (max_loop_iterations_int): Use wide-int interfaces.
3766 (max_stmt_executions): Take a widest_int pointer than than
3767 a double_int pointer.
3768 (estimated_stmt_executions): Likewise.
3769 (n_of_executions_at_most): Use wide-int interfaces.
3770 (scev_probably_wraps_p): Likewise.
3771 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3773 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3775 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3776 double_ints. Adjust for trailing_wide_ints <3> representation.
3777 (set_nonzero_bits): Likewise.
3778 (get_range_info): Return wide_ints rather than double_ints.
3779 Adjust for trailing_wide_ints <3> representation.
3780 (get_nonzero_bits): Likewise.
3781 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3783 * tree-ssanames.h (struct range_info_def): Replace min, max and
3784 nonzero_bits with a trailing_wide_ints <3>.
3785 (set_range_info): Use wide_int_refs rather than double_ints.
3786 (set_nonzero_bits): Likewise.
3787 (get_range_info): Return wide_ints rather than double_ints.
3788 (get_nonzero_bits): Likewise.
3789 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3790 * tree-ssa-pre.c (phi_translate_1): Likewise.
3791 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3792 (acceptable_pow_call): Likewise.
3793 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3795 (vn_reference_fold_indirect): Likewise.
3796 (vn_reference_maybe_forwprop_address): Likewise.
3797 (valueize_refs_1): Likewise.
3798 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3799 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3800 tree_int_cst_lt and tree_int_cst_le.
3801 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3803 (streamer_alloc_tree): Likewise.
3804 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3805 (streamer_write_tree_header): Likewise.
3806 (streamer_write_integer_cst): Likewise.
3807 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3808 (build_constructors): Likewise.
3809 (array_value_type): Likewise.
3810 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3811 (vect_check_gather): Likewise.
3812 * tree-vect-generic.c (build_replicated_const): Likewise.
3813 (expand_vector_divmod): Likewise.
3814 * tree-vect-loop.c (vect_transform_loop): Likewise.
3815 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3816 (vect_do_peeling_for_alignment): Likewise.
3817 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3818 * tree-vrp.c: Include wide-int.h.
3819 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3820 (extract_range_from_assert): Use wide-int interfaces.
3821 (vrp_int_const_binop): Likewise.
3822 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3823 double_int pointers.
3824 (ranges_from_anti_range): Use wide-int interfaces.
3825 (quad_int_cmp): Delete.
3826 (quad_int_pair_sort): Likewise.
3827 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3828 (extract_range_from_unary_expr_1): Likewise.
3829 (adjust_range_with_scev): Likewise.
3830 (masked_increment): Take and return wide_ints rather than double_ints.
3831 (register_edge_assert_for_2): Use wide-int interfaces.
3832 (check_array_ref): Likewise.
3833 (search_for_addr_array): Likewise.
3834 (maybe_set_nonzero_bits): Likewise.
3835 (union_ranges): Pass an integer of the correct type instead of
3836 using integer_one_node.
3837 (intersect_ranges): Likewise.
3838 (simplify_truth_ops_using_ranges): Likewise.
3839 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3840 (range_fits_type_p): Likewise.
3841 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3843 (simplify_conversion_using_ranges): Use wide-int interfaces.
3844 (simplify_float_conversion_using_ranges): Likewise.
3845 (vrp_finalize): Likewise.
3846 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3847 (gimple_stringops_transform): Likewise.
3848 * varasm.c (decode_addr_const): Likewise.
3849 (const_hash_1): Likewise.
3850 (const_rtx_hash_1): Likewise
3851 (output_constant): Likewise.
3852 (array_size_for_constructor): Likewise.
3853 (output_constructor_regular_field): Likewise.
3854 (output_constructor_bitfield): Likewise.
3855 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3856 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3858 * gencheck.c: Define BITS_PER_UNIT.
3861 * wide-int-print.cc: New.
3862 * wide-int-print.h: New.
3864 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3866 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3868 2014-05-06 Richard Biener <rguenther@suse.de>
3870 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3871 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3872 (TODO_verify_all): Adjust.
3873 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3874 TODO_verify_stmts and TODO_verify_rtl_sharing.
3875 * bb-reorder.c: Likewise.
3876 * cfgexpand.c: Likewise.
3877 * cprop.c: Likewise.
3879 * function.c: Likewise.
3880 * fwprop.c: Likewise.
3882 * gimple-ssa-isolate-paths.c: Likewise.
3883 * gimple-ssa-strength-reduction.c: Likewise.
3884 * ipa-split.c: Likewise.
3885 * loop-init.c: Likewise.
3886 * loop-unroll.c: Likewise.
3887 * lower-subreg.c: Likewise.
3888 * modulo-sched.c: Likewise.
3889 * postreload-gcse.c: Likewise.
3890 * predict.c: Likewise.
3891 * recog.c: Likewise.
3892 * sched-rgn.c: Likewise.
3893 * store-motion.c: Likewise.
3894 * tracer.c: Likewise.
3895 * trans-mem.c: Likewise.
3896 * tree-call-cdce.c: Likewise.
3897 * tree-cfg.c: Likewise.
3898 * tree-cfgcleanup.c: Likewise.
3899 * tree-complex.c: Likewise.
3900 * tree-eh.c: Likewise.
3901 * tree-emutls.c: Likewise.
3902 * tree-if-conv.c: Likewise.
3903 * tree-into-ssa.c: Likewise.
3904 * tree-loop-distribution.c: Likewise.
3905 * tree-object-size.c: Likewise.
3906 * tree-parloops.c: Likewise.
3907 * tree-pass.h: Likewise.
3908 * tree-sra.c: Likewise.
3909 * tree-ssa-ccp.c: Likewise.
3910 * tree-ssa-copy.c: Likewise.
3911 * tree-ssa-copyrename.c: Likewise.
3912 * tree-ssa-dce.c: Likewise.
3913 * tree-ssa-dom.c: Likewise.
3914 * tree-ssa-dse.c: Likewise.
3915 * tree-ssa-forwprop.c: Likewise.
3916 * tree-ssa-ifcombine.c: Likewise.
3917 * tree-ssa-loop-ch.c: Likewise.
3918 * tree-ssa-loop-ivcanon.c: Likewise.
3919 * tree-ssa-loop.c: Likewise.
3920 * tree-ssa-math-opts.c: Likewise.
3921 * tree-ssa-phiopt.c: Likewise.
3922 * tree-ssa-phiprop.c: Likewise.
3923 * tree-ssa-pre.c: Likewise.
3924 * tree-ssa-reassoc.c: Likewise.
3925 * tree-ssa-sink.c: Likewise.
3926 * tree-ssa-strlen.c: Likewise.
3927 * tree-ssa-tail-merge.c: Likewise.
3928 * tree-ssa-uncprop.c: Likewise.
3929 * tree-switch-conversion.c: Likewise.
3930 * tree-tailcall.c: Likewise.
3931 * tree-vect-generic.c: Likewise.
3932 * tree-vectorizer.c: Likewise.
3933 * tree-vrp.c: Likewise.
3935 * var-tracking.c: Likewise.
3936 * bt-load.c: Likewise.
3937 * cfgcleanup.c: Likewise.
3938 * combine-stack-adj.c: Likewise.
3939 * combine.c: Likewise.
3940 * compare-elim.c: Likewise.
3941 * config/epiphany/resolve-sw-modes.c: Likewise.
3942 * config/i386/i386.c: Likewise.
3943 * config/mips/mips.c: Likewise.
3944 * config/s390/s390.c: Likewise.
3945 * config/sh/sh_treg_combine.cc: Likewise.
3946 * config/sparc/sparc.c: Likewise.
3949 * final.c: Likewise.
3950 * ifcvt.c: Likewise.
3951 * mode-switching.c: Likewise.
3952 * passes.c: Likewise.
3953 * postreload.c: Likewise.
3955 * reg-stack.c: Likewise.
3956 * regcprop.c: Likewise.
3957 * regrename.c: Likewise.
3960 2014-05-06 Richard Biener <rguenther@suse.de>
3963 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3964 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3966 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3969 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3971 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3972 Tom de Vries <tom@codesourcery.com>
3974 * target.def (call_fusage_contains_non_callee_clobbers): New
3976 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3978 (@node Miscellaneous Register Hooks): New node.
3979 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3980 * doc/tm.texi: Regenerate.
3982 2014-05-05 Marek Polacek <polacek@redhat.com>
3985 * opts.c (common_handle_option): Call error_at instead of warning_at.
3987 2014-05-05 Richard Biener <rguenther@suse.de>
3989 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3990 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3991 under the TODO_verify_il umbrella.
3993 2014-05-05 Richard Biener <rguenther@suse.de>
3995 * passes.c (execute_function_todo): Move TODO_verify_flow under
3996 the TODO_verify_ul umbrella.
3998 2014-05-05 Richard Biener <rguenther@suse.de>
4001 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4002 X & CST away from a CST that is the mask of a mode.
4004 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4006 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4007 int argument to enum machine_mode.
4008 (picochip_class_max_nregs): Ditto.
4009 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4010 (picochip_class_max_nregs): Ditto.
4012 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4014 * target.def: Add new target hook.
4015 * doc/tm.texi: Regenerate.
4016 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4017 * targhooks.c (default_keep_leaf_when_profiled): New function.
4019 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4020 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4022 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4024 PR tree-optimization/60363
4025 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4026 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4027 (update_destination_phis): New parameter.
4028 (create_edge_and_update_destination_phis): Ditto.
4029 (ssa_fix_duplicate_block_edges): Pass new arguments.
4030 (thread_single_edge): Ditto.
4032 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4034 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4035 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4036 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4037 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4038 Use RS6000_BTM_HARD_FLOAT.
4039 (BU_MISC_2): Likewise.
4040 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4041 RS6000_BTM_HARD_FLOAT.
4042 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4044 (rs6000_invalid_builtin): Add hard floating builtin support.
4045 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4046 hard float builtins.
4047 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4049 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4051 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4052 Add missing function* argument.
4054 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4056 * lra-constraints.c (valid_address_p): Move earlier in file.
4057 Add a constraint argument to the address_info version.
4058 (satisfies_memory_constraint_p): New function.
4059 (satisfies_address_constraint_p): Likewise.
4060 (process_alt_operands, curr_insn_transform): Use them.
4061 (process_address): Pass the constraint to valid_address_p when
4062 checking address operands.
4064 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4066 * config/mips/mips.c (mips_isa_rev): New variable.
4067 (mips_set_architecture): Set it.
4068 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4070 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4071 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4072 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4073 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4074 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4075 conditions in terms of mips_isa_rev.
4076 (mips_isa_rev): Declare.
4078 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4080 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4081 (prob_unlikely, prob_likely): Make variables const.
4083 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4085 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4087 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4089 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4091 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4093 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4094 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4095 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4097 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4098 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4100 Replace usage of ROUND_REG with sh_round_reg.
4101 Use CEIL instead of ROUND_ADVANCE.
4103 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4106 * config/sh/sh.c: Include stdlib headers before everything else.
4108 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4110 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4111 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4112 (gimplify_adjust_omp_clauses): Simd region is never
4113 directly nested in combined parallel. Instead, for linear
4114 with copyin/copyout, if in combined for simd loop, make decl
4115 firstprivate/lastprivate on OMP_FOR.
4116 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4117 expand_omp_for_static_chunk): When setting endvar, also set
4118 fd->loop.v to the same value.
4120 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4122 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4124 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4126 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4129 2014-05-02 Marek Polacek <polacek@redhat.com>
4131 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4133 2014-05-02 Kito Cheng <kito@0xlab.org>
4135 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4136 to a C expression marco.
4137 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4138 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4139 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4140 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4141 HONOR_REG_ALLOC_ORDER.
4142 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4144 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4146 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4148 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4150 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4152 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4154 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4155 (convert_scalar_cond_reduction): Likewise.
4156 (predicate_scalar_phi): Add recognition and transformation
4157 of simple conditioanl reduction to be vectorizable.
4159 2014-05-01 Marek Polacek <polacek@redhat.com>
4162 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4164 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4166 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4167 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4168 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4169 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4170 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4171 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4172 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4173 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4175 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4177 * config/arc/arc.opt (mlra): Move comment above option name
4178 to avoid mis-parsing as language options.
4180 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4182 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4183 * config/sol2.h: ... here.
4184 * config/sol2-10.h: Remove.
4186 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4187 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4188 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4189 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4190 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4191 * config/sol2.h: ... here.
4192 (SECTION_NAME_FORMAT): Don't redefine.
4193 (STARTFILE_ARCH32_SPEC): Rename to ...
4194 (STARTFILE_ARCH_SPEC): ... this.
4195 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4196 * config/sparc/sol2.h: ... here.
4197 (SECTION_NAME_FORMAT): Don't undef.
4198 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4199 (SUBTARGET_EXTRA_SPECS): Remove.
4200 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4202 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4203 (MD_STARTFILE_PREFIX): Remove.
4204 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4205 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4206 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4207 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4208 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4209 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4210 * config/i386/sol2.h: ... here.
4211 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4212 * config/i386/sol2-bi.h: Remove.
4213 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4214 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4216 * config/i386/t-sol2-64: Rename to ...
4217 * config/i386/t-sol2: ... this.
4218 * config/sparc/t-sol2-64: Rename to ...
4219 * config/sparc/t-sol2: ... this.
4221 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4222 sol2_tm_file_head, sol2_tm_file_tail.
4223 Include ${cpu_type}/sol2.h before sol2.h.
4225 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4226 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4227 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4228 Reflect i386/t-sol2-64 renaming.
4229 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4230 Reflect sparc/t-sol2-64 renaming.
4232 2014-04-30 Richard Biener <rguenther@suse.de>
4234 * passes.c (execute_function_todo): Move TODO_verify_stmts
4235 and TODO_verify_ssa under the TODO_verify_il umbrella.
4236 * tree-ssa.h (verify_ssa): Adjust prototype.
4237 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4238 we should verify SSA operands.
4239 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4240 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4241 whether we should verify whether not throwing stmts have EH info.
4242 * graphite-scop-detection.c (create_sese_edges): Adjust.
4243 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4244 * tree-eh.c (lower_try_finally_switch): Do not add the
4245 default case label twice.
4247 2014-04-30 Marek Polacek <polacek@redhat.com>
4249 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4250 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4251 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4252 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4254 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4256 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4257 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4258 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4259 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4260 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4261 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4262 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4263 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4265 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4267 * tree-cfg.c (dump_function_to_file): Dump the return type of
4268 functions, in a line to itself before the function body, mimicking
4269 the layout of a C function.
4271 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4273 PR tree-optimization/60971
4274 * tree-tailcall.c (process_assignment): Reject conversions which
4277 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4279 * calls.c (initialize_argument_information): Always treat
4280 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4281 (expand_call): Likewise.
4282 (emit_library_call_calue_1): Likewise.
4283 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4284 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4287 2014-04-29 Nick Clifton <nickc@redhat.com>
4289 * config/msp430/msp430.md (umulsidi): Fix typo.
4290 (mulhisi3): Enable even inside interrupt handlers.
4291 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4292 bigger return address pushed in large mode.
4294 2014-04-29 Nick Clifton <nickc@redhat.com>
4296 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4297 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4299 * config/m32r/m32r.c (init_reg_tables): Likewise.
4300 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4301 enum to hold the modes.
4303 2014-04-29 Richard Biener <rguenther@suse.de>
4305 * dominance.c (free_dominance_info): Add overload with
4307 (dom_info_state): Likewise.
4308 (dom_info_available_p): Likewise.
4309 * basic-block.h (free_dominance_info, dom_info_state,
4310 dom_info_available_p): Declare overloads.
4311 * passes.c (execute_function_todo): Verify that verifiers
4312 don't change dominator info state. Drop dominator info
4313 for IPA pass invocations.
4314 * cgraph.c (release_function_body): Restore asserts that
4315 dominator information is released.
4317 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4319 * doc/invoke.texi: Fix typo.
4320 * tree-vrp.c: Fix typos.
4321 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4323 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4325 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4327 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4329 * config/aarch64/aarch64-builtins.c
4330 (aarch64_types_storestruct_lane_qualifiers): New.
4331 (TYPES_STORESTRUCT_LANE): Likewise.
4332 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4333 (st3_lane): Likewise.
4334 (st4_lane): Likewise.
4335 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4336 (vec_store_lanesci_lane<mode>): Likewise.
4337 (vec_store_lanesxi_lane<mode>): Likewise.
4338 (aarch64_st2_lane<VQ:mode>): Likewise.
4339 (aarch64_st3_lane<VQ:mode>): Likewise.
4340 (aarch64_st4_lane<VQ:mode>): Likewise.
4341 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4342 * config/aarch64/arm_neon.h
4343 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4344 use new macro arguments.
4345 (__ST3_LANE_FUNC): Likewise.
4346 (__ST4_LANE_FUNC): Likewise.
4347 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4348 (V_THREE_ELEM): Likewise.
4349 (V_FOUR_ELEM): Likewise.
4351 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4353 * doc/gimple.texi: Replace the description of the now-defunct
4354 union gimple_statement_d with a diagram showing the
4355 gimple_statement_base class hierarchy and its relationships to
4356 the GSS_ and GIMPLE_ enums.
4358 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4360 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4361 * config/aarch64/aarch64.c
4362 (aarch64_cannot_change_mode_class): Weaken conditions.
4363 (aarch64_modes_tieable_p): New.
4364 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4366 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4368 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4369 (loadsync_<mode>): Change mode.
4370 (load_quadpti, store_quadpti): New.
4371 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4372 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4374 2014-04-28 Martin Jambor <mjambor@suse.cz>
4376 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4377 same alias type as the original statement.
4378 (subreplacement_assignment_data): New type.
4379 (handle_unscalarized_data_in_subtree): New type of parameter,
4380 generate new memory accesses with same alias type as the original
4382 (load_assign_lhs_subreplacements): Likewise.
4383 (sra_modify_constructor_assign): Generate new memory accesses with
4384 same alias type as the original statement.
4386 2014-04-28 Richard Biener <rguenther@suse.de>
4388 * tree-pass.h (TODO_verify_il): Define.
4389 (TODO_verify_all): Complete properly.
4390 * passes.c (execute_function_todo): Move existing loop-closed
4391 SSA verification under TODO_verify_il.
4392 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4393 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4394 Fix tree sharing issue.
4396 2014-04-28 Richard Biener <rguenther@suse.de>
4399 * builtins.def (DEF_C11_BUILTIN): Add.
4400 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4401 * coretypes.h (enum function_class): Add function_c11_misc.
4402 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4403 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4404 (call_may_clobber_ref_p_1): Likewise.
4405 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4406 (mark_all_reaching_defs_necessary_1): Likewise.
4407 (propagate_necessity): Likewise.
4408 (eliminate_unnecessary_stmts): Likewise.
4409 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4411 2014-04-28 Richard Biener <rguenther@suse.de>
4413 * tree-vrp.c (vrp_var_may_overflow): Remove.
4414 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4415 with overflow immediately bump to one before that value and
4416 let iteration figure out overflow status.
4418 2014-04-28 Richard Biener <rguenther@suse.de>
4420 * configure.ac: Do valgrind header checks unconditionally.
4421 Add --enable-valgrind-annotations.
4422 * system.h: Guard valgrind header inclusion with
4423 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4424 * alloc-pool.c (pool_alloc, pool_free): Use
4425 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4426 to guard possibly dead code.
4427 * config.in: Regenerated.
4428 * configure: Likewise.
4430 2014-04-28 Jeff Law <law@redhat.com>
4432 PR tree-optimization/60902
4433 * tree-ssa-threadedge.c
4434 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4435 over real defs when invalidating outputs from statements that do not
4436 produce useful outputs for threading.
4438 2014-04-28 Richard Biener <rguenther@suse.de>
4440 PR tree-optimization/60979
4441 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4442 SCOPs that end in a block with a successor with abnormal
4445 2014-04-28 Richard Biener <rguenther@suse.de>
4447 * tree-pass.h (execute_pass_list): Adjust prototype.
4448 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4449 (do_per_function): Change callback signature, push all actual
4450 work to the callbals.
4451 (do_per_function_toporder): Likewise.
4452 (execute_function_dump): Adjust.
4453 (execute_function_todo): Likewise.
4454 (clear_last_verified): Likewise.
4455 (verify_curr_properties): Likewise.
4456 (update_properties_after_pass): Likewise.
4457 (execute_pass_list_1): Split out from ...
4458 (execute_pass_list): ... here. Adjust.
4459 (execute_ipa_pass_list): Likewise.
4460 * cgraphunit.c (cgraph_add_new_function): Adjust.
4461 (analyze_function): Likewise.
4462 (expand_function): Likewise.
4463 * cgraph.c (release_function_body): Free dominance info
4464 here instead of asserting it was magically freed elsewhere.
4466 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4468 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4469 * configure: Regenerate.
4470 * config/sparc/sparc.opt (muser-mode): New option.
4471 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4473 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4474 * doc/invoke.texi (SPARC options): Document -muser-mode.
4476 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4478 * cselib.c (find_slot_memmode): Delete.
4479 (cselib_hasher): Change compare_type to a struct.
4480 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4482 (preserve_constants_and_equivs): Adjust for new compare_type.
4483 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4484 (wrap_constant): Delete.
4485 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4487 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4489 * doc/install.texi (Building with profile feedback): Remove
4492 2014-04-26 Tom de Vries <tom@codesourcery.com>
4494 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4497 2014-04-25 Cary Coutant <ccoutant@google.com>
4500 * dwarf2out.c (should_move_die_to_comdat): A type definition
4501 can contain a subprogram definition, but don't move it to a
4503 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4504 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4506 (clone_tree_hash): Rename to...
4507 (clone_tree_partial): ...this; change callers. Copy
4508 DW_TAG_subprogram DIEs as declarations.
4509 (copy_decls_walk): Don't copy children of a declaration into a
4512 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4515 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4518 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4520 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4521 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4523 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4526 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4528 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4530 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4532 PR tree-optimization/60930
4533 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4534 creating a multiply candidate by folding two constant
4535 multiplicands when the result overflows.
4537 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4539 PR tree-optimization/60960
4540 * tree-vect-generic.c (expand_vector_operation): Only call
4541 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4543 2014-04-25 Tom de Vries <tom@codesourcery.com>
4545 * expr.c (clobber_reg_mode): New function.
4546 * expr.h (clobber_reg): New function.
4548 2014-04-25 Tom de Vries <tom@codesourcery.com>
4550 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4553 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4554 Tom de Vries <tom@codesourcery.com>
4556 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4558 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4559 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4560 new argument to find_all_hard_reg_sets call.
4562 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4564 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4565 Use HOST_WIDE_INT_C for mask literal.
4566 (aarch_rev16_shleft_mask_imm_p): Likewise.
4568 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4571 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4573 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4575 PR preprocessor/56540
4576 * config/i386/i386-c.c (ix86_target_macros): Define
4577 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4579 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4581 * configure.ac (tga_func): Remove.
4582 (LIB_TLS_SPEC): Remove.
4583 * configure: Regenerate.
4584 * config.in: Regenerate.
4585 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4587 2014-04-25 Richard Biener <rguenther@suse.de>
4590 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4591 call stmt use/clobber sets during stmt walk instead of
4592 walking the possibly incomplete set of caller edges.
4594 2014-04-25 Richard Biener <rguenther@suse.de>
4597 * passes.c (apply_ipa_transforms): Inline into only caller ...
4598 (execute_one_pass): ... here. Properly bring in function
4599 bodies for nodes we want to apply IPA transforms to.
4601 2014-04-24 Cong Hou <congh@google.com>
4603 PR tree-optimization/60896
4604 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4605 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4606 (vect_mark_pattern_stmts): Set the def type of all statements in
4607 PATTERN_DEF_SEQ as vect_internal_def.
4609 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4611 * doc/extend.texi (PowerPC Built-in Functions): Document new
4612 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4613 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4615 * config/rs6000/predicates.md (const_0_to_3_operand): New
4616 predicate to match 0..3 integer constants.
4618 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4619 to support adding miscellaneous builtin functions.
4620 (BU_DFP_MISC_2): Likewise.
4621 (BU_P7_MISC_1): Likewise.
4622 (BU_P7_MISC_2): Likewise.
4623 (BU_P8V_MISC_3): Likewise.
4624 (BU_MISC_1): Likewise.
4625 (BU_MISC_2): Likewise.
4626 (DIVWE): Add extended divide builtin functions.
4629 (DIVWEUO): Likewise.
4633 (DIVDEUO): Likewise.
4634 (DXEX): Add decimal floating-point builtin functions.
4637 (DDEDPDQ): Likewise.
4639 (DENBCDQ): Likewise.
4646 (CDTBCD): Add new BCD builtin functions.
4650 (BCDADD_LT): Likewise.
4651 (BCDADD_EQ): Likewise.
4652 (BCDADD_GT): Likewise.
4653 (BCDADD_OV): Likewise.
4655 (BCDSUB_LT): Likewise.
4656 (BCDSUB_EQ): Likewise.
4657 (BCDSUB_GT): Likewise.
4658 (BCDSUB_OV): Likewise.
4659 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4660 (UNPACK_TD): Likewise.
4661 (PACK_TF): Likewise.
4662 (UNPACK_TF): Likewise.
4663 (UNPACK_TF_0): Likewise.
4664 (UNPACK_TF_1): Likewise.
4665 (PACK_V1TI): Likewise.
4666 (UNPACK_V1TI): Likewise.
4668 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4669 support for decimal floating point builtin functions.
4670 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4671 functions that take constant arguments.
4672 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4673 (rs6000_init_builtins): Setup long double, _Decimal64, and
4674 _Decimal128 types for new builtin functions.
4675 (builtin_function_type): Set the unsigned flags appropriately for
4676 the new builtin functions.
4677 (rs6000_opt_masks): Add support for decimal floating point builtin
4680 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4681 floating point builtin functions.
4682 (RS6000_BTM_COMMON): Likewise.
4683 (RS6000_BTI_long_double): Likewise.
4684 (RS6000_BTI_dfloat64): Likewise.
4685 (RS6000_BTI_dfloat128): Likewise.
4686 (long_double_type_internal_node): Likewise.
4687 (dfloat64_type_internal_node): Likewise.
4688 (dfloat128_type_internal_node): Likewise.
4690 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4691 2.07 bcd arithmetic instructions.
4692 (UNSPEC_BCDSUB): Likewise.
4693 (UNSPEC_BCD_OVERFLOW): Likewise.
4694 (UNSPEC_BCD_ADD_SUB): Likewise.
4695 (bcd_add_sub): Likewise.
4696 (BCD_TEST): Likewise.
4697 (bcd<bcd_add_sub>): Likewise.
4698 (bcd<bcd_add_sub>_test): Likewise.
4699 (bcd<bcd_add_sub>_test2): Likewise.
4700 (bcd<bcd_add_sub>_<code>): Likewise.
4701 (peephole2 for combined bcd ops): Likewise.
4703 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4704 decimal floating point builtin functions.
4705 (UNSPEC_DENBCD): Likewise.
4706 (UNSPEC_DXEX): Likewise.
4707 (UNSPEC_DIEX): Likewise.
4708 (UNSPEC_DSCLI): Likewise.
4709 (UNSPEC_DSCRI): Likewise.
4710 (D64_D128): Likewise.
4711 (dfp_suffix): Likewise.
4712 (dfp_ddedpd_<mode>): Likewise.
4713 (dfp_denbcd_<mode>): Likewise.
4714 (dfp_dxex_<mode>): Likewise.
4715 (dfp_diex_<mode>): Likewise.
4716 (dfp_dscli_<mode>): Likewise.
4717 (dfp_dscri_<mode>): Likewise.
4719 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4721 (UNSPEC_CDTBCD): Likewise.
4722 (UNSPEC_CBCDTD): Likewise.
4723 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4724 (UNSPEC_DIVEO): Likewise.
4725 (UNSPEC_DIVEU): Likewise.
4726 (UNSPEC_DIVEUO): Likewise.
4727 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4728 pack/unpack 128-bit types.
4729 (UNSPEC_PACK_128BIT): Likewise.
4730 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4731 (udiv<mode>3): Use idiv_ldiv mode attribute.
4732 (div<mode>3): Likewise.
4733 (addg6s): Add new BCD builtin functions.
4736 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4737 (div_extend): Likewise.
4738 (div<div_extend>_<mode>"): Likewise.
4739 (FP128_64): Add support for new builtin functions to pack/unpack
4741 (unpack<mode>): Likewise.
4742 (unpacktf_0): Likewise.
4743 (unpacktf_1): Likewise.
4744 (unpack<mode>_dm): Likewise.
4745 (unpack<mode>_nodm): Likewise.
4746 (pack<mode>): Likewise.
4747 (unpackv1ti): Likewise.
4748 (packv1ti): Likewise.
4750 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4752 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4755 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4757 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4758 * gimplify.c (omp_is_private): Change last argument's type to int.
4759 Only diagnose lastprivate if the simd argument is 1, only diagnose
4760 linear if the simd argument is 2.
4761 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4762 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4763 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4764 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4765 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4766 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4767 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4768 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4769 * tree-nested.c (convert_nonlocal_omp_clauses,
4770 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4772 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4775 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4778 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4780 * flag-types.h (enum ivar_visibility): Add.
4782 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4784 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4785 function * argument.
4787 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4789 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4791 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4792 Tom de Vries <tom@codesourcery.com>
4794 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4795 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4797 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4798 * emit-rtl.c (try_split): Same.
4800 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4801 Tom de Vries <tom@codesourcery.com>
4803 * common.opt (fuse-caller-save): New option.
4805 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4807 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4808 elements for big-endian.
4810 2014-04-24 Richard Biener <rguenther@suse.de>
4812 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4813 during TER and instead use the sepops interface for expanding
4814 non-GIMPLE_SINGLE_RHS.
4816 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4818 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4819 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4821 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4823 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4824 assembler 64-bit option.
4825 * configure: Regenerate.
4827 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4829 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4830 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4831 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4832 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4833 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4835 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4837 * config/aarch64/aarch64-builtins.c
4838 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4839 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4840 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4841 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4843 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4844 (Vrevsuff): New mode attribute.
4846 2014-04-24 Terry Guo <terry.guo@arm.com>
4848 * config/arm/arm.h (machine_function): Define variable
4849 after_arm_reorg here.
4850 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4851 (arm_split_constant): Update the way to access variable
4854 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4856 2014-04-23 Tom de Vries <tom@codesourcery.com>
4858 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4860 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4862 * is-a.h: Update comments to reflect the following changes to the
4863 "pointerness" of the API, making the template parameter match the
4864 return type, allowing use of is-a.h with typedefs of pointers.
4865 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4866 that the return type matches the parameter to the is_a_helper.
4868 (dyn_cast): Likewise.
4870 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4871 pointer from the is-a.h API.
4873 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4874 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4876 (is_a_helper <varpool_node>::test): Likewise, convert to...
4877 (is_a_helper <varpool_node *>::test): ...this.
4879 (varpool_first_variable): Update for removal of implicit pointer
4880 from the is-a.h API.
4881 (varpool_next_variable): Likewise.
4882 (varpool_first_static_initializer): Likewise.
4883 (varpool_next_static_initializer): Likewise.
4884 (varpool_first_defined_variable): Likewise.
4885 (varpool_next_defined_variable): Likewise.
4886 (cgraph_first_defined_function): Likewise.
4887 (cgraph_next_defined_function): Likewise.
4888 (cgraph_first_function): Likewise.
4889 (cgraph_next_function): Likewise.
4890 (cgraph_first_function_with_gimple_body): Likewise.
4891 (cgraph_next_function_with_gimple_body): Likewise.
4892 (cgraph_alias_target): Likewise.
4893 (varpool_alias_target): Likewise.
4894 (cgraph_function_or_thunk_node): Likewise.
4895 (varpool_variable_node): Likewise.
4896 (symtab_real_symbol_p): Likewise.
4897 * cgraphunit.c (referred_to_p): Likewise.
4898 (analyze_functions): Likewise.
4899 (handle_alias_pairs): Likewise.
4900 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4901 * gimple-ssa.h (gimple_vuse_op): Likewise.
4902 (gimple_vdef_op): Likewise.
4903 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4904 * gimple.c (gimple_build_asm_1): Likewise.
4905 (gimple_build_try): Likewise.
4906 (gimple_build_resx): Likewise.
4907 (gimple_build_eh_dispatch): Likewise.
4908 (gimple_build_omp_for): Likewise.
4909 (gimple_omp_for_set_clauses): Likewise.
4911 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4912 (is_a_helper <gimple_statement_asm *>::test): ...this.
4913 (is_a_helper <gimple_statement_bind>::test): Convert to...
4914 (is_a_helper <gimple_statement_bind *>::test): ...this.
4915 (is_a_helper <gimple_statement_call>::test): Convert to...
4916 (is_a_helper <gimple_statement_call *>::test): ...this.
4917 (is_a_helper <gimple_statement_catch>::test): Convert to...
4918 (is_a_helper <gimple_statement_catch *>::test): ...this.
4919 (is_a_helper <gimple_statement_resx>::test): Convert to...
4920 (is_a_helper <gimple_statement_resx *>::test): ...this.
4921 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4922 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4923 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4924 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4925 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4926 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4927 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4928 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4929 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4930 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4931 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4932 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4933 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4934 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4935 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4936 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4937 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4938 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4939 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4940 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4941 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4942 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4943 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4944 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4945 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4946 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4947 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4948 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4949 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4950 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4951 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4952 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4953 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4954 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4955 (is_a_helper <gimple_statement_phi>::test): Convert to...
4956 (is_a_helper <gimple_statement_phi *>::test): ...this.
4957 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4958 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4959 (is_a_helper <gimple_statement_try>::test): Convert to...
4960 (is_a_helper <gimple_statement_try *>::test): ...this.
4961 (is_a_helper <gimple_statement_wce>::test): Convert to...
4962 (is_a_helper <gimple_statement_wce *>::test): ...this.
4963 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4964 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4965 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4966 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4967 (is_a_helper <const gimple_statement_call>::test): Convert to...
4968 (is_a_helper <const gimple_statement_call *>::test): ...this.
4969 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4970 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4971 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4972 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4973 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4974 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4975 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4976 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4977 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4979 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4981 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4983 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4985 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4986 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4987 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4989 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4990 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4992 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4993 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4994 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4995 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4996 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4997 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4999 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5000 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5001 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5002 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5004 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5005 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5006 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5007 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5008 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5009 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5010 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5011 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5012 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5013 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5014 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5015 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5016 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5017 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5018 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5019 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5021 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5023 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5024 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5026 (gimple_use_ops): Update for removal of implicit pointer from the
5028 (gimple_set_use_ops): Likewise.
5029 (gimple_vuse): Likewise.
5030 (gimple_vdef): Likewise.
5031 (gimple_vuse_ptr): Likewise.
5032 (gimple_vdef_ptr): Likewise.
5033 (gimple_set_vuse): Likewise.
5034 (gimple_set_vdef): Likewise.
5035 (gimple_omp_return_set_lhs): Likewise.
5036 (gimple_omp_return_lhs): Likewise.
5037 (gimple_omp_return_lhs_ptr): Likewise.
5038 (gimple_call_fntype): Likewise.
5039 (gimple_call_set_fntype): Likewise.
5040 (gimple_call_set_internal_fn): Likewise.
5041 (gimple_call_use_set): Likewise.
5042 (gimple_call_clobber_set): Likewise.
5043 (gimple_bind_vars): Likewise.
5044 (gimple_bind_set_vars): Likewise.
5045 (gimple_bind_body_ptr): Likewise.
5046 (gimple_bind_set_body): Likewise.
5047 (gimple_bind_add_stmt): Likewise.
5048 (gimple_bind_block): Likewise.
5049 (gimple_bind_set_block): Likewise.
5050 (gimple_asm_ninputs): Likewise.
5051 (gimple_asm_noutputs): Likewise.
5052 (gimple_asm_nclobbers): Likewise.
5053 (gimple_asm_nlabels): Likewise.
5054 (gimple_asm_input_op): Likewise.
5055 (gimple_asm_input_op_ptr): Likewise.
5056 (gimple_asm_output_op): Likewise.
5057 (gimple_asm_output_op_ptr): Likewise.
5058 (gimple_asm_set_output_op): Likewise.
5059 (gimple_asm_clobber_op): Likewise.
5060 (gimple_asm_set_clobber_op): Likewise.
5061 (gimple_asm_label_op): Likewise.
5062 (gimple_asm_set_label_op): Likewise.
5063 (gimple_asm_string): Likewise.
5064 (gimple_catch_types): Likewise.
5065 (gimple_catch_types_ptr): Likewise.
5066 (gimple_catch_handler_ptr): Likewise.
5067 (gimple_catch_set_types): Likewise.
5068 (gimple_catch_set_handler): Likewise.
5069 (gimple_eh_filter_types): Likewise.
5070 (gimple_eh_filter_types_ptr): Likewise.
5071 (gimple_eh_filter_failure_ptr): Likewise.
5072 (gimple_eh_filter_set_types): Likewise.
5073 (gimple_eh_filter_set_failure): Likewise.
5074 (gimple_eh_must_not_throw_fndecl): Likewise.
5075 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5076 (gimple_eh_else_n_body_ptr): Likewise.
5077 (gimple_eh_else_e_body_ptr): Likewise.
5078 (gimple_eh_else_set_n_body): Likewise.
5079 (gimple_eh_else_set_e_body): Likewise.
5080 (gimple_try_eval_ptr): Likewise.
5081 (gimple_try_cleanup_ptr): Likewise.
5082 (gimple_try_set_eval): Likewise.
5083 (gimple_try_set_cleanup): Likewise.
5084 (gimple_wce_cleanup_ptr): Likewise.
5085 (gimple_wce_set_cleanup): Likewise.
5086 (gimple_phi_capacity): Likewise.
5087 (gimple_phi_num_args): Likewise.
5088 (gimple_phi_result): Likewise.
5089 (gimple_phi_result_ptr): Likewise.
5090 (gimple_phi_set_result): Likewise.
5091 (gimple_phi_arg): Likewise.
5092 (gimple_phi_set_arg): Likewise.
5093 (gimple_resx_region): Likewise.
5094 (gimple_resx_set_region): Likewise.
5095 (gimple_eh_dispatch_region): Likewise.
5096 (gimple_eh_dispatch_set_region): Likewise.
5097 (gimple_omp_critical_name): Likewise.
5098 (gimple_omp_critical_name_ptr): Likewise.
5099 (gimple_omp_critical_set_name): Likewise.
5100 (gimple_omp_for_clauses): Likewise.
5101 (gimple_omp_for_clauses_ptr): Likewise.
5102 (gimple_omp_for_set_clauses): Likewise.
5103 (gimple_omp_for_collapse): Likewise.
5104 (gimple_omp_for_index): Likewise.
5105 (gimple_omp_for_index_ptr): Likewise.
5106 (gimple_omp_for_set_index): Likewise.
5107 (gimple_omp_for_initial): Likewise.
5108 (gimple_omp_for_initial_ptr): Likewise.
5109 (gimple_omp_for_set_initial): Likewise.
5110 (gimple_omp_for_final): Likewise.
5111 (gimple_omp_for_final_ptr): Likewise.
5112 (gimple_omp_for_set_final): Likewise.
5113 (gimple_omp_for_incr): Likewise.
5114 (gimple_omp_for_incr_ptr): Likewise.
5115 (gimple_omp_for_set_incr): Likewise.
5116 (gimple_omp_for_pre_body_ptr): Likewise.
5117 (gimple_omp_for_set_pre_body): Likewise.
5118 (gimple_omp_parallel_clauses): Likewise.
5119 (gimple_omp_parallel_clauses_ptr): Likewise.
5120 (gimple_omp_parallel_set_clauses): Likewise.
5121 (gimple_omp_parallel_child_fn): Likewise.
5122 (gimple_omp_parallel_child_fn_ptr): Likewise.
5123 (gimple_omp_parallel_set_child_fn): Likewise.
5124 (gimple_omp_parallel_data_arg): Likewise.
5125 (gimple_omp_parallel_data_arg_ptr): Likewise.
5126 (gimple_omp_parallel_set_data_arg): Likewise.
5127 (gimple_omp_task_clauses): Likewise.
5128 (gimple_omp_task_clauses_ptr): Likewise.
5129 (gimple_omp_task_set_clauses): Likewise.
5130 (gimple_omp_task_child_fn): Likewise.
5131 (gimple_omp_task_child_fn_ptr): Likewise.
5132 (gimple_omp_task_set_child_fn): Likewise.
5133 (gimple_omp_task_data_arg): Likewise.
5134 (gimple_omp_task_data_arg_ptr): Likewise.
5135 (gimple_omp_task_set_data_arg): Likewise.
5136 (gimple_omp_taskreg_clauses): Likewise.
5137 (gimple_omp_taskreg_clauses_ptr): Likewise.
5138 (gimple_omp_taskreg_set_clauses): Likewise.
5139 (gimple_omp_taskreg_child_fn): Likewise.
5140 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5141 (gimple_omp_taskreg_set_child_fn): Likewise.
5142 (gimple_omp_taskreg_data_arg): Likewise.
5143 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5144 (gimple_omp_taskreg_set_data_arg): Likewise.
5145 (gimple_omp_task_copy_fn): Likewise.
5146 (gimple_omp_task_copy_fn_ptr): Likewise.
5147 (gimple_omp_task_set_copy_fn): Likewise.
5148 (gimple_omp_task_arg_size): Likewise.
5149 (gimple_omp_task_arg_size_ptr): Likewise.
5150 (gimple_omp_task_set_arg_size): Likewise.
5151 (gimple_omp_task_arg_align): Likewise.
5152 (gimple_omp_task_arg_align_ptr): Likewise.
5153 (gimple_omp_task_set_arg_align): Likewise.
5154 (gimple_omp_single_clauses): Likewise.
5155 (gimple_omp_single_clauses_ptr): Likewise.
5156 (gimple_omp_single_set_clauses): Likewise.
5157 (gimple_omp_target_clauses): Likewise.
5158 (gimple_omp_target_clauses_ptr): Likewise.
5159 (gimple_omp_target_set_clauses): Likewise.
5160 (gimple_omp_target_child_fn): Likewise.
5161 (gimple_omp_target_child_fn_ptr): Likewise.
5162 (gimple_omp_target_set_child_fn): Likewise.
5163 (gimple_omp_target_data_arg): Likewise.
5164 (gimple_omp_target_data_arg_ptr): Likewise.
5165 (gimple_omp_target_set_data_arg): Likewise.
5166 (gimple_omp_teams_clauses): Likewise.
5167 (gimple_omp_teams_clauses_ptr): Likewise.
5168 (gimple_omp_teams_set_clauses): Likewise.
5169 (gimple_omp_sections_clauses): Likewise.
5170 (gimple_omp_sections_clauses_ptr): Likewise.
5171 (gimple_omp_sections_set_clauses): Likewise.
5172 (gimple_omp_sections_control): Likewise.
5173 (gimple_omp_sections_control_ptr): Likewise.
5174 (gimple_omp_sections_set_control): Likewise.
5175 (gimple_omp_for_set_cond): Likewise.
5176 (gimple_omp_for_cond): Likewise.
5177 (gimple_omp_atomic_store_set_val): Likewise.
5178 (gimple_omp_atomic_store_val): Likewise.
5179 (gimple_omp_atomic_store_val_ptr): Likewise.
5180 (gimple_omp_atomic_load_set_lhs): Likewise.
5181 (gimple_omp_atomic_load_lhs): Likewise.
5182 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5183 (gimple_omp_atomic_load_set_rhs): Likewise.
5184 (gimple_omp_atomic_load_rhs): Likewise.
5185 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5186 (gimple_omp_continue_control_def): Likewise.
5187 (gimple_omp_continue_control_def_ptr): Likewise.
5188 (gimple_omp_continue_set_control_def): Likewise.
5189 (gimple_omp_continue_control_use): Likewise.
5190 (gimple_omp_continue_control_use_ptr): Likewise.
5191 (gimple_omp_continue_set_control_use): Likewise.
5192 (gimple_transaction_body_ptr): Likewise.
5193 (gimple_transaction_label): Likewise.
5194 (gimple_transaction_label_ptr): Likewise.
5195 (gimple_transaction_set_body): Likewise.
5196 (gimple_transaction_set_label): Likewise.
5198 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5199 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5200 * ipa-ref.c (ipa_record_reference): Likewise.
5201 * ipa-reference.c (analyze_function): Likewise.
5202 (ipa_reference_write_optimization_summary): Likewise.
5203 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5204 (address_taken_from_non_vtable_p): Likewise.
5205 (comdat_can_be_unshared_p_1): Likewise.
5206 * lto-cgraph.c (lto_output_ref): Likewise.
5207 (add_references): Likewise.
5208 (compute_ltrans_boundary): Likewise.
5209 (output_symtab): Likewise.
5210 (input_ref): Likewise.
5211 (input_cgraph_1): Likewise.
5212 (output_cgraph_opt_summary): Likewise.
5213 * lto-streamer-out.c (lto_output): Likewise.
5214 (output_symbol_p): Likewise.
5215 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5216 (lsei_start_function_in_partition): Likewise.
5217 (lsei_next_variable_in_partition): Likewise.
5218 (lsei_start_variable_in_partition): Likewise.
5219 * symtab.c (insert_to_assembler_name_hash): Likewise.
5220 (unlink_from_assembler_name_hash): Likewise.
5221 (symtab_unregister_node): Likewise.
5222 (symtab_remove_node): Likewise.
5223 (dump_symtab_node): Likewise.
5224 (verify_symtab_base): Likewise.
5225 (verify_symtab_node): Likewise.
5226 (symtab_make_decl_local): Likewise.
5227 (symtab_alias_ultimate_target): Likewise.
5228 (symtab_resolve_alias): Likewise.
5229 (symtab_get_symbol_partitioning_class): Likewise.
5230 * tree-phinodes.c (allocate_phi_node): Likewise.
5231 (reserve_phi_args_for_new_edge): Likewise.
5232 (remove_phi_args): Likewise.
5233 * varpool.c (varpool_node_for_asm): Likewise.
5234 (varpool_remove_unreferenced_decls): Likewise.
5236 2014-04-23 Jeff Law <law@redhat.com>
5238 PR tree-optimization/60902
5239 * tree-ssa-threadedge.c
5240 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5241 invalidate outputs from statements that do not produce useful
5242 outputs for threading.
5244 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5246 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5247 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5248 machine descriptions for Stack Smashing Protector.
5250 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5252 * aarch64.md (<optab>_rol<mode>3): New pattern.
5253 (<optab>_rolsi3_uxtw): Likewise.
5254 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5256 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5258 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5259 (arm_cortex_a12_tune): Likewise.
5261 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5263 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5265 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5267 * config/arm/arm.md (arm_rev16si2): New pattern.
5268 (arm_rev16si2_alt): Likewise.
5269 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5271 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5273 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5274 (rev16<mode>2_alt): Likewise.
5275 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5276 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5277 (aarch_rev16_shleft_mask_imm_p): Likewise.
5278 (aarch_rev16_p_1): Likewise.
5279 (aarch_rev16_p): Likewise.
5280 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5281 (aarch_rev16_shright_mask_imm_p): Likewise.
5282 (aarch_rev16_shleft_mask_imm_p): Likewise.
5284 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5286 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5287 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5289 (cortex_a53_extra_costs): Likewise.
5290 (cortex_a57_extra_costs): Likewise.
5291 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5292 (cortexa7_extra_costs): Likewise.
5293 (cortexa8_extra_costs): Likewise.
5294 (cortexa12_extra_costs): Likewise.
5295 (cortexa15_extra_costs): Likewise.
5296 (v7m_extra_costs): Likewise.
5297 (arm_new_rtx_costs): Handle BSWAP.
5299 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5301 * config/arm/arm.c (cortexa8_extra_costs): New table.
5302 (arm_cortex_a8_tune): New tuning struct.
5303 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5305 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5307 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5309 2014-04-23 Richard Biener <rguenther@suse.de>
5311 * Makefile.in (OBJS): Remove loop-unswitch.o.
5312 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5313 * passes.def (pass_rtl_unswitch): Likewise.
5314 * loop-init.c (gate_rtl_unswitch): Likewise.
5315 (rtl_unswitch): Likewise.
5316 (pass_data_rtl_unswitch): Likewise.
5317 (pass_rtl_unswitch): Likewise.
5318 (make_pass_rtl_unswitch): Likewise.
5319 * rtl.h (reversed_condition): Likewise.
5320 (compare_and_jump_seq): Likewise.
5321 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5323 * loop-unroll.c (compare_and_jump_seq): Likewise.
5325 2014-04-23 Richard Biener <rguenther@suse.de>
5327 PR tree-optimization/60903
5328 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5329 commented code block.
5330 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5331 loop flags to newly created BBs and edges.
5333 2014-04-23 Nick Clifton <nickc@redhat.com>
5335 * config/msp430/msp430.c (msp430_handle_option): Move function
5337 (msp430_option_override): Simplify mcu and mcpu option handling.
5338 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5339 support for -mhwmult command line option.
5340 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5341 -mhwmult command line option.
5342 (msp430_hwmult_enabled): Delete.
5343 (msp43o_output_labelref): Add support for -mhwmult command line option.
5344 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5345 (umulsidi3): Likewise.
5346 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5347 (mcpu, mlarge, msmall): Likewise.
5348 (mhwmult): New option.
5349 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5351 (msp430_is_f5_mcu): Remove prototype.
5352 (msp430_use_f5_series_hwmult): Add prototype.
5353 * config/msp430/msp430-opts.h: New file.
5354 * common/config/msp430: New directory.
5355 * common/config/msp430/msp430-common.c: New file.
5356 * config.gcc (msp430): Remove target_has_targetm_common.
5357 * doc/invoke.texi: Document -mhwmult command line option.
5359 2014-04-23 Nick Clifton <nickc@redhat.com>
5361 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5362 default-manifest.o if it can be found in the search path.
5363 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5365 2014-04-23 Terry Guo <terry.guo@arm.com>
5367 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5369 2014-04-23 Richard Biener <rguenther@suse.de>
5372 * tree-inline.c (declare_return_variable): Use mark_addressable.
5374 2014-04-23 Richard Biener <rguenther@suse.de>
5377 * loop-init.c (loop_optimizer_init): Make sure to apply
5378 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5380 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5383 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5385 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5386 if flag_sanitize_undefined_trap_on_error.
5387 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5388 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5389 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5390 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5391 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5392 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5393 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5394 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5395 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5396 * ubsan.c (ubsan_instrument_unreachable): Return
5397 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5398 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5399 if flag_sanitize_undefined_trap_on_error and
5400 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5401 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5402 instrument_bool_enum_load): Emit __builtin_trap () if
5403 flag_sanitize_undefined_trap_on_error and
5404 __builtin_handle_*_abort () if !flag_sanitize_recover.
5405 * doc/invoke.texi (-fsanitize-recover,
5406 -fsanitize-undefined-trap-on-error): Document.
5408 2014-04-22 Christian Bruel <christian.bruel@st.com>
5410 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5411 Force immediates to SImode.
5413 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5415 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5417 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5418 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5419 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5420 (nios2_fpu_insn): Add entry for round.
5421 (N2FPU_NO_ERRNO_P): Define.
5422 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5424 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5426 2014-04-22 Richard Henderson <rth@redhat.com>
5428 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5429 (add<GPI>3_compare0): Remove leading * from name.
5430 (add<GPI>3_carryin): Likewise.
5431 (sub<GPI>3_compare0): Likewise.
5432 (sub<GPI>3_carryin): Likewise.
5433 (<su_optab>mulditi3): New expander.
5434 (multi3): New expander.
5435 (madd<GPI>): Remove leading * from name.
5437 2014-04-22 Martin Jambor <mjambor@suse.cz>
5439 * cgraphclones.c (cgraph_function_versioning): Copy
5440 ipa_transforms_to_apply instead of asserting it is empty.
5442 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5445 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5446 on count_exp to get mode.
5448 2014-04-22 Andrew Pinski <apinski@cavium.com>
5450 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5451 Handle TLS for ILP32.
5452 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5453 (tlsie_small_<mode>): this and handle PTR.
5454 (tlsie_small_sidi): New pattern.
5455 (tlsle_small): Change to an expand to handle ILP32.
5456 (tlsle_small_<mode>): New pattern.
5457 (tlsdesc_small): Rename to ...
5458 (tlsdesc_small_<mode>): this and handle PTR.
5460 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5462 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5464 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5466 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5467 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5468 (aarch64_types_signed_poly_qualifiers): Likewise.
5469 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5470 (aarch64_types_poly_signed_qualifiers): Likewise.
5471 (TYPES_REINTERP_SS): Type macro added.
5472 (TYPES_REINTERP_SU): Likewise.
5473 (TYPES_REINTERP_SP): Likewise.
5474 (TYPES_REINTERP_US): Likewise.
5475 (TYPES_REINTERP_PS): Likewise.
5476 (aarch64_fold_builtin): New expression folding added.
5477 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5478 Declarations removed.
5479 (REINTERP_SS): Declarations added.
5480 (REINTERP_US): Likewise.
5481 (REINTERP_PS): Likewise.
5482 (REINTERP_SU): Likewise.
5483 (REINTERP_SP): Likewise.
5484 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5485 (vreinterpretq_p8_f64): Likewise.
5486 (vreinterpret_p16_f64): Likewise.
5487 (vreinterpretq_p16_f64): Likewise.
5488 (vreinterpret_f32_f64): Likewise.
5489 (vreinterpretq_f32_f64): Likewise.
5490 (vreinterpret_f64_f32): Likewise.
5491 (vreinterpret_f64_p8): Likewise.
5492 (vreinterpret_f64_p16): Likewise.
5493 (vreinterpret_f64_s8): Likewise.
5494 (vreinterpret_f64_s16): Likewise.
5495 (vreinterpret_f64_s32): Likewise.
5496 (vreinterpret_f64_s64): Likewise.
5497 (vreinterpret_f64_u8): Likewise.
5498 (vreinterpret_f64_u16): Likewise.
5499 (vreinterpret_f64_u32): Likewise.
5500 (vreinterpret_f64_u64): Likewise.
5501 (vreinterpretq_f64_f32): Likewise.
5502 (vreinterpretq_f64_p8): Likewise.
5503 (vreinterpretq_f64_p16): Likewise.
5504 (vreinterpretq_f64_s8): Likewise.
5505 (vreinterpretq_f64_s16): Likewise.
5506 (vreinterpretq_f64_s32): Likewise.
5507 (vreinterpretq_f64_s64): Likewise.
5508 (vreinterpretq_f64_u8): Likewise.
5509 (vreinterpretq_f64_u16): Likewise.
5510 (vreinterpretq_f64_u32): Likewise.
5511 (vreinterpretq_f64_u64): Likewise.
5512 (vreinterpret_s64_f64): Likewise.
5513 (vreinterpretq_s64_f64): Likewise.
5514 (vreinterpret_u64_f64): Likewise.
5515 (vreinterpretq_u64_f64): Likewise.
5516 (vreinterpret_s8_f64): Likewise.
5517 (vreinterpretq_s8_f64): Likewise.
5518 (vreinterpret_s16_f64): Likewise.
5519 (vreinterpretq_s16_f64): Likewise.
5520 (vreinterpret_s32_f64): Likewise.
5521 (vreinterpretq_s32_f64): Likewise.
5522 (vreinterpret_u8_f64): Likewise.
5523 (vreinterpretq_u8_f64): Likewise.
5524 (vreinterpret_u16_f64): Likewise.
5525 (vreinterpretq_u16_f64): Likewise.
5526 (vreinterpret_u32_f64): Likewise.
5527 (vreinterpretq_u32_f64): Likewise.
5529 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5531 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5532 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5533 (vreinterpret_p8_s8): Likewise.
5534 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5535 (vreinterpret_p8_s16): Likewise.
5536 (vreinterpret_p8_s32): Likewise.
5537 (vreinterpret_p8_s64): Likewise.
5538 (vreinterpret_p8_f32): Likewise.
5539 (vreinterpret_p8_u8): Likewise.
5540 (vreinterpret_p8_u16): Likewise.
5541 (vreinterpret_p8_u32): Likewise.
5542 (vreinterpret_p8_u64): Likewise.
5543 (vreinterpret_p8_p16): Likewise.
5544 (vreinterpretq_p8_s8): Likewise.
5545 (vreinterpretq_p8_s16): Likewise.
5546 (vreinterpretq_p8_s32): Likewise.
5547 (vreinterpretq_p8_s64): Likewise.
5548 (vreinterpretq_p8_f32): Likewise.
5549 (vreinterpretq_p8_u8): Likewise.
5550 (vreinterpretq_p8_u16): Likewise.
5551 (vreinterpretq_p8_u32): Likewise.
5552 (vreinterpretq_p8_u64): Likewise.
5553 (vreinterpretq_p8_p16): Likewise.
5554 (vreinterpret_p16_s8): Likewise.
5555 (vreinterpret_p16_s16): Likewise.
5556 (vreinterpret_p16_s32): Likewise.
5557 (vreinterpret_p16_s64): Likewise.
5558 (vreinterpret_p16_f32): Likewise.
5559 (vreinterpret_p16_u8): Likewise.
5560 (vreinterpret_p16_u16): Likewise.
5561 (vreinterpret_p16_u32): Likewise.
5562 (vreinterpret_p16_u64): Likewise.
5563 (vreinterpret_p16_p8): Likewise.
5564 (vreinterpretq_p16_s8): Likewise.
5565 (vreinterpretq_p16_s16): Likewise.
5566 (vreinterpretq_p16_s32): Likewise.
5567 (vreinterpretq_p16_s64): Likewise.
5568 (vreinterpretq_p16_f32): Likewise.
5569 (vreinterpretq_p16_u8): Likewise.
5570 (vreinterpretq_p16_u16): Likewise.
5571 (vreinterpretq_p16_u32): Likewise.
5572 (vreinterpretq_p16_u64): Likewise.
5573 (vreinterpretq_p16_p8): Likewise.
5574 (vreinterpret_f32_s8): Likewise.
5575 (vreinterpret_f32_s16): Likewise.
5576 (vreinterpret_f32_s32): Likewise.
5577 (vreinterpret_f32_s64): Likewise.
5578 (vreinterpret_f32_u8): Likewise.
5579 (vreinterpret_f32_u16): Likewise.
5580 (vreinterpret_f32_u32): Likewise.
5581 (vreinterpret_f32_u64): Likewise.
5582 (vreinterpret_f32_p8): Likewise.
5583 (vreinterpret_f32_p16): Likewise.
5584 (vreinterpretq_f32_s8): Likewise.
5585 (vreinterpretq_f32_s16): Likewise.
5586 (vreinterpretq_f32_s32): Likewise.
5587 (vreinterpretq_f32_s64): Likewise.
5588 (vreinterpretq_f32_u8): Likewise.
5589 (vreinterpretq_f32_u16): Likewise.
5590 (vreinterpretq_f32_u32): Likewise.
5591 (vreinterpretq_f32_u64): Likewise.
5592 (vreinterpretq_f32_p8): Likewise.
5593 (vreinterpretq_f32_p16): Likewise.
5594 (vreinterpret_s64_s8): Likewise.
5595 (vreinterpret_s64_s16): Likewise.
5596 (vreinterpret_s64_s32): Likewise.
5597 (vreinterpret_s64_f32): Likewise.
5598 (vreinterpret_s64_u8): Likewise.
5599 (vreinterpret_s64_u16): Likewise.
5600 (vreinterpret_s64_u32): Likewise.
5601 (vreinterpret_s64_u64): Likewise.
5602 (vreinterpret_s64_p8): Likewise.
5603 (vreinterpret_s64_p16): Likewise.
5604 (vreinterpretq_s64_s8): Likewise.
5605 (vreinterpretq_s64_s16): Likewise.
5606 (vreinterpretq_s64_s32): Likewise.
5607 (vreinterpretq_s64_f32): Likewise.
5608 (vreinterpretq_s64_u8): Likewise.
5609 (vreinterpretq_s64_u16): Likewise.
5610 (vreinterpretq_s64_u32): Likewise.
5611 (vreinterpretq_s64_u64): Likewise.
5612 (vreinterpretq_s64_p8): Likewise.
5613 (vreinterpretq_s64_p16): Likewise.
5614 (vreinterpret_u64_s8): Likewise.
5615 (vreinterpret_u64_s16): Likewise.
5616 (vreinterpret_u64_s32): Likewise.
5617 (vreinterpret_u64_s64): Likewise.
5618 (vreinterpret_u64_f32): Likewise.
5619 (vreinterpret_u64_u8): Likewise.
5620 (vreinterpret_u64_u16): Likewise.
5621 (vreinterpret_u64_u32): Likewise.
5622 (vreinterpret_u64_p8): Likewise.
5623 (vreinterpret_u64_p16): Likewise.
5624 (vreinterpretq_u64_s8): Likewise.
5625 (vreinterpretq_u64_s16): Likewise.
5626 (vreinterpretq_u64_s32): Likewise.
5627 (vreinterpretq_u64_s64): Likewise.
5628 (vreinterpretq_u64_f32): Likewise.
5629 (vreinterpretq_u64_u8): Likewise.
5630 (vreinterpretq_u64_u16): Likewise.
5631 (vreinterpretq_u64_u32): Likewise.
5632 (vreinterpretq_u64_p8): Likewise.
5633 (vreinterpretq_u64_p16): Likewise.
5634 (vreinterpret_s8_s16): Likewise.
5635 (vreinterpret_s8_s32): Likewise.
5636 (vreinterpret_s8_s64): Likewise.
5637 (vreinterpret_s8_f32): Likewise.
5638 (vreinterpret_s8_u8): Likewise.
5639 (vreinterpret_s8_u16): Likewise.
5640 (vreinterpret_s8_u32): Likewise.
5641 (vreinterpret_s8_u64): Likewise.
5642 (vreinterpret_s8_p8): Likewise.
5643 (vreinterpret_s8_p16): Likewise.
5644 (vreinterpretq_s8_s16): Likewise.
5645 (vreinterpretq_s8_s32): Likewise.
5646 (vreinterpretq_s8_s64): Likewise.
5647 (vreinterpretq_s8_f32): Likewise.
5648 (vreinterpretq_s8_u8): Likewise.
5649 (vreinterpretq_s8_u16): Likewise.
5650 (vreinterpretq_s8_u32): Likewise.
5651 (vreinterpretq_s8_u64): Likewise.
5652 (vreinterpretq_s8_p8): Likewise.
5653 (vreinterpretq_s8_p16): Likewise.
5654 (vreinterpret_s16_s8): Likewise.
5655 (vreinterpret_s16_s32): Likewise.
5656 (vreinterpret_s16_s64): Likewise.
5657 (vreinterpret_s16_f32): Likewise.
5658 (vreinterpret_s16_u8): Likewise.
5659 (vreinterpret_s16_u16): Likewise.
5660 (vreinterpret_s16_u32): Likewise.
5661 (vreinterpret_s16_u64): Likewise.
5662 (vreinterpret_s16_p8): Likewise.
5663 (vreinterpret_s16_p16): Likewise.
5664 (vreinterpretq_s16_s8): Likewise.
5665 (vreinterpretq_s16_s32): Likewise.
5666 (vreinterpretq_s16_s64): Likewise.
5667 (vreinterpretq_s16_f32): Likewise.
5668 (vreinterpretq_s16_u8): Likewise.
5669 (vreinterpretq_s16_u16): Likewise.
5670 (vreinterpretq_s16_u32): Likewise.
5671 (vreinterpretq_s16_u64): Likewise.
5672 (vreinterpretq_s16_p8): Likewise.
5673 (vreinterpretq_s16_p16): Likewise.
5674 (vreinterpret_s32_s8): Likewise.
5675 (vreinterpret_s32_s16): Likewise.
5676 (vreinterpret_s32_s64): Likewise.
5677 (vreinterpret_s32_f32): Likewise.
5678 (vreinterpret_s32_u8): Likewise.
5679 (vreinterpret_s32_u16): Likewise.
5680 (vreinterpret_s32_u32): Likewise.
5681 (vreinterpret_s32_u64): Likewise.
5682 (vreinterpret_s32_p8): Likewise.
5683 (vreinterpret_s32_p16): Likewise.
5684 (vreinterpretq_s32_s8): Likewise.
5685 (vreinterpretq_s32_s16): Likewise.
5686 (vreinterpretq_s32_s64): Likewise.
5687 (vreinterpretq_s32_f32): Likewise.
5688 (vreinterpretq_s32_u8): Likewise.
5689 (vreinterpretq_s32_u16): Likewise.
5690 (vreinterpretq_s32_u32): Likewise.
5691 (vreinterpretq_s32_u64): Likewise.
5692 (vreinterpretq_s32_p8): Likewise.
5693 (vreinterpretq_s32_p16): Likewise.
5694 (vreinterpret_u8_s8): Likewise.
5695 (vreinterpret_u8_s16): Likewise.
5696 (vreinterpret_u8_s32): Likewise.
5697 (vreinterpret_u8_s64): Likewise.
5698 (vreinterpret_u8_f32): Likewise.
5699 (vreinterpret_u8_u16): Likewise.
5700 (vreinterpret_u8_u32): Likewise.
5701 (vreinterpret_u8_u64): Likewise.
5702 (vreinterpret_u8_p8): Likewise.
5703 (vreinterpret_u8_p16): Likewise.
5704 (vreinterpretq_u8_s8): Likewise.
5705 (vreinterpretq_u8_s16): Likewise.
5706 (vreinterpretq_u8_s32): Likewise.
5707 (vreinterpretq_u8_s64): Likewise.
5708 (vreinterpretq_u8_f32): Likewise.
5709 (vreinterpretq_u8_u16): Likewise.
5710 (vreinterpretq_u8_u32): Likewise.
5711 (vreinterpretq_u8_u64): Likewise.
5712 (vreinterpretq_u8_p8): Likewise.
5713 (vreinterpretq_u8_p16): Likewise.
5714 (vreinterpret_u16_s8): Likewise.
5715 (vreinterpret_u16_s16): Likewise.
5716 (vreinterpret_u16_s32): Likewise.
5717 (vreinterpret_u16_s64): Likewise.
5718 (vreinterpret_u16_f32): Likewise.
5719 (vreinterpret_u16_u8): Likewise.
5720 (vreinterpret_u16_u32): Likewise.
5721 (vreinterpret_u16_u64): Likewise.
5722 (vreinterpret_u16_p8): Likewise.
5723 (vreinterpret_u16_p16): Likewise.
5724 (vreinterpretq_u16_s8): Likewise.
5725 (vreinterpretq_u16_s16): Likewise.
5726 (vreinterpretq_u16_s32): Likewise.
5727 (vreinterpretq_u16_s64): Likewise.
5728 (vreinterpretq_u16_f32): Likewise.
5729 (vreinterpretq_u16_u8): Likewise.
5730 (vreinterpretq_u16_u32): Likewise.
5731 (vreinterpretq_u16_u64): Likewise.
5732 (vreinterpretq_u16_p8): Likewise.
5733 (vreinterpretq_u16_p16): Likewise.
5734 (vreinterpret_u32_s8): Likewise.
5735 (vreinterpret_u32_s16): Likewise.
5736 (vreinterpret_u32_s32): Likewise.
5737 (vreinterpret_u32_s64): Likewise.
5738 (vreinterpret_u32_f32): Likewise.
5739 (vreinterpret_u32_u8): Likewise.
5740 (vreinterpret_u32_u16): Likewise.
5741 (vreinterpret_u32_u64): Likewise.
5742 (vreinterpret_u32_p8): Likewise.
5743 (vreinterpret_u32_p16): Likewise.
5744 (vreinterpretq_u32_s8): Likewise.
5745 (vreinterpretq_u32_s16): Likewise.
5746 (vreinterpretq_u32_s32): Likewise.
5747 (vreinterpretq_u32_s64): Likewise.
5748 (vreinterpretq_u32_f32): Likewise.
5749 (vreinterpretq_u32_u8): Likewise.
5750 (vreinterpretq_u32_u16): Likewise.
5751 (vreinterpretq_u32_u64): Likewise.
5752 (vreinterpretq_u32_p8): Likewise.
5753 (vreinterpretq_u32_p16): Likewise.
5755 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5757 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5759 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5761 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5762 (vqnegd_s64): Likewise.
5763 (vqabs_s64): Likewise.
5764 (vqabsd_s64): Likewise.
5766 2014-04-22 Richard Henderson <rth@redhat.com>
5768 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5769 computation to the top of the loop.
5771 2014-04-22 Renlin <renlin.li@arm.com>
5772 Jiong Wang <jiong.wang@arm.com>
5774 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5775 * config/aarch64/aarch64.c (aarch64_layout_frame)
5776 (aarch64_initial_elimination_offset): Likewise.
5778 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5780 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5783 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5785 * machmode.h (bitwise_mode_for_mode): Declare.
5786 * stor-layout.h (bitwise_type_for_mode): Likewise.
5787 * stor-layout.c (bitwise_mode_for_mode): New function.
5788 (bitwise_type_for_mode): Likewise.
5789 * builtins.c (fold_builtin_memory_op): Use it instead of
5790 int_mode_for_mode and build_nonstandard_integer_type.
5792 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5794 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5795 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5796 (*-*-solaris2*): Simplify.
5797 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5798 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5799 *-*-solaris2.9* handling.
5801 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5803 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5804 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5806 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5807 * configure: Regenerate.
5809 * config/i386/sol2-9.h: Remove.
5811 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5812 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5813 Remove Solaris 9 references.
5815 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5817 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5818 (floatuns<GPI:mode><GPF:mode>2): Remove.
5819 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5820 and floatuns conversions.
5821 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5822 and floatuns conversions.
5823 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5824 (w1,w2): New mode attributes for inequal width conversions.
5826 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5828 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5829 the output asm format.
5831 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5833 * config/aarch64/aarch64-simd.md
5834 (aarch64_cm<optab>di): Always split.
5835 (*aarch64_cm<optab>di): New.
5836 (aarch64_cmtstdi): Always split.
5837 (*aarch64_cmtstdi): New.
5839 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5841 PR tree-optimization/60823
5842 * omp-low.c (ipa_simd_modify_function_body): Go through
5843 all SSA_NAMEs and for those refering to vector arguments
5844 which are going to be replaced adjust SSA_NAME_VAR and,
5845 if it is a default definition, change it into a non-default
5846 definition assigned at the beginning of function from new_decl.
5847 (ipa_simd_modify_stmt_ops): Rewritten.
5848 * tree-dfa.c (set_ssa_default_def): When removing default def,
5849 check for NULL loc instead of NULL *loc.
5851 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5853 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5854 restrictions on core registers for DImode values in Thumb2.
5856 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5858 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5859 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5861 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5863 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5864 (*iordi_notzesidi_di): Likewise.
5865 (*iordi_notsesidi_di): Likewise.
5867 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5869 * config/arm/arm-protos.h (tune_params): New struct members.
5870 * config/arm/arm.c: Initialise tune_params per processor.
5871 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5872 for speed, based on new tune_params.
5874 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5876 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5877 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5878 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5879 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5880 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5881 (vrnda_f64): Likewise.
5882 (vrndi_f64): Likewise.
5883 (vrndm_f64): Likewise.
5884 (vrndn_f64): Likewise.
5885 (vrndp_f64): Likewise.
5886 (vrndx_f64): Likewise.
5888 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5890 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5891 GET_MODE_SIZE argument is enum machine_mode.
5893 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5896 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5897 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5899 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5902 * asan.c (asan_emit_stack_protection): Force the base to align to
5903 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5904 appropriate bits if STRICT_ALIGNMENT.
5905 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5907 (expand_used_vars): Leave a space in the stack frame for alignment
5908 if STRICT_ALIGNMENT.
5910 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5912 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5914 (gimple_store_p): Likewise.
5915 (gimple_assign_load_p): Likewise.
5916 (gimple_assign_cast_p): Likewise.
5917 (gimple_clobber_p): Likewise.
5919 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5920 rather than a gimple.
5921 (gimple_assign_cast_p): Likewise.
5923 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5926 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5927 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5929 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5930 more debug information for E500 if -mdebug=reg.
5932 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5935 * config/i386/i386.c (ix86_expand_builtin)
5936 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5937 register for target RTX.
5938 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5940 2014-04-18 Cong Hou <congh@google.com>
5942 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5943 the widen-mult pattern by handling two operands with different sizes,
5944 and operands whose size is smaller than half of the result type.
5946 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5948 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5949 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5950 (do_estimate_edge_time): Compute it.
5951 * ipa-inline.c (want_inline_small_function_p): Bypass
5952 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5954 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5956 * ipa-inline.c (spec_rem): New static variable.
5957 (dump_overall_stats): New function.
5958 (dump_inline_stats): New function.
5960 2014-04-18 Richard Henderson <rth@redhat.com>
5962 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5963 to GET_MODE_SIZE, not a reg_class_t.
5965 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5967 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5968 (vsx_xxmrglw_<mode>): Likewise.
5970 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5973 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5974 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5975 (rs6000_init_hard_regno_mode_ok): Likewise.
5977 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5979 * ipa-inline.c (inline_small_functions): Account only non-cold
5981 * doc/invoke.texi (inline-unit-growth): Update documentation.
5983 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5985 * config/rs6000/rs6000.md (addti3, subti3): New.
5987 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5990 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5991 comment. Check optimize_insn_for_size_p instead of
5992 optimize_insn_for_speed_p.
5994 2014-04-17 Martin Jambor <mjambor@suse.cz>
5996 * gimple-iterator.c (gsi_start_edge): New function.
5997 * gimple-iterator.h (gsi_start_edge): Declare.
5998 * tree-sra.c (single_non_eh_succ): New function.
5999 (disqualify_ops_if_throwing_stmt): Renamed to
6000 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6001 having one non-EH successor BB.
6002 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6003 generate loads into replacements.
6004 (sra_modify_assign): Likewise and and also use the simple path for
6006 (sra_modify_function_body): Commit statements on edges.
6008 2014-04-17 Richard Biener <rguenther@suse.de>
6011 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6012 comparison results and add clarifying comment.
6014 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6016 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6017 (blank_mode): Initialize it.
6018 (emit_mode_size_inline, emit_mode_nunits_inline,
6019 emit_mode_inner_inline): New functions.
6020 (emit_insn_modes_h): Call them and surround their output with
6021 #if GCC_VERSION >= 4001 ... #endif.
6022 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6023 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6024 mode_* arrays if the argument is __builtin_constant_p.
6025 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6026 is enum machine_mode.
6028 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6030 * passes.c (opt_pass::execute): Adjust.
6031 (pass_manager::execute_pass_mode_switching): Likewise.
6032 (early_local_passes::execute): Likewise.
6033 (execute_one_pass): Pass cfun to the pass's execute method.
6034 * tree-pass.h (opt_pass::execute): Add function * argument.
6035 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6036 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6037 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6038 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6039 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6040 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6041 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6042 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6043 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6044 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6045 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6046 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6047 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6048 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6049 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6050 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6051 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6052 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6053 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6054 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6055 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6056 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6057 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6058 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6059 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6060 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6061 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6062 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6063 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6066 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6068 * passes.c (opt_pass::gate): Take function * argument.
6069 (gate_all_early_local_passes): Merge into
6070 (early_local_passes::gate): this.
6071 (gate_all_early_optimizations): Merge into
6072 (all_early_optimizations::gate): this.
6073 (gate_all_optimizations): Mege into
6074 (all_optimizations::gate): this.
6075 (gate_all_optimizations_g): Merge into
6076 (all_optimizations_g::gate): this.
6077 (gate_rest_of_compilation): Mege into
6078 (rest_of_compilation::gate): this.
6079 (gate_postreload): Merge into
6080 (postreload::gate): this.
6081 (dump_one_pass): Pass cfun to the pass's gate method.
6082 (execute_ipa_summary_passes): Likewise.
6083 (execute_one_pass): Likewise.
6084 (ipa_write_summaries_2): Likewise.
6085 (ipa_write_optimization_summaries_1): Likewise.
6086 (ipa_read_summaries_1): Likewise.
6087 (ipa_read_optimization_summaries_1): Likewise.
6088 (execute_ipa_stmt_fixups): Likewise.
6089 * tree-pass.h (opt_pass::gate): Add function * argument.
6090 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6091 combine-stack-adj.c, combine.c, compare-elim.c,
6092 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6093 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6094 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6095 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6096 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6097 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6098 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6099 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6100 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6101 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6102 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6103 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6104 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6105 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6106 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6107 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6108 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6109 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6110 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6111 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6112 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6113 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6114 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6115 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6116 var-tracking.c, vtable-verify.c, web.c: Adjust.
6118 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6120 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6121 * configure: Regenerate.
6123 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6125 * passes.c (dump_one_pass): don't check pass->has_gate.
6126 (execute_ipa_summary_passes): Likewise.
6127 (execute_one_pass): Likewise.
6128 (ipa_write_summaries_2): Likewise.
6129 (ipa_write_optimization_summaries_1): Likewise.
6130 (ipa_read_optimization_summaries_1): Likewise.
6131 (execute_ipa_stmt_fixups): Likewise.
6132 * tree-pass.h (pass_data::has_gate): Remove.
6133 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6134 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6135 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6136 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6137 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6138 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6139 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6140 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6141 gimple-low.c, gimple-ssa-isolate-paths.c,
6142 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6143 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6144 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6145 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6146 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6147 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6148 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6149 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6150 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6151 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6152 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6153 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6154 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6155 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6156 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6157 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6158 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6159 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6160 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6161 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6162 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6163 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6166 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6168 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6170 * passes.c (pass_manager::register_dump_files_1): Merge into
6171 (pass_manager::register_dump_files): this, and remove its handling of
6172 properties since the pass always has the properties anyway.
6173 (pass_manager::pass_manager): Adjust.
6175 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6177 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6178 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6179 dealing with properties.
6180 (pass_manager::register_dump_files): Adjust.
6182 2014-03-20 Mark Wielaard <mjw@redhat.com>
6184 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6185 then represent the bound as normal constant value.
6187 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6190 Forward port from 4.8 branch
6191 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6193 * config/i386/bmiintrin.h (_blsi_u32): New.
6197 (_blsmsk_u32): Ditto.
6198 (_blsmsk_u64): Ditto.
6199 (_tzcnt_u32): Ditto.
6200 (_tzcnt_u64): Ditto.
6202 2014-04-17 Kito Cheng <kito@0xlab.org>
6204 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6206 2014-04-17 Richard Biener <rguenther@suse.de>
6209 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6210 boolean results for comparisons.
6212 2014-04-17 Richard Biener <rguenther@suse.de>
6214 PR tree-optimization/60836
6215 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6216 initial PHI args to be gimple values.
6218 2014-04-17 Richard Biener <rguenther@suse.de>
6220 PR tree-optimization/60841
6221 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6222 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6223 of stmts to SLP build.
6224 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6225 (vect_analyze_slp): Likewise.
6226 (vect_analyze_slp_instance): Likewise.
6227 (vect_build_slp_tree): Limit overall SLP tree growth.
6228 * tree-vectorizer.h (vect_analyze_data_refs,
6229 vect_analyze_slp): Adjust prototypes.
6231 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6233 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6236 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6238 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6239 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6240 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6241 for TARGET_SLOW_PSHUFB
6243 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6245 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6246 * config/i386/i386.c (intel_cost): Ditto.
6248 2014-04-17 Joey Ye <joey.ye@arm.com>
6250 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6252 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6254 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6257 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6259 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6260 (type_all_derivations_known_p): New predicate.
6261 (type_all_ctors_visible_p): New predicate.
6262 (type_possibly_instantiated_p): New predicate.
6263 (get_odr_type): Compute all_derivations_known.
6264 (dump_odr_type): Dump the flag.
6265 (maybe_record_type): Cleanup.
6266 (record_target_from_binfo): Add bases_to_consider array;
6267 record bases for types w/o instances and skip CXX destructor.
6268 (possible_polymorphic_call_targets_1): Add bases_to_consider
6269 and consider_construction parameters; check if type may have instance.
6270 (get_polymorphic_call_info): Set maybe_in_construction to true
6271 when we know nothing.
6272 (record_targets_from_bases): Skip CXX destructors; they are
6273 never called for types in construction.
6274 (possible_polymorphic_call_targets): Do not record target when
6275 type may not have instance.
6277 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6280 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6281 external aliases alive, too.
6283 2014-04-16 Andrew Pinski <apinski@cavium.com>
6285 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6288 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6290 * final.c (compute_alignments): Do not apply loop alignment to a block
6291 falling through to the exit.
6293 2014-04-16 Catherine Moore <clm@codesourcery.com>
6295 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6296 Adjust constraints for microMIPS store patterns.
6298 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6300 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6302 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6304 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6305 (append_use): Run at -O0.
6306 (append_vdef): Likewise.
6307 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6308 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6310 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6312 PR tree-optimization/60844
6313 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6314 (propagate_op_to_single_use, remove_visited_stmt_chain,
6315 linearize_expr, repropagate_negates, reassociate_bb): Use it
6316 instead of gsi_remove.
6318 2014-04-16 Martin Jambor <mjambor@suse.cz>
6320 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6321 ipa_transforms_to_apply.
6322 (cgraph_function_versioning): Assert that old_node has empty
6323 ipa_transforms_to_apply.
6324 * trans-mem.c (ipa_tm_create_version): Likewise.
6325 * tree-inline.c (tree_function_versioning): Do not duplicate
6326 ipa_transforms_to_apply.
6328 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6331 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6333 Pass necessary as flags on 64-bit Solaris/x86.
6334 Use lowercase relocs for x86_64-*-*.
6335 * configure: Regenerate.
6337 2014-04-15 Jan Hubicka <jh@suse.cz>
6339 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6340 (maybe_record_node, likely_target_p): Use it.
6342 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6345 Revert following patch
6347 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6350 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6351 software floating point or no floating point registers, do not
6352 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6353 in GPRs that occurs after we tested for GPRs that would never be
6356 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6357 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6358 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6359 specifically allow DDmode, since that does not use the SPE SIMD
6362 2014-03-21 Mark Wielaard <mjw@redhat.com>
6364 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6365 as unsigned or int depending on type and value used.
6367 2014-04-15 Richard Biener <rguenther@suse.de>
6369 PR rtl-optimization/56965
6370 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6371 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6373 * alias.c (true_dependence_1): Do not call
6374 nonoverlapping_component_refs_p.
6375 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6376 nonoverlapping_component_refs_p.
6377 (indirect_refs_may_alias_p): Likewise.
6379 2014-04-15 Teresa Johnson <tejohnson@google.com>
6381 * cfg.c (dump_bb_info): Fix flags check.
6382 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6384 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6386 PR rtl-optimization/60663
6387 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6390 2014-04-15 Richard Biener <rguenther@suse.de>
6392 * lto-streamer.h (LTO_major_version): Bump to 4.
6394 2014-04-15 Richard Biener <rguenther@suse.de>
6396 * common.opt (lto_partition_model): New enum.
6397 (flto-partition=): Merge separate options with a single with argument,
6398 add -flto-partition=one support.
6399 * flag-types.h (enum lto_partition_model): Declare.
6400 * opts.c (finish_options): Remove duplicate -flto-partition=
6402 * lto-wrapper.c (run_gcc): Adjust.
6404 2014-04-15 Richard Biener <rguenther@suse.de>
6406 * alias.c (ncr_compar): New function.
6407 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6409 2014-04-15 Richard Biener <rguenther@suse.de>
6411 * alias.c (record_component_aliases): Do not walk BINFOs.
6413 2014-04-15 Richard Biener <rguenther@suse.de>
6415 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6416 Add struct function argument and adjust.
6417 (find_func_aliases_for_call): Likewise.
6418 (find_func_aliases): Likewise.
6419 (find_func_clobbers): Likewise.
6420 (intra_create_variable_infos): Likewise.
6421 (compute_points_to_sets): Likewise.
6422 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6424 2014-04-15 Richard Biener <rguenther@suse.de>
6426 * tree.c (iterative_hash_expr): Use enum tree_code_class
6427 to store TREE_CODE_CLASS.
6428 (tree_block): Likewise.
6429 (tree_set_block): Likewise.
6430 * tree.h (fold_build_pointer_plus_loc): Use
6431 convert_to_ptrofftype_loc.
6433 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6436 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6439 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6441 * cfgloop.h (struct loop): Move force_vectorize down.
6442 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6443 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6444 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6445 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6446 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6447 * tree-core.h (enum annot_expr_kind): Add new kind values.
6448 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6449 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6451 * tree.def (ANNOTATE_EXPR): Tweak comment.
6453 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6455 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6458 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6460 * tree.h (TYPE_IDENTIFIER): Declare.
6461 * tree.c (subrange_type_for_debug_p): Use it.
6462 * godump.c (go_format_type): Likewise.
6463 * dwarf2out.c (is_cxx_auto, modified_type_die,
6464 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6465 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6467 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6470 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6472 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6474 * config/i386/i386.c (examine_argument): Return bool. Return true if
6475 parameter should be passed in memory.
6476 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6477 (construct_container): Update calls to examine_argument.
6478 (function_arg_advance_64): Ditto.
6479 (return_in_memory_32): Merge with ix86_return_in_memory.
6480 (return_in_memory_64): Ditto.
6481 (return_in_memory_ms_64): Ditto.
6483 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6485 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6486 * coverage.c (coverage_compute_profile_id): Handle externally visible
6489 2014-04-14 Martin Jambor <mjambor@suse.cz>
6491 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6492 DECL_DISREGARD_INLINE_LIMITS functions.
6494 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6497 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6499 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6502 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6503 optimize_insn_for_speed_p instead of
6504 optimize_function_for_speed_p.
6506 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6508 * doc/invoke.texi (free): Document AArch64.
6510 2014-04-14 Richard Biener <rguenther@suse.de>
6512 PR tree-optimization/60042
6513 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6514 (insert_into_preds_of_block): Do not prevent PHI insertion
6515 for REFERENCE exprs here ...
6516 (eliminate_dom_walker::before_dom_children): ... but prevent
6517 their use here under similar conditions when applied to the
6518 IL after PRE optimizations.
6520 2014-04-14 Richard Biener <rguenther@suse.de>
6522 * passes.def: Move early points-to after early SRA.
6524 2014-04-14 Richard Biener <rguenther@suse.de>
6526 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6527 check for which sign-changes we allow when forwarding
6528 a converted value into a switch.
6530 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6532 * stor-layout.c (place_field): Finalize non-constant offset for the
6535 2014-04-14 Richard Biener <rguenther@suse.de>
6537 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6539 (expand_switch_using_bit_tests_p): Likewise.
6540 (process_switch): Compute and pass on speed_p based on the
6542 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6543 optimize_bb_for_speed_p.
6545 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6547 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6548 * function.h (struct function): Rename has_force_vect_loops into
6549 has_force_vectorize_loops.
6550 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6551 (input_struct_function_base): Likewise.
6552 * lto-streamer-out.c (output_cfg): Likewise.
6553 (output_struct_function_base): Likewise.
6554 * omp-low.c (expand_omp_simd): Likewise.
6555 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6556 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6557 (version_loop_for_if_conversion): Likewise.
6558 (tree_if_conversion): Likewise.
6559 (main_tree_if_conversion): Likewise.
6560 (gate_tree_if_conversion): Likewise.
6561 * tree-inline.c (copy_loops): Likewise.
6562 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6563 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6564 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6565 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6566 * tree-vectorizer.c (vectorize_loops): Likewise.
6567 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6569 2014-04-14 Richard Biener <rguenther@suse.de>
6572 * lto-streamer-out.c (wrap_refs): New function.
6573 (lto_output): Wrap symbol references in global initializes in
6574 type-preserving MEM_REFs.
6576 2014-04-14 Christian Bruel <christian.bruel@st.com>
6578 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6580 2014-04-14 Christian Bruel <christian.bruel@st.com>
6582 * config/sh/sh.md (setmemqi): New expand pattern.
6583 * config/sh/sh.h (CLEAR_RATIO): Define.
6584 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6585 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6587 2014-04-14 Richard Biener <rguenther@suse.de>
6590 * fold-const.c (negate_expr_p): Don't negate directional rounding
6592 (fold_negate_expr): Likewise.
6594 2014-04-14 Richard Biener <rguenther@suse.de>
6596 PR tree-optimization/59817
6597 PR tree-optimization/60453
6598 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6599 recursion to catch all CHRECs in the scalar evolution and restrict
6600 the predicate for the remains appropriately.
6602 2014-04-12 Catherine Moore <clm@codesourcery.com>
6604 * config/mips/constraints.md: Add new register constraint "kb".
6605 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6606 (*movhi_internal): Likewise.
6607 (*movqi_internal): Likewise.
6608 * config/mips/mips.h (M16_STORE_REGS): New register class.
6609 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6610 (REG_CLASS_CONTENTS): Likewise.
6611 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6613 2014-04-11 Tobias Burnus <burnus@net-b.de>
6616 * doc/invoke.texi (-Wformat-signedness): Document it.
6617 (Wformat=2): Mention that this enables -Wformat-signedness.
6619 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6621 * common/config/epiphany/epiphany-common.c
6622 (epiphany_option_optimization_table): Enable section anchors by
6623 default at -O1 or higher.
6624 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6625 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6626 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6627 carries no extra cost.
6628 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6629 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6630 * config/epiphany/predicates.md (memclob_operand): New predicate.
6631 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6632 Use memclob_operand predicate and X constraint for operand 3.
6634 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6636 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6637 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6640 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6642 PR rtl-optimization/60651
6643 * mode-switching.c (optimize_mode_switching): Make sure to emit
6644 sets of a lower numbered entity before sets of a higher numbered
6645 entity to a mode of the same or lower priority.
6646 When creating a seginfo for a basic block that starts with a code
6647 label, move the insertion point past the code label.
6648 (new_seginfo): Document and enforce requirement that
6649 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6650 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6651 * doc/tm.texi: Regenerate.
6653 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6656 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6658 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6660 * BASE-VER: Set to 4.10.0.
6662 2014-04-11 Tobias Burnus <burnus@net-b.de>
6665 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6666 * doc/gcc.texi (Service): Update description in the @menu
6667 * doc/invoke.texi (Option Summary): Remove misplaced and
6670 2014-04-11 Steve Ellcey <sellcey@mips.com>
6671 Jakub Jelinek <jakub@redhat.com>
6674 * expr.c (convert_move): Use emit_store_flag_force instead of
6675 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6678 2014-04-11 Richard Biener <rguenther@suse.de>
6681 * varasm.c (assemble_alias): Avoid endless error reporting
6682 recursion by setting TREE_ASM_WRITTEN.
6684 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6686 * config/s390/s390.md: Add a splitter for NOT rtx.
6688 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6690 PR rtl-optimization/60663
6691 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6693 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6694 Jakub Jelinek <jakub@redhat.com>
6697 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6698 flag from decl_node to node.
6700 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6703 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6704 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6705 ameliorating the cases where it can be.
6707 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6710 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6712 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6713 (loadsync_<mode>): Change mode.
6714 (load_quadpti, store_quadpti): New.
6715 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6716 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6717 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6719 2014-04-09 Cong Hou <congh@google.com>
6722 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6725 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6727 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6728 instead of vnor to exploit possible fusion opportunity in the
6730 (altivec_expand_vec_perm_const_le): Likewise.
6732 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6734 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6735 (loadsync_<mode>): Change mode.
6736 (load_quadpti, store_quadpti): New.
6737 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6738 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6740 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6743 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6744 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6745 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6747 2014-04-08 Richard Biener <rguenther@suse.de>
6750 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6751 a 64bit widest int print double-int similar to on HWI64 hosts.
6753 2014-04-08 Richard Biener <rguenther@suse.de>
6755 PR tree-optimization/60785
6756 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6757 default defs properly.
6759 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6761 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6762 (Weffc++): Likewise.
6764 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6766 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6767 set completep to false rather than true.
6769 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6772 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6773 ARM_TARGET2_DWARF_FORMAT.
6775 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6778 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6779 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6782 2014-04-07 Richard Biener <rguenther@suse.de>
6784 PR tree-optimization/60766
6785 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6786 (may_eliminate_iv): Convert cand_value_at result to desired type.
6788 2014-04-07 Jason Merrill <jason@redhat.com>
6791 * common.opt (-fno-gnu-unique): Add.
6792 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6794 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6796 * haifa-sched.c: Fix outdated function reference and minor
6797 grammar errors in introductory comment.
6799 2014-04-07 Richard Biener <rguenther@suse.de>
6802 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6804 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6806 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6809 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6810 size accounting for thunks.
6811 (pa_asm_output_mi_thunk): Use final_start_function() and
6812 final_end_function() to output function start and end directives.
6814 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6816 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6817 device specific ISA/ feature information. Remove short_sp and
6818 errata_skip ds. Add avr_device_specific_features enum to have device
6820 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6821 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6822 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6823 updated device specific info.
6824 * config/avr/avr-mcus.def: Merge device specific details to
6825 dev_attribute field.
6826 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6828 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6829 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6830 assembler if RMW isa supported by current device.
6831 * config/avr/genmultilib.awk: Update as device info structure changed.
6832 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6834 2014-04-04 Cong Hou <congh@google.com>
6836 PR tree-optimization/60656
6837 * tree-vect-stmts.c (supportable_widening_operation):
6838 Fix a bug that elements in a vector with vect_used_by_reduction
6839 property are incorrectly reordered when the operation on it is not
6840 consistant with the one in reduction operation.
6842 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6844 PR rtl-optimization/60155
6845 * gcse.c (record_set_data): New function.
6846 (single_set_gcse): New function.
6847 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6848 (hoist_code): Likewise.
6849 (get_pressure_class_and_nregs): Likewise.
6851 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6853 * explow.c (probe_stack_range): Emit a final optimization blockage.
6855 2014-04-04 Anthony Green <green@moxielogic.com>
6857 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6860 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6863 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6864 flags are set only during streaming.
6865 * ipa.c (process_references, walk_polymorphic_call_targets,
6866 symtab_remove_unreachable_nodes): Drop bodies of always inline
6867 after early inlining.
6868 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6870 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6871 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6874 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6877 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6880 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6882 (cortex_a53_fdivd): Likewise.
6884 2014-04-04 Martin Jambor <mjambor@suse.cz>
6887 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6889 * cgraph.c (clone_of_p): Also return true if thunks match.
6890 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6891 cgraph_function_or_thunk_node and an obsolete comment.
6892 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6894 (build_function_decl_skip_args): Likewise.
6895 (set_new_clone_decl_and_node_flags): New function.
6896 (duplicate_thunk_for_node): Likewise.
6897 (redirect_edge_duplicating_thunks): Likewise.
6898 (cgraph_clone_node): New parameter args_to_skip, pass it to
6899 redirect_edge_duplicating_thunks which is called instead of
6900 cgraph_redirect_edge_callee.
6901 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6902 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6904 2014-04-04 Jeff Law <law@redhat.com>
6907 * config/arm/predicates.md (const_int_I_operand): New predicate.
6908 (const_int_M_operand): Similarly.
6909 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6911 (insv_t2, extv_reg, extzv_t2): Likewise.
6912 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6913 (pop_multiple_with_writeback_and_return): Likewise.
6914 (vfp_pop_multiple_with_writeback): Likewise
6916 2014-04-04 Richard Biener <rguenther@suse.de>
6919 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6920 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6922 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6923 * gimplify.c (gimple_add_tmp_var_fn): New function.
6924 * gimple-expr.h (create_tmp_reg_fn): Declare.
6925 * gimple-expr.c (create_tmp_reg_fn): New function.
6926 * gimple-low.c (record_vars_into): Don't change cfun.
6927 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6928 code generation without cfun.
6930 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6933 * Makefile.in (install-driver): Fix shell scripting.
6935 2014-04-03 Cong Hou <congh@google.com>
6937 PR tree-optimization/60505
6938 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6939 threshold of number of iterations below which no vectorization
6941 * tree-vect-loop.c (new_loop_vec_info):
6942 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6943 * tree-vect-loop.c (vect_analyze_loop_operations):
6944 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6945 * tree-vect-loop.c (vect_transform_loop):
6946 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6947 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6948 of iterations of the loop and see if we should build the epilogue.
6950 2014-04-03 Richard Biener <rguenther@suse.de>
6952 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6953 (streamer_tree_cache_create): Adjust.
6954 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6955 to allow optional nodes array.
6956 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6957 (streamer_tree_cache_append): Likewise.
6958 (streamer_tree_cache_create): Create nodes array optionally
6959 as specified by parameter.
6960 * lto-streamer-out.c (create_output_block): Avoid maintaining
6961 the node array in the writer cache.
6962 (DFS_write_tree): Remove assertion.
6963 (produce_asm_for_decls): Free the out decl state hash table early.
6964 * lto-streamer-in.c (lto_data_in_create): Adjust for
6965 streamer_tree_cache_create prototype change.
6967 2014-04-03 Richard Biener <rguenther@suse.de>
6969 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6970 set TREE_CHAIN to NULL_TREE.
6972 2014-04-03 Richard Biener <rguenther@suse.de>
6974 PR tree-optimization/60740
6975 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6976 over all GIMPLE_COND operands.
6978 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6980 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6981 (Weffc++): Remove Scott's numbering, merge lists and reference
6984 2014-04-03 Nick Clifton <nickc@redhat.com>
6986 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6989 2014-04-03 Martin Jambor <mjambor@suse.cz>
6991 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6992 mention gcc_unreachable before failing.
6993 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6996 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6999 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7000 inconsistent code and instead mark the context inconsistent.
7001 (possible_polymorphic_call_targets): For inconsistent contexts
7002 return empty complete list.
7004 2014-04-02 Anthony Green <green@moxielogic.com>
7006 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7007 (extendqisi2, extendhisi2): Define.
7008 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7009 (WCHAR_TYPE): Change to unsigned int.
7011 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7013 PR tree-optimization/60733
7014 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7015 insertion point for PHI candidates to be the end of the feeding
7016 block for the PHI argument.
7018 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7020 PR rtl-optimization/60650
7021 * lra-constraints.c (process_alt_operands): Decrease reject for
7022 earlyclobber matching.
7024 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7026 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7028 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7030 * config/spu/spu.c (pad_bb): Do not crash when the last
7031 insn is CODE_FOR_blockage.
7033 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7035 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7036 lies outside the target mode.
7038 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7041 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7042 software floating point or no floating point registers, do not
7043 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7044 in GPRs that occurs after we tested for GPRs that would never be
7047 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7048 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7049 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7050 specifically allow DDmode, since that does not use the SPE SIMD
7053 2014-04-02 Richard Biener <rguenther@suse.de>
7056 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7057 MODE_INTs. Properly use negv_optab.
7058 (expand_abs): Likewise.
7060 2014-04-02 Richard Biener <rguenther@suse.de>
7063 * Makefile.in (install-driver): Guard extra installs with special
7066 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7068 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7071 2014-04-01 Richard Henderson <rth@redhat.com>
7074 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7075 alternative enabled before register allocation.
7077 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7079 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7080 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7082 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7083 (nios2_got_address): Update nios2_large_got_address call site.
7084 (nios2_delegitimize_address): New function.
7085 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7086 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7087 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7089 2014-04-01 Martin Husemann <martin@duskware.de>
7091 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7094 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7096 PR rtl-optimization/60604
7097 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7098 check from register_operand.
7099 (register_operand): Redefine in terms of general_operand.
7100 (nonmemory_operand): Use register_operand for the non-constant cases.
7102 2014-04-01 Richard Biener <rguenther@suse.de>
7104 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7106 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7108 * doc/invoke.texi (mapp-regs): Clarify.
7110 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7112 * config/i386/avx512fintrin.h (__v32hi): Define type.
7113 (__v64qi): Likewise.
7114 (_mm512_set1_epi8): Define.
7115 (_mm512_set1_epi16): Define.
7116 (_mm512_set4_epi32): Define.
7117 (_mm512_set4_epi64): Define.
7118 (_mm512_set4_pd): Define.
7119 (_mm512_set4_ps): Define.
7120 (_mm512_setr4_epi64): Define.
7121 (_mm512_setr4_epi32): Define.
7122 (_mm512_setr4_pd): Define.
7123 (_mm512_setr4_ps): Define.
7124 (_mm512_setzero_epi32): Define.
7126 2014-03-31 Martin Jambor <mjambor@suse.cz>
7129 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7130 callsite_arguments_match_p. Updated all callers. Also check types of
7131 corresponding formal parameters and actual arguments.
7132 (not_all_callers_have_enough_arguments_p) Renamed to
7133 some_callers_have_mismatched_arguments_p.
7135 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7137 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7139 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7142 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7145 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7147 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7148 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7150 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7151 Use FMAMODE_NOVF512 mode iterator.
7152 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7153 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7154 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7155 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7157 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7158 Use VF_128_256 mode iterator.
7159 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7162 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7164 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7165 static chain if needed.
7167 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7170 * lra-constraints.c (index_part_to_reg): New.
7171 (process_address): Use it.
7173 2014-03-27 Jeff Law <law@redhat.com>
7174 Jakub Jelinek <jakub@redhat.com>
7177 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7178 gen_rtx_{PLUS,MULT} to build up the address expression.
7180 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7181 creating non-canonical RTL.
7183 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7186 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7187 functions; reorganize to make cheap checks first.
7188 (inline_small_functions): Do not estimate growth when dumping;
7190 * ipa-inline.h (inline_summary): Add min_size.
7191 (growth_likely_positive): New function.
7192 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7193 (set_cond_stmt_execution_predicate): Cleanup.
7194 (estimate_edge_size_and_time): Compute min_size.
7195 (estimate_calls_size_and_time): Likewise.
7196 (estimate_node_size_and_time): Likewise.
7197 (inline_update_overall_summary): Update min_size.
7198 (do_estimate_edge_time): Likewise.
7199 (do_estimate_edge_size): Update.
7200 (do_estimate_edge_hints): Update.
7201 (growth_likely_positive): New function.
7203 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7206 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7207 also if addr has VOIDmode.
7209 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7211 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7212 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7214 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7215 instructions as well as AdvancedSIMD loads.
7217 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7219 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7220 Use crypto_aese type.
7221 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7222 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7223 crypto_aese, crypto_aesmc. Move to types.md.
7224 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7226 * config/arm/iterators.md (crypto_type): Likewise.
7228 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7230 * cgraph.c: Include expr.h and tree-dfa.h.
7231 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7234 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7237 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7238 regs from checking multi-reg pseudos.
7240 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7242 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7244 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7246 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7247 if it would clobber the stack pointer, even temporarily.
7249 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7251 * mode-switching.c: Make small adjustments to the top comment.
7253 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7255 * config/rs6000/constraints.md (wD constraint): New constraint to
7256 match the constant integer to get the top DImode/DFmode out of a
7257 vector in a VSX register.
7259 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7260 match the constant integer to get the top DImode/DFmode out of a
7261 vector in a VSX register.
7263 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7266 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7269 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7270 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7272 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7273 Optimize vec_extract of 64-bit values, where the value being
7274 extracted is in the top word, where we can use scalar
7275 instructions. Add direct move and store support. Combine the big
7276 endian/little endian vector select load support into a single insn.
7277 (vsx_extract_<mode>_internal1): Likewise.
7278 (vsx_extract_<mode>_internal2): Likewise.
7279 (vsx_extract_<mode>_load): Likewise.
7280 (vsx_extract_<mode>_store): Likewise.
7281 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7282 combined into vsx_extract_<mode>_load.
7283 (vsx_extract_<mode>_one_le): Likewise.
7285 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7286 define the top 64-bit vector element.
7288 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7291 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7292 Document vec_vbpermq builtin.
7295 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7296 enable use of xxsldwi and xxpermdi builtin functions.
7297 (vec_xxpermdi): Likewise.
7299 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7300 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7302 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7304 PR rtl-optimization/60650
7305 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7307 (find_spills_for): New.
7308 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7309 Spill all pseudos on the second iteration.
7311 2014-03-27 Marek Polacek <polacek@redhat.com>
7314 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7317 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7319 * config/s390/s390.c (s390_can_use_return_insn): Check for
7320 call-saved FPRs on 31 bit.
7322 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7325 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7326 if they need regimplification, just drop them instead of
7327 calling gimple_regimplify_operands on them.
7329 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7332 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7333 (aarch64_frame_pointer_required): Adjust logic.
7334 (aarch64_can_eliminate): Adjust logic.
7335 (aarch64_override_options_after_change): Adjust logic.
7337 2014-03-27 Dehao Chen <dehao@google.com>
7339 * ipa-inline.c (early_inliner): Update node's inline info.
7341 2014-03-26 Dehao Chen <dehao@google.com>
7343 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7344 compiler inserted conditional jumps for NAN float check.
7346 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7348 * ubsan.h (ubsan_create_data): Change second argument's type
7349 to const location_t *.
7350 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7352 (ubsan_create_data): Change second argument to const location_t *PLOC.
7353 Create Loc field whenever PLOC is non-NULL.
7354 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7355 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7359 * real.c (real_to_integer2): Change type of low to UHWI.
7361 2014-03-26 Tobias Burnus <burnus@net-b.de>
7363 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7364 (CILK_SELF_SPECS): New define.
7365 (driver_self_specs): Use it.
7367 2014-03-26 Richard Biener <rguenther@suse.de>
7369 * tree-pretty-print.c (percent_K_format): Implement special
7370 case for LTO and its stripped down BLOCK tree.
7372 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7375 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7377 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7378 one range is range_int_cst_p, but not both, at least optimize
7379 addition/subtraction of 0 and multiplication by 0 or 1.
7380 * gimple-fold.c (gimple_fold_call): Fold
7381 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7382 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7383 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7385 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7387 PR rtl-optimization/60452
7388 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7389 <case REG>: Return 1 for invalid offsets from the frame pointer.
7391 2014-03-26 Marek Polacek <polacek@redhat.com>
7394 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7397 2014-03-26 Marek Polacek <polacek@redhat.com>
7400 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7403 2014-03-26 Marek Polacek <polacek@redhat.com>
7406 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7407 multiplication in unsigned type.
7409 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7411 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7413 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7415 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7417 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7420 * cif-code.def (UNREACHABLE) New code.
7421 * ipa-inline.c (inline_small_functions): Skip edges to
7422 __builtlin_unreachable.
7423 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7424 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7425 predicate to __bulitin_unreachable.
7426 (set_cond_stmt_execution_predicate): Fix issue when
7427 invert_tree_comparison returns ERROR_MARK.
7428 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7429 propagate to inline clones.
7430 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7432 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7433 * cgraphclones.c (cgraph_clone_node): If call destination is already
7434 ureachable, do not redirect it back.
7435 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7438 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7440 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7441 Do not modify inline clones.
7443 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7445 * config/i386/i386.md (general_sext_operand): New mode attr.
7446 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7447 don't generate (sign_extend (const_int)).
7448 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7449 operands[2]. Use We constraint instead of <i> and
7450 <general_sext_operand> predicate instead of <general_operand>.
7451 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7452 * config/i386/constraints.md (We): New constraint.
7453 * config/i386/predicates.md (x86_64_sext_operand,
7454 sext_operand): New predicates.
7456 2014-03-25 Martin Jambor <mjambor@suse.cz>
7459 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7460 inconsistent devirtualizations to __builtin_unreachable.
7462 2014-03-25 Marek Polacek <polacek@redhat.com>
7465 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7467 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7469 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7470 order of elements for big-endian.
7472 2014-03-25 Richard Biener <rguenther@suse.de>
7475 * gimplify-me.c (gimple_regimplify_operands): Update the
7478 2014-03-25 Martin Jambor <mjambor@suse.cz>
7481 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7482 (lto_output_varpool_node): Likewise.
7483 (input_overwrite_node): Likewise.
7484 (input_varpool_node): Likewise.
7486 2014-03-25 Richard Biener <rguenther@suse.de>
7488 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7489 (run_gcc): Likewise.
7491 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7493 * combine.c (simplify_compare_const): Add MODE argument.
7494 Handle mode_width 0 as very large mode_width.
7495 (try_combine, simplify_comparison): Adjust callers.
7497 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7498 type to avoid signed integer overflow.
7499 * explow.c (plus_constant): Likewise.
7501 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7503 * doc/generic.texi: Correct typos.
7505 2014-03-24 Tobias Burnus <burnus@net-b.de>
7507 * doc/invoke.texi (-flto): Expand section about
7508 using static libraries with LTO.
7510 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7512 PR rtl-optimization/60501
7513 * optabs.def (addptr3_optab): New optab.
7514 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7515 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7516 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7518 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7520 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7522 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7524 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7527 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7528 (_mm256_undefined_ps): Define.
7529 (_mm256_undefined_pd): Define.
7530 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7531 (_mm_undefined_pd): Define.
7532 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7533 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7534 (_mm512_undefined_ps): Define.
7535 (_mm512_undefined_pd): Define.
7536 Use _mm*_undefined_*.
7537 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7539 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7541 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7542 (lshr_simd): DI mode added.
7543 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7544 (aarch64_ushr_simddi): Likewise.
7545 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7546 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7547 (vshrd_n_u64): Likewise.
7549 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7551 * Makefile.in (s-macro_list): Depend on cc1.
7553 2014-03-23 Teresa Johnson <tejohnson@google.com>
7555 * ipa-utils.c (ipa_print_order): Use specified dump file.
7557 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7559 PR rtl-optimization/60601
7560 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7562 * gcc.c (eval_spec_function): Initialize save_growing_value.
7564 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7567 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7568 code == MINUS_EXPR, never swap op0 with op1.
7570 * toplev.c (init_local_tick): Avoid signed integer multiplication
7572 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7573 shift by first operand's bitsize.
7575 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7578 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7580 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7581 TARGET_ISA_64BIT_P(x).
7583 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7585 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7586 pattern for vector nor instead of subtract from splat(-1).
7587 (altivec_expand_vec_perm_const_le): Likewise.
7589 2014-03-21 Richard Henderson <rth@twiddle.net>
7592 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7593 related insns after epilogue_completed.
7595 2014-03-21 Martin Jambor <mjambor@suse.cz>
7598 * cgraph.h (symtab_node): New flag body_removed.
7599 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7600 when removing bodies.
7601 * symtab.c (dump_symtab_base): Dump body_removed flag.
7602 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7603 had their bodies removed.
7605 2014-03-21 Martin Jambor <mjambor@suse.cz>
7608 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7611 2014-03-21 Richard Biener <rguenther@suse.de>
7613 PR tree-optimization/60577
7614 * tree-core.h (struct tree_base): Document nothrow_flag use
7615 in VAR_DECL_NONALIASED.
7616 * tree.h (VAR_DECL_NONALIASED): New.
7617 (may_be_aliased): Adjust.
7618 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7620 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7622 * expr.c (expand_expr_real_1): Remove outdated comment.
7624 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7627 * ira.c (adjust_cleared_regs): Call copy_rtx on
7628 *reg_equiv[REGNO (loc)].src_p before passing it to
7629 simplify_replace_fn_rtx.
7632 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7633 into CONST, put pic register as first operand of PLUS. Use
7634 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7636 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7638 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7640 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7642 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7643 around for store forwarding issue in the FPU on the UT699.
7644 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7645 loads and operations if -mfix-ut699 is specified.
7646 (divtf3_hq): Tweak attribute.
7647 (sqrttf2_hq): Likewise.
7649 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7651 * calls.c (store_one_arg): Remove incorrect const qualification on the
7652 type of the temporary.
7653 * cfgexpand.c (expand_return): Likewise.
7654 * expr.c (expand_constructor): Likewise.
7655 (expand_expr_real_1): Likewise.
7657 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7659 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7662 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7665 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7667 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7669 * config/arm/aarch-common-protos.h
7670 (alu_cost_table): Fix spelling of "extend".
7671 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7673 2014-03-19 Richard Biener <rguenther@suse.de>
7676 * tree-core.h (tree_type_common): Re-order pointer members
7677 to reduce recursion depth during GC walks.
7679 2014-03-19 Marek Polacek <polacek@redhat.com>
7682 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7683 before accessing it.
7685 2014-03-19 Richard Biener <rguenther@suse.de>
7688 * lto-streamer-in.c (input_function): In WPA stage do not drop
7691 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7693 PR tree-optimization/60559
7694 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7695 with build_zero_cst assignment.
7697 2014-03-18 Kai Tietz <ktietz@redhat.com>
7699 PR rtl-optimization/56356
7700 * sdbout.c (sdbout_parms): Verify that parms'
7701 incoming argument is valid.
7702 (sdbout_reg_parms): Likewise.
7704 2014-03-18 Richard Henderson <rth@redhat.com>
7707 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7708 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7709 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7711 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7713 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7714 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7715 Italicize plugin event names in description. Explain that
7716 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7717 Remind that no GCC functions should be called after PLUGIN_FINISH.
7718 Explain what pragmas with expansion are.
7720 2014-03-18 Martin Liska <mliska@suse.cz>
7722 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7723 gimple call statement is update.
7724 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7725 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7727 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7730 * ubsan.c (ubsan_instrument_unreachable): Call
7731 initialize_sanitizer_builtins.
7732 (ubsan_pass): Likewise.
7735 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7736 varpool_finalize_decl instead of rest_of_decl_compilation.
7738 2014-03-18 Richard Biener <rguenther@suse.de>
7740 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7741 by using bitmap_and_compl instead of bitmap_and_compl_into.
7742 (df_rd_transfer_function): Likewise.
7744 2014-03-18 Richard Biener <rguenther@suse.de>
7746 * doc/lto.texi (fresolution): Fix typo.
7748 2014-03-18 Richard Biener <rguenther@suse.de>
7750 * doc/invoke.texi (flto): Update for changes in 4.9.
7752 2014-03-18 Richard Biener <rguenther@suse.de>
7754 * doc/loop.texi: Remove section on the removed lambda framework.
7755 Update loop docs with recent changes in preserving loop structure.
7757 2014-03-18 Richard Biener <rguenther@suse.de>
7759 * doc/lto.texi (-fresolution): Document.
7761 2014-03-18 Richard Biener <rguenther@suse.de>
7763 * doc/contrib.texi: Adjust my name.
7765 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7768 * internal-fn.c: Include diagnostic-core.h.
7769 (expand_BUILTIN_EXPECT): New function.
7770 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7771 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7772 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7773 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7775 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7776 Revert 3 argument __builtin_expect code.
7777 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7778 * gimple-fold.c (gimple_fold_call): Likewise.
7779 * tree.h (fold_builtin_expect): New prototype.
7780 * builtins.c (build_builtin_expect_predicate): Add predictor
7781 argument, if non-NULL, create 3 argument __builtin_expect.
7782 (fold_builtin_expect): No longer static. Add ARG2 argument,
7783 pass it through to build_builtin_expect_predicate.
7784 (fold_builtin_2): Adjust caller.
7785 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7786 * internal-fn.def (BUILTIN_EXPECT): New.
7788 2014-03-18 Tobias Burnus <burnus@net-b.de>
7791 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7792 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7793 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7795 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7798 * predict.c (combine_predictions_for_bb): Fix up formatting.
7799 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7800 fill what it points to if non-NULL.
7801 (tree_predict_by_opcode): Adjust caller, use the predictor.
7802 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7804 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7806 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7807 proper constant for the store mode.
7809 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7811 * symtab.c (change_decl_assembler_name): Fix transparent alias
7814 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7816 * config/aarch64/aarch64.c: Correct the comments about the
7817 aarch64 stack layout.
7819 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7821 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7822 check for GF_OMP_FOR_KIND_FOR.
7824 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7826 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7827 ymm and zmm register names.
7829 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7832 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7833 note creation for the 2010-08-31 changes.
7835 2014-03-17 Marek Polacek <polacek@redhat.com>
7838 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7839 as -fno-tree-loop-vectorize.
7840 (expand_omp_simd): Likewise.
7842 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7844 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7845 (eligible_for_call_delay): New prototype.
7846 * config/sparc/sparc.c (tls_call_delay): Rename into...
7847 (eligible_for_call_delay): ...this. Return false if the instruction
7848 cannot be put in the delay slot of a branch.
7849 (eligible_for_restore_insn): Simplify.
7850 (eligible_for_return_delay): Return false if the instruction cannot be
7851 put in the delay slot of a branch and simplify.
7852 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7853 put in the delay slot of a branch.
7854 * config/sparc/sparc.md (fix_ut699): New attribute.
7855 (tls_call_delay): Delete.
7856 (in_call_delay): Reimplement.
7857 (eligible_for_sibcall_delay): Rename into...
7858 (in_sibcall_delay): ...this.
7859 (eligible_for_return_delay): Rename into...
7860 (in_return_delay): ...this.
7861 (in_branch_delay): Reimplement.
7862 (in_uncond_branch_delay): Delete.
7863 (in_annul_branch_delay): Delete.
7865 2014-03-14 Richard Henderson <rth@redhat.com>
7868 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7869 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7870 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7871 (floathi splitters): Remove.
7872 (float<SWI48x>xf2): New pattern.
7873 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7874 code that tried to handle DImode for 32-bit, but which was excluded
7875 by the pattern's condition. Drop allocation of stack temporary.
7876 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7877 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7878 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7879 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7880 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7881 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7882 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7883 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7884 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7885 (*float<SWI48x><X87MODEF>2_i387): Remove.
7886 (all float _with_temp splitters): Remove.
7887 (*float<SWI48x><MODEF>2_i387): New pattern.
7888 (*float<SWI48><MODEF>2_sse): New pattern.
7889 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7890 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7892 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7893 Marek Polacek <polacek@redhat.com>
7896 * common.opt (dump_base_name_prefixed): New Variable.
7897 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7898 if x_dump_base_name_prefixed is already set, set it at the end.
7900 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7902 PR rtl-optimization/60508
7903 * lra-constraints.c (get_reload_reg): Add new parameter
7905 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7906 Pass the new parameter values.
7908 2014-03-14 Richard Biener <rguenther@suse.de>
7910 * common.opt: Revert unintented changes from r205065.
7913 2014-03-14 Richard Biener <rguenther@suse.de>
7916 * cfghooks.c (split_block): Properly adjust all loops the
7917 block was a latch of.
7919 2014-03-14 Martin Jambor <mjambor@suse.cz>
7922 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7925 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7928 * config/avr/avr.c (avr_set_current_function): Pass function name
7929 through default_strip_name_encoding before sanity checking instead
7930 of skipping the first char of the assembler name.
7932 2014-03-13 Richard Henderson <rth@redhat.com>
7935 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7936 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7937 * config/i386/i386-protos.h: Likewise.
7938 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7939 in the expander instead of a splitter.
7940 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7941 any possibility of requiring a memory.
7942 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7943 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7944 (fp branch splitters): Update for ix86_split_fp_branch.
7945 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7946 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7947 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7948 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7949 (*fop_<MODEF>_3_i387): Likewise.
7950 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7951 (splitters for the fop_* register patterns): Remove.
7952 (fscalexf4_i387): Rename from *fscalexf4_i387.
7953 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7955 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7957 PR tree-optimization/59779
7958 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7959 type for bitsize and maxsize instead of HOST_WIDE_INT.
7961 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7963 PR rtl-optimization/57320
7964 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7965 the CFG after thread_prologue_and_epilogue_insns.
7967 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7969 PR rtl-optimization/57189
7970 * lra-constraints.c (process_alt_operands): Disfavor spilling
7973 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7975 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7977 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7979 PR tree-optimization/59025
7981 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7982 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7984 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7987 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7988 calls of avr_out_plus_1.
7990 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7992 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7993 BB's single pred and update the father loop's latch info later.
7995 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7997 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8001 (VEC_base): Likewise.
8002 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8003 registers, we need to swap double words in little endian mode.
8005 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8006 to be a container mode for 128-bit integer operations added in ISA
8007 2.07. Unlike TImode and PTImode, the preferred register set is
8008 the Altivec/VMX registers for the 128-bit operations.
8010 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8012 (rs6000_split_128bit_ok_p): Likewise.
8014 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8015 macros for creating ISA 2.07 normal and overloaded builtin
8016 functions with 3 arguments.
8017 (BU_P8V_OVERLOAD_3): Likewise.
8018 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8019 for use as overloaded functions.
8020 (VPERM_1TI_UNS): Likewise.
8021 (VSEL_1TI): Likewise.
8022 (VSEL_1TI_UNS): Likewise.
8023 (ST_INTERNAL_1ti): Likewise.
8024 (LD_INTERNAL_1ti): Likewise.
8025 (XXSEL_1TI): Likewise.
8026 (XXSEL_1TI_UNS): Likewise.
8027 (VPERM_1TI): Likewise.
8028 (VPERM_1TI_UNS): Likewise.
8029 (XXPERMDI_1TI): Likewise.
8030 (SET_1TI): Likewise.
8031 (LXVD2X_V1TI): Likewise.
8032 (STXVD2X_V1TI): Likewise.
8033 (VEC_INIT_V1TI): Likewise.
8034 (VEC_SET_V1TI): Likewise.
8035 (VEC_EXT_V1TI): Likewise.
8036 (EQV_V1TI): Likewise.
8037 (NAND_V1TI): Likewise.
8038 (ORC_V1TI): Likewise.
8039 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8040 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8042 (VADDUQM): Likewise.
8043 (VSUBCUQ): Likewise.
8044 (VADDEUQM): Likewise.
8045 (VADDECUQ): Likewise.
8046 (VSUBEUQM): Likewise.
8047 (VSUBECUQ): Likewise.
8049 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8050 __int128_t and __uint128_t types.
8051 (__uint128_type): Likewise.
8052 (altivec_categorize_keyword): Add support for vector __int128_t,
8053 vector __uint128_t, vector __int128, and vector unsigned __int128
8054 as a container type for TImode operations that need to be done in
8055 VSX/Altivec registers.
8056 (rs6000_macro_to_expand): Likewise.
8057 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8058 to support 128-bit integer instructions vaddcuq, vadduqm,
8059 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8060 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8062 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8063 for V1TImode, and set up preferences to use VSX/Altivec registers.
8064 Setup VSX reload handlers.
8065 (rs6000_debug_reg_global): Likewise.
8066 (rs6000_init_hard_regno_mode_ok): Likewise.
8067 (rs6000_preferred_simd_mode): Likewise.
8068 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8069 (easy_altivec_constant): Likewise.
8070 (output_vec_const_move): Likewise.
8071 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8073 (rs6000_expand_vector_extract): Likewise.
8074 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8076 (rs6000_const_vec): Add support for V1TImode.
8077 (rs6000_emit_le_vsx_load): Swap double words when loading or
8078 storing TImode/V1TImode.
8079 (rs6000_emit_le_vsx_store): Likewise.
8080 (rs6000_emit_le_vsx_move): Likewise.
8081 (rs6000_emit_move): Add support for V1TImode.
8082 (altivec_expand_ld_builtin): Likewise.
8083 (altivec_expand_st_builtin): Likewise.
8084 (altivec_expand_vec_init_builtin): Likewise.
8085 (altivec_expand_builtin): Likewise.
8086 (rs6000_init_builtins): Add support for V1TImode type. Add
8087 support for ISA 2.07 128-bit integer builtins. Define type names
8088 for the VSX/Altivec vector types.
8089 (altivec_init_builtins): Add support for overloaded vector
8090 functions with V1TImode type.
8091 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8092 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8094 (rs6000_split_128bit_ok_p): Likewise.
8095 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8096 __int128_t and vector __uint128_t.
8098 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8099 and mode attributes.
8106 (VS_scalar): Likewise.
8107 (VS_double): Likewise.
8108 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8110 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8111 we support the ISA 2.07 128-bit integer arithmetic instructions.
8112 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8113 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8114 and TImode types for use with the builtin functions.
8115 (V1TI_type_node): Likewise.
8116 (unsigned_V1TI_type_node): Likewise.
8117 (intTI_type_internal_node): Likewise.
8118 (uintTI_type_internal_node): Likewise.
8120 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8121 128-bit builtin functions.
8122 (UNSPEC_VADDEUQM): Likewise.
8123 (UNSPEC_VADDECUQ): Likewise.
8124 (UNSPEC_VSUBCUQ): Likewise.
8125 (UNSPEC_VSUBEUQM): Likewise.
8126 (UNSPEC_VSUBECUQ): Likewise.
8127 (VM): Add V1TImode to vector mode iterators.
8129 (VI_unit): Likewise.
8130 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8131 (altivec_vaddcuq): Likewise.
8132 (altivec_vsubuqm): Likewise.
8133 (altivec_vsubcuq): Likewise.
8134 (altivec_vaddeuqm): Likewise.
8135 (altivec_vaddecuq): Likewise.
8136 (altivec_vsubeuqm): Likewise.
8137 (altivec_vsubecuq): Likewise.
8139 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8141 (BOOL_128): Likewise.
8142 (BOOL_REGS_OUTPUT): Likewise.
8143 (BOOL_REGS_OP1): Likewise.
8144 (BOOL_REGS_OP2): Likewise.
8145 (BOOL_REGS_UNARY): Likewise.
8146 (BOOL_REGS_AND_CR0): Likewise.
8148 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8149 128-bit integer builtin support.
8150 (vec_vadduqm): Likewise.
8151 (vec_vaddecuq): Likewise.
8152 (vec_vaddeuqm): Likewise.
8153 (vec_vsubecuq): Likewise.
8154 (vec_vsubeuqm): Likewise.
8155 (vec_vsubcuq): Likewise.
8156 (vec_vsubuqm): Likewise.
8158 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8159 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8160 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8161 128-bit integer add/subtract to ISA 2.07.
8163 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8165 * config/arc/arc.c (arc_predicate_delay_insns):
8166 Fix third argument passed to conditionalize_nonjump.
8168 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8170 * config/aarch64/aarch64-builtins.c
8171 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8172 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8173 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8174 instead of __builtin_lfloor.
8175 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8177 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8179 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8180 (tree_ssa_ifcombine_bb_1): New function.
8181 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8182 is an empty forwarder block to then_bb or vice versa and then_bb
8183 and else_bb are effectively swapped.
8185 2014-03-12 Christian Bruel <christian.bruel@st.com>
8188 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8189 REG_CFA_DEF_CFA note.
8190 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8191 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8193 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8195 PR tree-optimization/60454
8196 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8198 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8200 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8201 Do not define target_cpu_default2 to generic.
8202 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8203 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8204 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8206 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8207 Marc Glisse <marc.glisse@inria.fr>
8209 PR tree-optimization/60502
8210 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8211 instead of build_low_bits_mask.
8213 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8216 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8217 if there are multiple uses, but op doesn't live on E edge.
8218 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8219 clobber stmts before __builtin_unreachable.
8221 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8223 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8224 hard_frame_pointer_rtx.
8225 * cse.c (cse_insn): Remove volatile check.
8226 * cselib.c (cselib_process_insn): Likewise.
8227 * dse.c (scan_insn): Likewise.
8229 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8231 * config/arc/arc.c (conditionalize_nonjump): New function,
8233 (arc_ifcvt): ... this.
8234 (arc_predicate_delay_insns): Use it.
8236 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8238 * config/arc/predicates.md (extend_operand): During/after reload,
8239 allow const_int_operand.
8240 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8241 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8242 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8244 (umulsi3_highpart_i): Likewise.
8246 2014-03-11 Richard Biener <rguenther@suse.de>
8248 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8249 Add asserts to guard possible wrong-code bugs.
8251 2014-03-11 Richard Biener <rguenther@suse.de>
8253 PR tree-optimization/60429
8254 PR tree-optimization/60485
8255 * tree-ssa-structalias.c (set_union_with_increment): Properly
8256 take into account all fields that overlap the shifted vars.
8257 (do_sd_constraint): Likewise.
8258 (do_ds_constraint): Likewise.
8259 (get_constraint_for_ptr_offset): Likewise.
8261 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8263 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8264 (nios2_compute_frame_layout):
8265 Add calculation of cfun->machine->fp_save_offset.
8266 (nios2_expand_prologue): Correct setting of frame pointer register
8268 (nios2_expand_epilogue): Update recovery of stack pointer from
8269 frame pointer accordingly.
8270 (nios2_initial_elimination_offset): Update calculation of offset
8271 for eliminating to HARD_FRAME_POINTER_REGNUM.
8273 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8276 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8277 cgraph_get_create_node on VAR_DECLs.
8279 2014-03-10 Richard Biener <rguenther@suse.de>
8282 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8284 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8286 * config/vms/vms.opt (vms_float_format): New variable.
8288 2014-03-08 Tobias Burnus <burnus@net-b.de>
8290 * doc/invoke.texi (-fcilkplus): Update implementation status.
8292 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8293 Richard Biener <rguenther@suse.de>
8295 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8296 consistently accross all TUs.
8297 (run_gcc): Enable -fshort-double automatically at link at link-time
8298 and disallow override.
8300 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8303 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8304 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8305 if they can't be used.
8307 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8309 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8310 for Solaris 11/x86 ld.
8311 * configure: Regenerate.
8313 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8315 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8316 (LIB_TLS_SPEC): Save as ld_tls_libs.
8317 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8318 (HAVE_AS_IX86_TLSLDM): New test.
8319 * configure, config.in: Regenerate.
8320 * config/i386/i386.c (legitimize_tls_address): Fall back to
8321 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8322 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8323 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8324 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8326 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8328 * common.opt (fira-loop-pressure): Mark as optimization.
8330 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8332 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8333 an OpenMP mappable type.
8335 2014-03-06 Matthias Klose <doko@ubuntu.com>
8337 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8338 MULTILIB_OSDIRNAMES is not defined.
8340 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8341 Meador Inge <meadori@codesourcery.com>
8344 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8345 (arm_legitimize_address): Call legitimize_tls_address for any
8346 arm_tls_referenced_p expression, handle constant addend. Call it
8347 before testing for !TARGET_ARM.
8348 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8350 2014-03-06 Richard Biener <rguenther@suse.de>
8356 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8358 * tree-streamer.c (record_common_node): Assert we don't record
8359 nodes with type double.
8360 (preload_common_node): Skip type double, complex double and double
8361 pointer since it is now frontend dependent due to fshort-double option.
8363 2014-03-06 Richard Biener <rguenther@suse.de>
8365 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8366 or -fno-lto is specified and the linker has full plugin support.
8367 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8368 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8369 * lto-wrapper.c (merge_and_complain): Merge compile-time
8370 optimization levels.
8371 (run_gcc): And pass it through to the link options.
8373 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8377 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8379 * cselib.c (remove_useless_values): Skip to avoid quadratic
8380 behavior if the condition moved from...
8381 (cselib_process_insn): ... here holds.
8383 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8386 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8387 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8390 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8391 (TM_H): Add x86-tune.def.
8393 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8395 * config/aarch64/aarch64.c (generic_tunings):
8396 Use cortexa57_extra_costs.
8398 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8401 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8402 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8403 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8406 2014-03-04 Heiher <r@hev.cc>
8408 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8409 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8411 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8413 * config/i386/predicates.md (const2356_operand): Change to ...
8414 (const2367_operand): ... this.
8415 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8417 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8418 (*avx512pf_scatterpf<mode>sf): Ditto.
8419 (avx512pf_scatterpf<mode>df): Ditto.
8420 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8421 (*avx512pf_scatterpf<mode>df): Ditto.
8422 * config/i386/i386.c (ix86_expand_builtin): Update
8423 incorrect hint operand error message.
8425 2014-03-04 Richard Biener <rguenther@suse.de>
8427 * lto-section-in.c (lto_get_section_data): Fix const cast.
8429 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8431 * tree-streamer.c (record_common_node): Assert we don't record
8432 nodes with type double.
8433 (preload_common_node): Skip type double, complex double and double
8434 pointer since it is now frontend dependent due to fshort-double option.
8436 2014-03-04 Richard Biener <rguenther@suse.de>
8439 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8440 (lto_input_toplevel_asms): Likewise.
8441 * lto-section-in.c (lto_get_section_data): Instead do it here
8444 2014-03-04 Richard Biener <rguenther@suse.de>
8446 PR tree-optimization/60382
8447 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8448 dead PHIs a reduction.
8450 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8452 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8454 (_mm_prefetch): Move out of GCC target("sse") pragma.
8455 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8456 GCC target("prfchw") pragma.
8457 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8459 * config/i386/i386.c (ix86_option_override_internal): Enable
8460 -mprfchw with -mprefetchwt1.
8462 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8464 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8467 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8469 * opts.h (CL_PCH_IGNORE): Define.
8470 * targhooks.c (option_affects_pch_p):
8471 Return false for options that have CL_PCH_IGNORE set.
8472 * opt-functions.awk: Process PchIgnore.
8473 * doc/options.texi: Document PchIgnore.
8475 * config/arc/arc.opt (misize): Add PchIgnore property.
8477 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8479 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8480 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8481 constraint on constants to permit them being loaded into
8482 GENERAL_REGS or BASE_REGS.
8484 2014-03-03 Nick Clifton <nickc@redhat.com>
8486 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8487 anti-cacnonical alternatives.
8488 (negandhi3_real): New pattern.
8489 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8491 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8493 * config/avr/avr-mcus.def: Remove atxmega16x1.
8494 * config/avr/avr-tables.opt: Regenerate.
8495 * config/avr/t-multilib: Regenerate.
8496 * doc/avr-mmcu.texi: Regenerate.
8498 2014-03-03 Tobias Grosser <tobias@grosser.es>
8499 Mircea Namolaru <mircea.namolaru@inria.fr>
8501 PR tree-optimization/58028
8502 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8505 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8507 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8508 not handled by recognizers.
8510 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8513 * function.c (expand_function_end): Don't emit
8514 clobber_return_register sequence if clobber_after is a BARRIER.
8515 * cfgexpand.c (construct_exit_block): Append instructions before
8516 return_label to prev_bb.
8518 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8520 * config/rs6000/constraints.md: Document reserved use of "wc".
8522 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8525 * ipa.c (function_and_variable_visibility): When dissolving comdat
8526 group, also set all symbols to local.
8528 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8533 2013-12-14 Jan Hubicka <jh@suse.cz>
8535 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8537 2014-03-02 Jon Beniston <jon@beniston.com>
8543 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8544 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8545 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8546 (simple_return, *simple_return): New patterns
8547 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8548 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8550 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8552 * dwarf2out.c (gen_subprogram_die): Tidy.
8554 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8557 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8558 (*mov_t_msb_neg_negc): ... this new insn.
8560 2014-02-28 Jason Merrill <jason@redhat.com>
8563 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8566 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8569 * dwarf2out.c (decltype_auto_die): New static.
8570 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8571 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8572 (is_cxx_auto): Likewise.
8574 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8576 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8577 we are not using general regs only.
8579 2014-02-28 Richard Biener <rguenther@suse.de>
8582 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8583 previous fix and only allow to remove trivial pre-headers
8584 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8585 (remove_forwarder_block): Properly update the latch of a loop.
8587 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8590 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8591 (cselib_preserved_hash_table): New.
8592 (preserve_constants_and_equivs): Move preserved vals to it.
8593 (cselib_find_slot): Look it up first.
8594 (cselib_init): Initialize it.
8595 (cselib_finish): Release it.
8596 (dump_cselib_table): Dump it.
8598 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8601 * cselib.c (remove_useless_values): Skip to avoid quadratic
8602 behavior if the condition moved from...
8603 (cselib_process_insn): ... here holds.
8605 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8608 * var-tracking.c (vt_initialize): Apply the same condition to
8609 preserve the CFA base value.
8611 2014-02-28 Joey Ye <joey.ye@arm.com>
8614 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8615 if reload in progress or completed.
8617 2014-02-28 Tobias Burnus <burnus@net-b.de>
8620 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8623 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8625 * doc/tm.texi.in (Condition Code Status): Update documention for
8626 relative locations of cc0-setter and cc0-user.
8628 2014-02-27 Jeff Law <law@redhat.com>
8630 PR rtl-optimization/52714
8631 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8632 into two independent simple sets, if I3 is a jump, ensure the
8633 pattern we place into I3 is a (set (pc) ...).
8635 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8636 Jeff Law <law@redhat.com>
8638 PR rtl-optimization/49847
8639 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8640 are in different blocks.
8641 * doc/tm.texi (Condition Code Status): Update documention for
8642 relative locations of cc0-setter and cc0-user.
8644 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8647 * lra.c (lra_emit_add): Check SUBREG too.
8649 2014-02-27 Andreas Schwab <schwab@suse.de>
8651 * config/m68k/m68k.c (m68k_option_override): Disable
8652 -flive-range-shrinkage for classic m68k.
8653 (m68k_override_options_after_change): Likewise.
8655 2014-02-27 Marek Polacek <polacek@redhat.com>
8658 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8659 -Wmaybe-uninitialized.
8661 2014-02-27 Alan Modra <amodra@gmail.com>
8664 * reload1.c (emit_input_reload_insns): When reload_override_in,
8665 set old to rl->in_reg when rl->in_reg is a subreg.
8667 2014-02-26 Richard Biener <rguenther@suse.de>
8670 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8672 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8674 * common/config/i386/predicates.md (const1256_operand): Remove.
8675 (const2356_operand): New.
8676 (const_1_to_2_operand): Remove.
8677 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8678 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8679 (*avx512pf_gatherpf<mode>sf): Ditto.
8680 (avx512pf_gatherpf<mode>df): Ditto.
8681 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8682 (*avx512pf_gatherpf<mode>df): Ditto.
8683 (avx512pf_scatterpf<mode>sf): Ditto.
8684 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8685 (*avx512pf_scatterpf<mode>sf): Ditto.
8686 (avx512pf_scatterpf<mode>df): Ditto.
8687 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8688 (*avx512pf_scatterpf<mode>df): Ditto.
8689 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8691 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8693 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8694 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8695 (_mm512_mask_testn_epi64_mask): Move to ...
8696 * config/i386/avx512cdintrin.h: Here.
8697 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8698 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8699 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8700 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8701 TARGET_AVX512F from TARGET_AVX512CD.
8703 2014-02-26 Richard Biener <rguenther@suse.de>
8706 * ipa.c (walk_polymorphic_call_targets): Properly guard
8707 call to inline_update_overall_summary.
8709 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8712 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8713 and latches only if requested. Fix latch if it is removed.
8714 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8715 LOOPS_HAVE_PREHEADERS.
8717 2014-02-25 Andrew Pinski <apinski@cavium.com>
8719 * builtins.c (expand_builtin_thread_pointer): Create a new target
8720 when the target is NULL.
8722 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8724 PR rtl-optimization/60317
8725 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8726 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8727 * lra-assigns.c: Include params.h.
8728 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8729 other reload pseudos considerations.
8731 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8733 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8734 to use canonical form for nor<mode>3.
8736 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8739 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8742 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8744 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8745 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8746 (ix86_handle_option): Handle OPT_mprefetchwt1.
8747 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8748 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8750 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8751 OPTION_MASK_ISA_PREFETCHWT1.
8752 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8753 (PTA_PREFETCHWT1): New.
8754 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8755 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8756 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8757 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8758 (*prefetch_avx512pf_<mode>_: Change into ...
8759 (*prefetch_prefetchwt1_<mode>: This.
8760 * config/i386/i386.opt (mprefetchwt1): New.
8761 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8762 (_mm_prefetch): Handle intent to write.
8763 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8765 2014-02-25 Richard Biener <rguenther@suse.de>
8768 * emit-rtl.c (mem_attrs_htab): Remove.
8769 (mem_attrs_htab_hash): Likewise.
8770 (mem_attrs_htab_eq): Likewise.
8771 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8772 (init_emit_once): Do not allocate mem_attrs_htab.
8774 2014-02-25 Richard Biener <rguenther@suse.de>
8777 * lto-opts.c (lto_write_options): Output non-explicit conservative
8778 -fwrapv, -fno-trapv and -fno-strict-overflow.
8779 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8780 (run_gcc): And pass them through.
8782 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8784 * sel-sched.c (calculate_new_fences): New parameter ptime.
8785 Calculate it as a maximum over all fence cycles.
8786 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8787 Print the final schedule timing when sched_verbose.
8789 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8791 PR rtl-optimization/60292
8792 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8793 bit fot the fence instruction.
8795 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8797 * calls.h: Fix typo in comment.
8799 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8801 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8802 adjusting offsetable addresses.
8804 2014-02-24 Guozhi Wei <carrot@google.com>
8806 * sparseset.h (sparseset_pop): Fix the wrong index.
8808 2014-02-24 Walter Lee <walt@tilera.com>
8810 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8811 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8813 * common/config/tilegx/tilegx-common.c
8814 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8815 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8817 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8818 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8819 (tilegx_gimplify_va_arg_expr): Handle big endian.
8820 (tilegx_expand_unaligned_load): Ditto.
8821 (tilegx_expand_unaligned_store): Ditto.
8822 (TARGET_RETURN_IN_MSB): New.
8823 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8824 (TARGET_ENDIAN_DEFAULT): New.
8825 (TARGET_BIG_ENDIAN): Handle big endian.
8826 (BYTES_BIG_ENDIAN): Ditto.
8827 (WORDS_BIG_ENDIAN): Ditto.
8828 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8831 * config/tilegx/tilegx.md (extv): Handle big endian.
8833 (insn_st<n>): Ditto.
8834 (insn_st<n>_add<bitsuffix>): Ditto.
8835 (insn_stnt<n>): Ditto.
8836 (insn_stnt<n>_add<bitsuffix>):Ditto.
8837 (vec_interleave_highv8qi): Handle big endian.
8838 (vec_interleave_highv8qi_be): New.
8839 (vec_interleave_highv8qi_le): New.
8840 (insn_v1int_h): Handle big endian.
8841 (vec_interleave_lowv8qi): Handle big endian.
8842 (vec_interleave_lowv8qi_be): New.
8843 (vec_interleave_lowv8qi_le): New.
8844 (insn_v1int_l): Handle big endian.
8845 (vec_interleave_highv4hi): Handle big endian.
8846 (vec_interleave_highv4hi_be): New.
8847 (vec_interleave_highv4hi_le): New.
8848 (insn_v2int_h): Handle big endian.
8849 (vec_interleave_lowv4hi): Handle big endian.
8850 (vec_interleave_lowv4hi_be): New.
8851 (vec_interleave_lowv4hi_le): New.
8852 (insn_v2int_l): Handle big endian.
8853 (vec_interleave_highv2si): Handle big endian.
8854 (vec_interleave_highv2si_be): New.
8855 (vec_interleave_highv2si_le): New.
8856 (insn_v4int_h): Handle big endian.
8857 (vec_interleave_lowv2si): Handle big endian.
8858 (vec_interleave_lowv2si_be): New.
8859 (vec_interleave_lowv2si_le): New.
8860 (insn_v4int_l): Handle big endian.
8861 * config/tilegx/tilegx.opt (mbig-endian): New option.
8862 (mlittle-endian): New option.
8863 * doc/install.texi: Document tilegxbe-linux.
8864 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8866 2014-02-24 Martin Jambor <mjambor@suse.cz>
8869 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8870 there are no parameter descriptors.
8872 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8874 PR rtl-optimization/60268
8875 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8876 initialization to ...
8877 (sched_rgn_init): ... here.
8878 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8880 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8882 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8885 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8887 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8890 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8892 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8893 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8895 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8897 * config/microblaze/predicates.md: Add cmp_op predicate.
8898 * config/microblaze/microblaze.md: Add branch_compare instruction
8899 which uses cmp_op predicate and emits cmp insn before branch.
8900 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8901 to microblaze_expand_conditional_branch and consolidate logic.
8902 (microblaze_expand_conditional_branch): emit branch_compare
8903 insn instead of handling cmp op separate from branch insn.
8905 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8907 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8910 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8912 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8913 define_insn with define_expand and new define_insn
8914 *altivec_lve<VI_char>x_internal.
8915 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8916 and new define_insn *altivec_stve<VI_char>x_internal.
8917 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8919 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8921 (altivec_expand_stvex_be): New function.
8923 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8925 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8926 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8927 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8928 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8930 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8933 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8934 instead of emit_move_insn.
8936 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8938 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8940 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8941 gen_altivec_vsumsws.
8943 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8945 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8946 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8947 (altivec_lvxl_<mode>): New define_expand incorporating
8948 -maltivec=be semantics where needed.
8949 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8950 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8951 semantics where needed.
8952 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8953 (altivec_stvx_<mode>): New define_expand incorporating
8954 -maltivec=be semantics where needed.
8955 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8956 VM2 iterator instead of V4SI.
8957 (altivec_stvxl_<mode>): New define_expand incorporating
8958 -maltivec=be semantics where needed.
8959 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8960 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8961 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8962 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8963 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8964 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8965 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8966 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8967 ALTIVEC_BUILTIN_STVXL.
8968 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8969 (altivec_expand_stvx_be): Likewise.
8970 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8971 (altivec_expand_lvx_be): Likewise.
8972 (altivec_expand_stvx_be): Likewise.
8973 (altivec_expand_builtin): Add cases for
8974 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8975 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8976 (altivec_init_builtins): Add definitions for
8977 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8978 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8980 2014-02-21 Catherine Moore <clm@codesourcery.com>
8982 * doc/invoke.texi (mvirt, mno-virt): Document.
8983 * config/mips/mips.opt (mvirt): New option.
8984 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8986 2014-02-21 Richard Biener <rguenther@suse.de>
8988 PR tree-optimization/60276
8989 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8990 (STMT_VINFO_MIN_NEG_DIST): New macro.
8991 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8992 STMT_VINFO_MIN_NEG_DIST.
8993 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8994 made for negative dependence distances still hold.
8996 2014-02-21 Richard Biener <rguenther@suse.de>
8999 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9000 DECL_INITIAL for globals not in the current function context.
9002 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9004 PR tree-optimization/56490
9005 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9006 * tree-ssa-uninit.c: Include params.h.
9007 (compute_control_dep_chain): Add num_calls argument, return false
9008 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9009 num_calls to recursive call.
9010 (find_predicates): Change dep_chain into normal array,
9011 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9012 variable and adjust compute_control_dep_chain caller.
9013 (find_def_preds): Likewise.
9015 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9017 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9018 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9020 2014-02-21 Nick Clifton <nickc@redhat.com>
9022 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9023 (pushhi1): Likewise.
9024 (popqi1): Add mode to pre_dec.
9027 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9029 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9030 mode for mask of V8SFmode permutation.
9032 2014-02-20 Richard Henderson <rth@redhat.com>
9035 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9036 a new pseudo for OLDVAL.
9038 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9041 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9042 gen_reg_rtx if d->testing_p.
9043 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9044 if d->testing_p and we will certainly return true.
9045 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9048 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9050 * emit-rtl.c (gen_reg_rtx): Assert that
9051 crtl->emit.regno_pointer_align_length is non-zero.
9053 2014-02-20 Richard Henderson <rth@redhat.com>
9056 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9057 on failure the store back into EXPECT.
9059 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9060 Sandra Loosemore <sandra@codesourcery.com>
9062 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9063 * config/nios2/nios2.c (nios2_function_profiler): Add
9064 -fPIC (flag_pic == 2) support.
9065 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9066 (nios2_large_offset_p): New function.
9067 (nios2_unspec_reloc_p): Move up position, update to use
9068 nios2_large_offset_p.
9069 (nios2_unspec_address): Remove function.
9070 (nios2_unspec_offset): New function.
9071 (nios2_large_got_address): New function.
9072 (nios2_got_address): Add large offset support.
9073 (nios2_legitimize_tls_address): Update usage of removed and new
9075 (nios2_symbol_binds_local_p): New function.
9076 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9077 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9078 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9079 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9080 processing for (const (unspec ...)).
9081 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9083 2014-02-20 Richard Biener <rguenther@suse.de>
9085 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9086 doing the substitution.
9087 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9089 2014-02-20 Martin Jambor <mjambor@suse.cz>
9092 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9093 info when checking whether lattices are bottom.
9095 2014-02-20 Richard Biener <rguenther@suse.de>
9098 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9101 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9104 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9105 parameter specifying the scaling.
9106 (inline_call): Update.
9107 (want_inline_recursively): Guard division by zero.
9108 (recursive_inlining): Update.
9109 * ipa-inline.h (clone_inlined_nodes): Update.
9111 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9114 * config/i386/i386.c (classify_argument): Pass structures of size
9115 64 bytes or less in register.
9117 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9118 Kirill Yukhin <kirill.yukhin@intel.com>
9120 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9121 (_mm_rcp28_round_ss): Ditto.
9122 (_mm_rsqrt28_round_sd): Ditto.
9123 (_mm_rsqrt28_round_ss): Ditto.
9124 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9125 (_mm_rcp14_round_ss): Ditto.
9126 (_mm_rsqrt14_round_sd): Ditto.
9127 (_mm_rsqrt14_round_ss): Ditto.
9128 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9129 the first input operand, get rid of match_dup.
9130 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9132 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9134 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9135 operand as the first input operand, set type attribute.
9136 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9138 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9139 operand as the first input operand, set type attribute.
9141 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9143 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9146 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9149 * config/i386/i386.c (construct_container): Remove TFmode check
9150 for X86_64_INTEGER_CLASS.
9152 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9155 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9156 only when -Wpsabi is enabled.
9158 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9161 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9162 passing arrays in registers are the same as for structs, so remove the
9163 special case for them.
9165 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9167 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9168 destination type, extract only the valid bits if the source type is not
9169 integral and has a different mode.
9171 2014-02-19 Richard Biener <rguenther@suse.de>
9174 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9177 2014-02-19 Richard Biener <rguenther@suse.de>
9180 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9181 (ipa_modify_call_arguments): Emit an argument load explicitely and
9182 preserve virtual SSA form there and for the replacement call.
9183 Do not update SSA form nor free dominance info.
9185 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9187 * ipa.c (function_and_variable_visibility): Also clear WEAK
9188 flag when disolving COMDAT_GROUP.
9190 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9192 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9193 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9195 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9196 do more sanity checks.
9197 (detect_type_change): Return true when giving up early.
9198 (compute_complex_assign_jump_func): Fix type parameter of
9199 ipa_set_ancestor_jf.
9200 (compute_complex_ancestor_jump_func): Likewise.
9201 (update_jump_functions_after_inlining): Fix updating of
9203 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9205 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9207 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9208 inline clones when edge disappears.
9210 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9213 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9214 Split 64-bit moves into 2 patterns. Do not allow the use of
9215 direct move for TDmode in little endian, since the decimal value
9216 has little endian bytes within a word, but the 64-bit pieces are
9217 ordered in a big endian fashion, and normal subreg's of TDmode are
9219 (mov<mode>_64bit_dm): Likewise.
9220 (movtd_64bit_nodm): Likewise.
9222 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9224 PR tree-optimization/60174
9225 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9226 statement of an SSA_NAME that occurs in an abnormal PHI node.
9228 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9231 * final.c (SEEN_BB): Remove.
9232 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9233 (final_scan_insn): Don't force_source_line on second
9234 NOTE_INSN_BASIC_BLOCK.
9236 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9239 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9240 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9241 (type_natural_mode): Warn ABI change when %zmm register is not
9242 available for AVX512F vector value passing.
9244 2014-02-18 Kai Tietz <ktietz@redhat.com>
9247 * config/i386/i386.c (ix86_expand_prologue): Use value in
9248 rax register as displacement when restoring %r10 or %rax.
9249 Fix wrong offset when restoring both registers.
9251 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9253 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9254 assertion with conditional return.
9256 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9257 Uros Bizjak <ubizjak@gmail.com>
9260 * config/i386/driver-i386.c (host_detect_local_cpu): If
9261 YMM state is not saved by the OS, also clear has_f16c. Move
9262 CPUID 0x80000001 handling before YMM state saving checking.
9264 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9266 PR rtl-optimization/58960
9267 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9268 factored out from ...
9269 (sched_init): ... here.
9270 (free_global_sched_pressure_data): New, factored out from ...
9271 (sched_finish): ... here.
9272 * sched-int.h (free_global_sched_pressure_data): Declare.
9273 * sched-rgn.c (nr_regions_initial): New static global.
9274 (haifa_find_rgns): Initialize it.
9275 (schedule_region): Disable sched-pressure for the newly
9278 2014-02-17 Richard Biener <rguenther@suse.de>
9280 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9281 release SSA defs of pattern stmts.
9283 2014-02-17 Richard Biener <rguenther@suse.de>
9285 * tree-inline.c (expand_call_inline): Release the virtual
9286 operand defined by the call we are about to inline.
9288 2014-02-17 Richard Biener <rguenther@suse.de>
9290 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9292 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9293 Ilya Tocar <ilya.tocar@intel.com>
9295 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9296 arguments order in builtin.
9297 (_mm512_permutexvar_epi64): Ditto.
9298 (_mm512_mask_permutexvar_epi64): Ditto
9299 (_mm512_maskz_permutexvar_epi32): Ditto
9300 (_mm512_permutexvar_epi32): Ditto
9301 (_mm512_mask_permutexvar_epi32): Ditto
9303 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9305 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9306 (p8_vmrgow): Likewise.
9308 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9310 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9313 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9316 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9317 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9318 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9319 using direct move instructions on ISA 2.07. Also adjust
9320 instruction length for 64-bit.
9321 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9322 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9324 2014-02-15 Alan Modra <amodra@gmail.com>
9328 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9329 find_replacement on parts of insn rtl that might be reloaded.
9331 2014-02-15 Richard Biener <rguenther@suse.de>
9333 PR tree-optimization/60183
9334 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9335 (tree_ssa_phiprop): Calculate and free post-dominators.
9337 2014-02-14 Jeff Law <law@redhat.com>
9339 PR rtl-optimization/60131
9340 * ree.c (get_extended_src_reg): New function.
9341 (combine_reaching_defs): Use it rather than assuming location of REG.
9342 (find_and_remove_re): Verify first operand of extension is
9343 a REG before adding the insns to the copy list.
9345 2014-02-14 Roland McGrath <mcgrathr@google.com>
9347 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9348 * configure: Regenerated.
9349 * config.in: Regenerated.
9350 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9351 instead of ASM_SHORT.
9353 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9354 Richard Earnshaw <rearnsha@arm.com>
9356 PR rtl-optimization/59535
9357 * lra-constraints.c (process_alt_operands): Encourage alternative
9358 when unassigned pseudo class is superset of the alternative class.
9359 (inherit_reload_reg): Don't inherit when optimizing for code size.
9360 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9361 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9362 modes not less than 4 for Thumb1.
9364 2014-02-14 Kyle McMartin <kyle@redhat.com>
9367 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9369 2014-02-14 Richard Biener <rguenther@suse.de>
9371 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9372 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9373 Do not strip INDIRECT_REFs.
9375 2014-02-14 Richard Biener <rguenther@suse.de>
9378 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9379 DECL_FUNCTION_SPECIFIC_TARGET.
9380 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9381 * tree-streamer-out.c (pack_ts_target_option): Remove.
9382 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9383 (write_ts_function_decl_tree_pointers): Do not stream
9384 DECL_FUNCTION_SPECIFIC_TARGET.
9385 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9386 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9387 (lto_input_ts_function_decl_tree_pointers): Do not stream
9388 DECL_FUNCTION_SPECIFIC_TARGET.
9390 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9392 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9393 (get_initial_def_for_induction, vectorizable_induction): Ignore
9394 debug stmts when looking for exit_phi.
9395 (vectorizable_live_operation): Fix up condition.
9397 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9399 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9400 nreverse() because it changes the content of original tree list.
9402 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9404 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9405 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9407 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9409 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9410 GNU coding standards.
9412 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9415 * dwarf2out.c (gen_subprogram_die): Don't call
9416 add_calling_convention_attribute if subr_die is old_die.
9418 2014-02-13 Sharad Singhai <singhai@google.com>
9420 * doc/optinfo.texi: Fix order of nodes.
9422 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9424 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9425 operands[2], not operands[3].
9427 2014-02-13 Richard Biener <rguenther@suse.de>
9430 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9431 mention the possibility of an in-tree build.
9432 (CLooG): Update recommended version to 0.18.1, mention the
9433 possibility of an in-tree build and clarify that the ISL
9434 bundled with CLooG does not work.
9436 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9439 * expr.c (compress_float_constant): If x is a hard register,
9440 extend into a pseudo and then move to x.
9442 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9444 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9445 caused by bad second argument to warning_at() with -mhotpatch and
9446 nested functions (e.g. with gfortran).
9448 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9450 * opts.c (option_name): Remove "enabled by default" rider.
9452 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9454 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9456 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9457 Uros Bizjak <ubizjak@gmail.com>
9460 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9461 * configure: Regenerated.
9463 2014-02-12 Richard Biener <rguenther@suse.de>
9465 * vec.c (vec_prefix::calculate_allocation): Move as
9466 inline variant to vec.h.
9467 (vec_prefix::calculate_allocation_1): New out-of-line version.
9468 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9469 (vec_prefix::m_has_auto_buf): Rename to ...
9470 (vec_prefix::m_using_auto_storage): ... this.
9471 (vec_prefix::calculate_allocation): Inline the easy cases
9472 and dispatch to calculate_allocation_1 which doesn't need the
9474 (va_heap::reserve): Use gcc_checking_assert.
9475 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9476 m_using_auto_storage.
9477 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9479 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9480 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9481 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9483 2014-02-12 Richard Biener <rguenther@suse.de>
9485 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9486 when we found a dependence.
9488 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9490 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9492 (maybe_fold_stmt): ... into this new function.
9493 * omp-low.c (lower_omp): Update comment.
9495 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9498 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9501 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9503 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9504 identifiers in comments.
9505 (cortexa53_extra_costs): Likewise.
9506 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9507 (cortexa7_extra_costs): Likewise.
9508 (cortexa12_extra_costs): Likewise.
9509 (cortexa15_extra_costs): Likewise.
9510 (v7m_extra_costs): Likewise.
9512 2014-02-12 Richard Biener <rguenther@suse.de>
9515 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9516 of posix_memalign being successful.
9517 (lower_stmt): Restrict lowering of posix_memalign to when
9518 -ftree-bit-ccp is enabled.
9520 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9522 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9524 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9526 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9528 PR rtl-optimization/60116
9529 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9530 other_insn once the combination has been validated.
9532 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9535 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9537 * ipa-devirt.c: Include demangle.h
9538 (odr_violation_reported): New static variable.
9539 (add_type_duplicate): Update odr_violations.
9540 (maybe_record_node): Add completep parameter; update it.
9541 (record_target_from_binfo): Add COMPLETEP parameter;
9542 update it as needed.
9543 (possible_polymorphic_call_targets_1): Likewise.
9544 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9545 rename FINAL to COMPLETE.
9546 (record_targets_from_bases): Sanity check we found the binfo;
9547 fix COMPLETEP updating.
9548 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9549 parameter, fix computing of COMPLETEP.
9550 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9551 at LTO time do demangling.
9552 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9553 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9555 (gimple_get_virt_method_for_binfo): Likewise.
9556 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9557 gimple_get_virt_method_for_vtable): Update prototypes.
9559 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9562 * genautomata.c (add_presence_absence): Fix typo with
9563 {final_}presence_list.
9565 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9568 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9569 for VSX/Altivec vectors that land in GPR registers.
9571 2014-02-11 Richard Henderson <rth@redhat.com>
9572 Jakub Jelinek <jakub@redhat.com>
9575 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9576 around drhs if type conversion to lacc->type is not useless.
9578 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9580 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9582 (cortex-a57.cortex-a53): Likewise.
9583 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9585 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9587 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9590 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9592 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9593 add_options_for_arm_vfp3.
9595 2014-02-11 Jeff Law <law@redhat.com>
9598 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9599 object with an undesirable mode.
9601 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9604 * config/i386/sol2-9.h: New file.
9605 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9606 *-*-solaris2.9*): Use it.
9608 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9610 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9611 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9613 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9615 * config/microblaze/microblaze.c: Extend mcpu version format
9617 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9619 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9621 2014-02-10 Richard Henderson <rth@redhat.com>
9624 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9625 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9626 ms-abi vs -mno-accumulate-outgoing-args.
9627 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9628 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9631 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9634 * cfgexpand.c (expand_asm_operands): Attach source location to
9635 ASM_INPUT rtx objects.
9636 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9638 2014-02-10 Nick Clifton <nickc@redhat.com>
9640 * config/mn10300/mn10300.c (popcount): New function.
9641 (mn10300_expand_prologue): Include saved registers in stack usage
9644 2014-02-10 Jeff Law <law@redhat.com>
9647 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9648 when changing the SET_DEST of a prior insn to avoid an input reload.
9650 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9652 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9653 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9654 -mcall-openbsd, or -mcall-linux.
9655 (CC1_ENDIAN_BIG_SPEC): Remove.
9656 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9657 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9658 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9659 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9660 and %cc1_endian_default.
9661 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9663 2014-02-10 Richard Biener <rguenther@suse.de>
9665 PR tree-optimization/60115
9666 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9667 MEM_REF handling. Properly verify that the accesses are not
9668 out of the objects bound.
9670 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9672 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9675 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9677 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9678 proper constants and fix formatting.
9679 (possible_polymorphic_call_targets): Fix formatting.
9681 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9682 Ilya Tocar <ilya.tocar@intel.com>
9684 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9685 (_mm512_loadu_epi32): Renamed into...
9686 (_mm512_loadu_si512): This.
9687 (_mm512_storeu_epi32): Renamed into...
9688 (_mm512_storeu_si512): This.
9689 (_mm512_maskz_ceil_ps): Removed.
9690 (_mm512_maskz_ceil_pd): Ditto.
9691 (_mm512_maskz_floor_ps): Ditto.
9692 (_mm512_maskz_floor_pd): Ditto.
9693 (_mm512_floor_round_ps): Ditto.
9694 (_mm512_floor_round_pd): Ditto.
9695 (_mm512_ceil_round_ps): Ditto.
9696 (_mm512_ceil_round_pd): Ditto.
9697 (_mm512_mask_floor_round_ps): Ditto.
9698 (_mm512_mask_floor_round_pd): Ditto.
9699 (_mm512_mask_ceil_round_ps): Ditto.
9700 (_mm512_mask_ceil_round_pd): Ditto.
9701 (_mm512_maskz_floor_round_ps): Ditto.
9702 (_mm512_maskz_floor_round_pd): Ditto.
9703 (_mm512_maskz_ceil_round_ps): Ditto.
9704 (_mm512_maskz_ceil_round_pd): Ditto.
9705 (_mm512_expand_pd): Ditto.
9706 (_mm512_expand_ps): Ditto.
9707 * config/i386/i386.c (ix86_builtins): Remove
9708 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9709 (bdesc_args): Ditto.
9710 * config/i386/predicates.md (const1256_operand): New.
9711 (const_1_to_2_operand): Ditto.
9712 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9713 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9714 (*avx512pf_gatherpf<mode>sf): Ditto.
9715 (avx512pf_gatherpf<mode>df): Ditto.
9716 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9717 (*avx512pf_gatherpf<mode>df): Ditto.
9718 (avx512pf_scatterpf<mode>sf): Ditto.
9719 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9720 (*avx512pf_scatterpf<mode>sf): Ditto.
9721 (avx512pf_scatterpf<mode>df): Ditto.
9722 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9723 (*avx512pf_scatterpf<mode>df): Ditto.
9724 (avx512f_expand<mode>): Removed.
9725 (<shift_insn><mode>3<mask_name>): Change predicate type.
9727 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9729 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9730 not at the end of datarefs vector use ordered_remove to avoid
9731 reordering datarefs vector.
9734 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9735 mark local addressable non-static vars as GOVD_PRIVATE
9736 instead of GOVD_LOCAL.
9737 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9738 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9742 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9743 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9744 assume_aligned or alloc_align attributes.
9745 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9746 arguments. Handle also assume_aligned and alloc_align attributes.
9747 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9748 calls to functions with assume_aligned or alloc_align attributes.
9749 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9751 2014-02-08 Terry Guo <terry.guo@arm.com>
9753 * doc/invoke.texi: Document ARM -march=armv7e-m.
9755 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9757 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9758 flag on __cilkrts_rethrow builtin.
9761 * ipa-cp.c (determine_versionability): Fail at -O0
9762 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9763 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9766 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9769 * tree-inline.c (copy_forbidden): Fail for
9770 __attribute__((optimize (0))) functions.
9772 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9774 * varpool.c: Include pointer-set.h.
9775 (varpool_remove_unreferenced_decls): Variables in other partitions
9776 will not be output; be however careful to not lose information
9779 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9781 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9782 lookup in the vtable constructor.
9784 2014-02-07 Jeff Law <law@redhat.com>
9787 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9788 define_insn_and_split.
9790 * ipa-inline.c (inline_small_functions): Fix typos.
9792 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9794 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9795 (s390_can_use_return_insn): Declare.
9796 * config/s390/s390.h (EPILOGUE_USES): Define.
9797 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9799 (s390_chunkify_start): Handle return JUMP_LABELs.
9800 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9801 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9802 (s390_can_use_return_insn): New functions.
9803 (s390_fix_long_loop_prediction): Handle conditional returns.
9804 (TARGET_SET_UP_BY_PROLOGUE): Define.
9805 * config/s390/s390.md (ANY_RETURN): New code iterator.
9806 (*creturn, *csimple_return, return, simple_return): New patterns.
9808 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9810 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9811 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9812 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9813 REG_CFA_RESTORE list when deciding not to restore a register.
9815 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9817 * config/s390/s390.c: Include tree-pass.h and context.h.
9818 (s390_early_mach): New function, split out from...
9819 (s390_emit_prologue): ...here.
9820 (pass_data_s390_early_mach): New pass structure.
9821 (pass_s390_early_mach): New class.
9822 (s390_option_override): Create and register early_mach pass.
9823 Move to end of file.
9825 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9827 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9828 to match for the exit block.
9830 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9832 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9833 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9834 Reject misaligned operands.
9836 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9838 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9840 2014-02-07 Richard Biener <rguenther@suse.de>
9843 * gimple-low.c (lower_builtin_posix_memalign): New function.
9844 (lower_stmt): Call it to lower posix_memalign in a way
9845 to make alignment info accessible.
9847 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9850 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9851 __builtin_setjmp_receiver.
9853 2014-02-07 Richard Biener <rguenther@suse.de>
9856 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9857 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9858 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9859 Handle BUILT_IN_POSIX_MEMALIGN.
9860 (find_func_clobbers): Likewise.
9861 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9862 (call_may_clobber_ref_p_1): Likewise.
9864 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9867 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9870 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9873 * lto-cgraph.c (lto_output_node): Use
9874 symtab_get_symbol_partitioning_class.
9875 (lto_output_varpool_node): likewise.
9876 (symtab_get_symbol_partitioning_class): Move here from
9878 * cgraph.h (symbol_partitioning_class): Likewise.
9879 (symtab_get_symbol_partitioning_class): Declare.
9881 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9883 * ggc.h (ggc_internal_cleared_alloc): New macro.
9884 * vec.h (vec_safe_copy): Handle memory stats.
9885 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9886 * target-globals.c (save_target_globals): Likewise.
9888 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9891 * expr.c (emit_move_resolve_push): Export; be bit more selective
9892 on when to clear alias set.
9893 * expr.h (emit_move_resolve_push): Declare.
9894 * function.h (struct function): Add tail_call_marked.
9895 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9896 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9897 * config/i386/i386.md (TImode move expander): De not call
9899 (FP push expanders): Preserve memory attributes.
9900 * config/i386/sse.md (push<mode>1): Remove.
9901 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9902 (ix86_expand_push): Remove.
9903 * config/i386/mmx.md (push<mode>1): Remove.
9905 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9907 PR rtl-optimization/60030
9908 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9909 lopart with paradoxical subreg before shifting it up by hprec.
9911 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9913 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9914 Remove extra newline at end of file.
9915 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9916 (arm_issue_rate): Handle cortexa57.
9917 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9918 (cortex-a57.cortex-a53): Likewise.
9920 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9923 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9924 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9926 (arm_expand_prologue): Adjust all callers.
9927 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9928 info, registers also at the lowest numbered registers side. Use
9929 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9933 * var-tracking.c (adjust_mems): Before adding a SET to
9934 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9936 2014-02-06 Alan Modra <amodra@gmail.com>
9939 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9940 change SDmode to DDmode when lra_in_progress.
9942 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9945 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9946 free_data_ref on the dr first, and before goto again also set dr
9947 to the next dr. For simd_lane_access, free old datarefs[i] before
9948 overwriting it. For get_vectype_for_scalar_type failure, don't
9949 free_data_ref if simd_lane_access.
9951 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9954 * tree.h (opts_for_fn): New inline function.
9955 (opt_for_fn): Define.
9956 * config/i386/i386.c (ix86_function_regparm): Use
9957 opt_for_fn (decl, optimize) instead of optimize.
9959 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9961 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9962 for SYMBOL_REF in large memory model.
9964 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9966 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9968 (cortex-a57): Likewise.
9969 (cortex-a57.cortex-a53): Likewise.
9971 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9972 Kugan Vivekanandarajah <kuganv@linaro.org>
9974 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9976 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9978 2014-02-06 Richard Biener <rguenther@suse.de>
9980 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9981 set_loop_copy and initialize_original_copy_tables.
9983 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9985 * config/aarch64/aarch64-simd.md
9986 (aarch64_ashr_simddi): Change QI to SI.
9988 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9989 Jakub Jelinek <jakub@redhat.com>
9992 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9995 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9997 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9998 CODE_FOR_altivec_vpku[hw]um to
9999 CODE_FOR_altivec_vpku[hw]um_direct.
10000 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10001 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10002 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10003 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10005 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10007 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10008 generation for -maltivec=be.
10009 (altivec_vsumsws): Simplify redundant test.
10011 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10013 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10014 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10015 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10016 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10017 gen_altivec_vpkuwum.
10018 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10020 (altivec_vpks<VI_char>ss): Likewise.
10021 (altivec_vpks<VI_char>us): Likewise.
10022 (altivec_vpku<VI_char>us): Likewise.
10023 (altivec_vpku<VI_char>um): Likewise.
10024 (altivec_vpku<VI_char>um_direct): New (copy of
10025 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10027 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10028 target is little endian and -maltivec=be is not specified.
10029 (*altivec_vupkhs<VU_char>_direct): New (copy of
10030 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10031 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10032 target is little endian and -maltivec=be is not specified.
10033 (*altivec_vupkls<VU_char>_direct): New (copy of
10034 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10035 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10036 little endian and -maltivec=be is not specified.
10037 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10038 little endian and -maltivec=be is not specified.
10040 2014-02-05 Richard Henderson <rth@redhat.com>
10043 * combine-stack-adj.c: Revert r206943.
10044 * sched-int.h (struct deps_desc): Add last_args_size.
10045 * sched-deps.c (init_deps): Initialize it.
10046 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10047 contain REG_ARGS_SIZE notes.
10049 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10051 * lto-cgraph.c (asm_nodes_output): Make global.
10052 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10053 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10054 (driver_handle_option): Handle OPT_fwpa.
10056 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10059 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10060 a comment typo and formatting issue. If odr_hash hasn't been
10061 created, return vNULL and set *completep to false.
10063 PR middle-end/57499
10064 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10065 bb with no successors.
10067 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10070 * doc/invoke.texi (-march): Clarify documentation for ARM.
10071 (-mtune): Likewise.
10074 2014-02-05 Richard Biener <rguenther@suse.de>
10076 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10077 when not vectorizing because of too many alias checks.
10078 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10079 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10081 2014-02-05 Nick Clifton <nickc@redhat.com>
10083 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10084 accept extended registers in any mode when compiling for the MN10300.
10086 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10088 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10089 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10090 sanitization attributes.
10091 (can_inline_edge_p): Likewise.
10092 (sanitize_attrs_match_for_inline_p): New function.
10094 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10096 * ipa-prop.c (detect_type_change): Shor circuit testing of
10097 type changes on THIS pointer.
10099 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10102 * config/pa/pa.c (legitimize_tls_address): Return original address
10103 if not passed a SYMBOL_REF rtx.
10104 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10106 (pa_emit_move_sequence): Simplify TLS source operands.
10107 (pa_legitimate_constant_p): Reject all TLS constants.
10108 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10109 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10111 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10113 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10114 groups when we know they are controlled by LTO.
10115 * varasm.c (default_binds_local_p_1): If object is in other partition,
10116 it will be resolved locally.
10118 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10120 * config/host-linux.c (linux_gt_pch_use_address): Don't
10121 use SSIZE_MAX because it is not always defined.
10123 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10126 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10127 threshold for pseudo splitting.
10128 (update_ebb_live_info): Process call argument hard registers and
10129 hard registers from insn definition too.
10130 (max_small_class_regs_num): New constant.
10131 (inherit_in_ebb): Update live hard regs through EBBs. Update
10132 reloads_num only for small register classes. Don't split for
10135 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10138 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10141 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10143 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10144 visibility is safe.
10146 2014-02-04 Marek Polacek <polacek@redhat.com>
10148 * gdbinit.in (pel): Define.
10150 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10152 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10155 2014-02-04 Richard Biener <rguenther@suse.de>
10158 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10159 in function context local.
10160 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10161 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10162 similar to LTO_imported_decl_ref.
10164 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10166 PR tree-optimization/60002
10167 * cgraphclones.c (build_function_decl_skip_args): Clear
10168 DECL_LANG_SPECIFIC.
10170 PR tree-optimization/60023
10171 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10172 false to gsi_replace.
10173 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10174 has been in some EH region and vec_stmt could throw, add
10175 vec_stmt into the same EH region.
10176 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10177 has no lhs, ignore it.
10178 * internal-fn.c (expand_MASK_LOAD): Likewise.
10181 * tree-inline.c (copy_forbidden): Fail for
10182 __attribute__((optimize (0))) functions.
10185 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10186 is set, copy one less argument.
10187 (expand_simd_clones): Don't subtract clone_info->inbranch
10188 from simd_clone_struct_alloc argument.
10190 PR rtl-optimization/57915
10191 * recog.c (simplify_while_replacing): If all unary/binary/relational
10192 operation arguments are constant, attempt to simplify those.
10194 PR middle-end/59261
10195 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10196 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10198 2014-02-04 Richard Biener <rguenther@suse.de>
10200 PR tree-optimization/60012
10201 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10202 TBAA disambiguation to all DDRs.
10204 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10207 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10208 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10210 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10213 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10215 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10217 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10218 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10220 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10223 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10224 to figure out targets of polymorphic calls with known decl.
10225 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10226 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10227 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10228 (get_polymorphic_call_info): ... here.
10229 (get_polymorphic_call_info_from_invariant): New function.
10231 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10233 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10234 lookup via vtable pointer; check for type consistency
10235 and turn inconsitent facts into UNREACHABLE.
10236 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10237 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10238 type inconsistent querries; return UNREACHABLE instead.
10240 2014-02-03 Richard Henderson <rth@twiddle.net>
10243 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10244 already processed this node.
10245 (normalize_one_pred_1): Pass along mark_set.
10246 (normalize_one_pred): Create and destroy a pointer_set_t.
10247 (normalize_one_pred_chain): Likewise.
10249 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10251 PR gcov-profile/58602
10252 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10254 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10257 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10258 -fno-devirtualize; try to devirtualize by the knowledge of
10259 virtual table pointer given by aggregate propagation.
10260 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10261 (ipa_print_node_jump_functions): Dump also offset that
10262 is relevant for polymorphic calls.
10263 (determine_known_aggregate_parts): Add arg_type parameter; use it
10264 instead of determining the type from pointer type.
10265 (ipa_compute_jump_functions_for_edge): Update call of
10266 determine_known_aggregate_parts.
10267 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10268 (gimple_get_virt_method_for_binfo): ... here; simplify using
10269 vtable_pointer_value_to_vtable.
10270 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10271 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10272 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10273 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10275 (vtable_pointer_value_to_binfo): ... here.
10276 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10278 2014-02-03 Teresa Johnson <tejohnson@google.com>
10280 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10281 redef of outer loop index variable.
10283 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10287 * doc/extend.texi (Function Attributes): Typo.
10289 2014-02-03 Cong Hou <congh@google.com>
10291 PR tree-optimization/60000
10292 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10293 if the vectorized statement is a store. A store statement can only
10294 appear at the end of pattern statements.
10296 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10298 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10299 (ix86_option_override_internal): Default long double to 64-bit for
10300 32-bit Bionic and to 128-bit for 64-bit Bionic.
10302 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10303 TARGET_LONG_DOUBLE_128 is true.
10304 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10306 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10307 (mlong-double-64): Negate -mlong-double-128.
10308 (mlong-double-128): New option.
10310 * config/i386/i386-c.c (ix86_target_macros): Define
10311 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10313 * doc/invoke.texi: Document -mlong-double-128.
10315 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10317 PR rtl-optimization/60024
10318 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10320 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10322 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10324 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10326 PR rtl-optimization/57662
10327 * sel-sched.c (code_motion_path_driver): Do not mark already not
10328 existing blocks in the visiting bitmap.
10330 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10332 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10333 on the insn being emitted.
10335 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10336 Will Deacon <will.deacon@arm.com>
10338 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10340 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10342 * config/arm/arm-tables.opt: Regenerate.
10344 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10346 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10347 for vector types other than V16QImode.
10348 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10349 define_expand, and call altivec_expand_vec_perm_le when producing
10350 code with little endian element order.
10351 (*altivec_vperm_<mode>_internal): New insn having previous
10352 behavior of altivec_vperm_<mode>.
10353 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10354 altivec_expand_vec_perm_le when producing code with little endian
10356 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10357 behavior of altivec_vperm_<mode>_uns.
10359 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10361 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10362 (altivec_vsumsws): Add handling for -maltivec=be with a little
10364 (altivec_vsumsws_direct): New.
10365 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10366 gen_altivec_vsumsws.
10368 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10370 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10371 vtable_pointer_value_to_binfo): New functions.
10372 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10373 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10375 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10377 * config/nios2/nios2.md (load_got_register): Initialize GOT
10378 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10379 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10381 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10383 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10384 preserverd by passthrough, do not propagate the type.
10386 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10388 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10389 (mips_atomic_assign_expand_fenv): New function.
10390 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10392 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10394 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10395 (__builtin_mips_set_fcsr): Likewise.
10396 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10397 MIPS_USI_FTYPE_VOID.
10398 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10399 (mips16_expand_set_fcsr): Likewise.
10400 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10401 (mips16_set_fcsr_stub): Likewise.
10402 (mips16_get_fcsr_one_only_stub): New class.
10403 (mips16_set_fcsr_one_only_stub): Likewise.
10404 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10405 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10406 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10407 (hard_float): New availability predicate.
10408 (mips_builtins): Add get_fcsr and set_fcsr.
10409 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10410 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10411 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10412 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10413 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10416 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10418 * config/mips/mips.c (mips_one_only_stub): New class.
10419 (mips_need_mips16_rdhwr_p): Replace with...
10420 (mips16_rdhwr_stub): ...this new variable.
10421 (mips16_stub_call_address): New function.
10422 (mips16_rdhwr_one_only_stub): New class.
10423 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10424 (mips_output_mips16_rdhwr): Delete.
10425 (mips_finish_stub): New function.
10426 (mips_code_end): Use it to handle rdhwr stubs.
10428 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10431 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10432 when calculating size of integer atomic types.
10434 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10436 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10438 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10440 PR tree-optimization/60003
10441 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10442 * profile.c (branch_prob): Use gimple_call_builtin_p
10443 to check for BUILT_IN_SETJMP_RECEIVER.
10444 * tree-inline.c (copy_bb): Call notice_special_calls.
10446 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10449 * lra-constraints.c (process_alt_operands): Update reload_sum only
10452 2014-01-31 Richard Henderson <rth@redhat.com>
10454 PR middle-end/60004
10455 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10456 until after else_eh is processed.
10458 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10460 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10461 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10462 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10463 in smmintrin.h, remove them.
10464 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10465 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10466 * config/i386/i386.md (ROUND_SAE): Fix value.
10467 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10468 (const48_operand): New.
10469 * config/i386/subst.md (round), (round_expand): Use
10470 const_4_or_8_to_11_operand.
10471 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10473 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10475 * config/i386/constraints.md (Yk): Swap meaning with k.
10476 * config/i386/i386.md (movhi_internal): Change Yk to k.
10477 (movqi_internal): Ditto.
10478 (*k<logic><mode>): Ditto.
10481 (kandn<mode>): Ditto.
10482 (*<code>hi_1): Ditto.
10483 (*<code>qi_1): Ditto.
10484 (kxnor<mode>): Ditto.
10485 (kortestzhi): Ditto.
10486 (kortestchi): Ditto.
10488 (*one_cmplhi2_1): Ditto.
10489 (*one_cmplqi2_1): Ditto.
10490 * config/i386/sse.md (): Change k to Yk.
10491 (avx512f_load<mode>_mask): Ditto.
10492 (avx512f_blendm<mode>): Ditto.
10493 (avx512f_store<mode>_mask): Ditto.
10494 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10495 (avx512f_storedqu<mode>_mask): Ditto.
10496 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10498 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10499 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10500 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10501 (avx512f_maskcmp<mode>3): Ditto.
10502 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10503 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10504 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10505 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10506 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10507 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10508 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10509 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10510 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10511 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10512 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10513 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10514 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10515 (vec_extract_lo_<mode>_maskm): Ditto.
10516 (vec_extract_hi_<mode>_maskm): Ditto.
10517 (avx512f_vternlog<mode>_mask): Ditto.
10518 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10519 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10520 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10521 (avx512f_<code>v8div16qi2_mask): Ditto.
10522 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10523 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10524 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10525 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10526 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10527 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10528 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10529 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10530 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10531 (avx512cd_maskb_vec_dupv8di): Ditto.
10532 (avx512cd_maskw_vec_dupv16si): Ditto.
10533 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10534 (avx512f_vpermi2var<mode>3_mask): Ditto.
10535 (avx512f_vpermi2var<mode>3_mask): Ditto.
10536 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10537 (*avx512f_gathersi<mode>): Ditto.
10538 (*avx512f_gathersi<mode>_2): Ditto.
10539 (*avx512f_gatherdi<mode>): Ditto.
10540 (*avx512f_gatherdi<mode>_2): Ditto.
10541 (*avx512f_scattersi<mode>): Ditto.
10542 (*avx512f_scatterdi<mode>): Ditto.
10543 (avx512f_compress<mode>_mask): Ditto.
10544 (avx512f_compressstore<mode>_mask): Ditto.
10545 (avx512f_expand<mode>_mask): Ditto.
10546 * config/i386/subst.md (mask): Change k to Yk.
10547 (mask_scalar_merge): Ditto.
10550 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10552 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10554 2014-01-31 Richard Biener <rguenther@suse.de>
10556 PR middle-end/59990
10557 * builtins.c (fold_builtin_memory_op): Make sure to not
10558 use a floating-point mode or a boolean or enumeral type for
10559 the copy operation.
10561 2014-01-30 DJ Delorie <dj@redhat.com>
10563 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10564 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10565 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10566 whenever main() has an epilogue.
10568 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10570 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10571 unused variable "field".
10572 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10573 (vsx_mergeh_<mode>): Likewise.
10574 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10575 (altivec_vmrghh): Likewise.
10576 (altivec_vmrghw): Likewise.
10577 (altivec_vmrglb): Likewise.
10578 (altivec_vmrglh): Likewise.
10579 (altivec_vmrglw): Likewise.
10580 (altivec_vspltb): Add missing uses.
10581 (altivec_vsplth): Likewise.
10582 (altivec_vspltw): Likewise.
10583 (altivec_vspltsf): Likewise.
10585 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10588 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10589 frame related instructions.
10591 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10593 PR rtl-optimization/59959
10594 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10595 any reload of register whose subreg is invalid.
10597 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10599 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10600 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10601 Add missing return type - void.
10603 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10605 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10606 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10607 remove element index adjustment for endian (now handled in vsx.md
10609 (altivec_expand_vec_perm_const): Use
10610 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10611 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10612 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10613 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10614 define_expand and a new define_insn *altivec_vspltb_internal;
10615 adjust for -maltivec=be on a little endian target.
10616 (altivec_vspltb_direct): New.
10617 (altivec_vsplth): Divide into a define_expand and a new
10618 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10619 little endian target.
10620 (altivec_vsplth_direct): New.
10621 (altivec_vspltw): Divide into a define_expand and a new
10622 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10623 little endian target.
10624 (altivec_vspltw_direct): New.
10625 (altivec_vspltsf): Divide into a define_expand and a new
10626 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10627 a little endian target.
10629 2014-01-30 Richard Biener <rguenther@suse.de>
10631 PR tree-optimization/59993
10632 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10633 can propagate form the earlier stmt and avoid the transform
10634 when the intermediate result is needed.
10636 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10638 * README.Portability: Fix typo.
10640 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10642 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10643 comparison_operator with ordered_comparison_operator.
10645 2014-01-30 Nick Clifton <nickc@redhat.com>
10647 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10648 Rename to mn10300_store_multiple_regs.
10649 * config/mn10300/mn10300.c: Likewise.
10650 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10651 store_multiple_regs.
10652 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10653 Call mn10300_store_multiple_regs.
10655 2014-01-30 Nick Clifton <nickc@redhat.com>
10656 DJ Delorie <dj@redhat.com>
10658 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10659 %fp 2 to keep registers after it properly word-aligned.
10660 (rl78_alloc_physical_registers_umul): Handle the case where both
10661 input operands are the same.
10663 2014-01-30 Richard Biener <rguenther@suse.de>
10665 PR tree-optimization/59903
10666 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10669 2014-01-30 Jason Merrill <jason@redhat.com>
10672 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10675 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10677 2014-01-30 Richard Biener <rguenther@suse.de>
10679 PR tree-optimization/59951
10680 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10682 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10685 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10686 SFmode to DFmode case.
10688 2014-01-29 DJ Delorie <dj@redhat.com>
10690 * config/msp430/msp430.opt (-minrt): New.
10691 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10693 (ENDFILE_SPEC): Likewise.
10695 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10697 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10698 (estimate_function_body_sizes): Use it.
10700 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10703 * dwarf2out.c (is_cxx_auto): New.
10704 (is_base_type): Use it.
10705 (gen_type_die_with_usage): Likewise.
10707 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10709 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10710 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10711 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10712 -maltivec=be with LE targets.
10713 (vsx_mergeh_<mode>): Likewise.
10714 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10715 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10716 (altivec_vmrghb): Replace with define_expand and new
10717 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10718 (altivec_vmrghb_direct): New define_insn.
10719 (altivec_vmrghh): Replace with define_expand and new
10720 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10721 (altivec_vmrghh_direct): New define_insn.
10722 (altivec_vmrghw): Replace with define_expand and new
10723 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10724 (altivec_vmrghw_direct): New define_insn.
10725 (*altivec_vmrghsf): Adjust for endianness.
10726 (altivec_vmrglb): Replace with define_expand and new
10727 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10728 (altivec_vmrglb_direct): New define_insn.
10729 (altivec_vmrglh): Replace with define_expand and new
10730 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10731 (altivec_vmrglh_direct): New define_insn.
10732 (altivec_vmrglw): Replace with define_expand and new
10733 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10734 (altivec_vmrglw_direct): New define_insn.
10735 (*altivec_vmrglsf): Adjust for endianness.
10736 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10737 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10738 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10739 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10740 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10741 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10742 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10743 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10745 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10747 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10748 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10751 2014-01-29 Richard Biener <rguenther@suse.de>
10753 PR tree-optimization/58742
10754 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10755 associate_pointerplus_align.
10756 (associate_pointerplus_diff): New function.
10757 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10758 and associate_pointerplus_diff.
10760 2014-01-29 Richard Biener <rguenther@suse.de>
10762 * lto-streamer.h (LTO_major_version): Bump to 3.
10763 (LTO_minor_version): Reset to 0.
10765 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10767 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10768 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10769 (arm_file_start): Generate correct asm header for armv7ve.
10770 * config/arm/bpabi.h: Add multilib support for armv7ve.
10771 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10772 and cortex-a15 to armv7ve.
10773 * config/arm/t-aprofile: Add multilib support for armv7ve.
10774 * doc/invoke.texi: Document -march=armv7ve.
10776 2014-01-29 Richard Biener <rguenther@suse.de>
10778 PR tree-optimization/58742
10779 * tree-ssa-forwprop.c (associate_plusminus): Return true
10780 if we changed sth, defer EH cleanup to ...
10781 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10782 (simplify_mult): New function.
10784 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10786 PR middle-end/59917
10787 PR tree-optimization/59920
10788 * tree.c (build_common_builtin_nodes): Remove
10789 __builtin_setjmp_dispatcher initialization.
10790 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10791 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10792 instead of gsi_after_labels + manually skipping debug stmts.
10793 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10794 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10795 * tree-inline.c (copy_edges_for_bb): Remove
10796 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10797 argument. Ignore computed_goto_p stmts. Don't call
10798 make_abnormal_goto_edges. If a call might need abnormal edges
10799 for non-local gotos, see if it already has an edge to
10800 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10801 with true argument, don't do anything then, otherwise add
10802 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10803 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10805 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10806 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10807 (lower_stmt): Don't set data->calls_builtin_setjmp.
10808 (lower_builtin_setjmp): Adjust comment.
10809 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10810 * tree-cfg.c (found_computed_goto): Remove.
10811 (factor_computed_gotos): Remove.
10812 (make_goto_expr_edges): Return bool, true for computed gotos.
10813 Don't call make_abnormal_goto_edges.
10814 (build_gimple_cfg): Don't set found_computed_goto, don't call
10815 factor_computed_gotos.
10816 (computed_goto_p): No longer static.
10817 (make_blocks): Don't set found_computed_goto.
10818 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10819 (make_edges): If make_goto_expr_edges returns true, push bb
10820 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10821 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10822 vector. Record mapping between bbs and OpenMP regions if there
10823 are any, adjust make_gimple_omp_edges caller. Call
10824 handle_abnormal_edges.
10825 (make_abnormal_goto_edges): Remove.
10826 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10827 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10828 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10829 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10830 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10831 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10832 filling *region also set *region_idx to (*region)->entry->index.
10835 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10836 For REGs set ORIGINAL_REGNO.
10838 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10840 * doc/md.texi: Mention that a target shouldn't implement
10841 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10844 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10846 PR tree-optimization/59594
10847 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10848 a copy of the datarefs vector rather than the vector itself.
10850 2014-01-28 Jason Merrill <jason@redhat.com>
10853 * dwarf2out.c (auto_die): New static.
10854 (gen_type_die_with_usage): Handle C++1y 'auto'.
10855 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10858 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10861 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10862 (SPEC_X32): Likewise.
10863 (SPEC_64): Likewise.
10864 * config/i386/i386.c (ix86_option_override_internal): Turn off
10865 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10867 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10868 * config/i386/i386.h (TARGET_16BIT): New macro.
10869 (TARGET_16BIT_P): Likewise.
10870 * config/i386/i386.opt: Add m16.
10871 * doc/invoke.texi: Document -m16.
10873 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10875 PR preprocessor/59935
10876 * input.c (location_get_source_line): Bail out on when line number
10877 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10879 2014-01-28 Richard Biener <rguenther@suse.de>
10881 PR tree-optimization/58742
10882 * tree-ssa-forwprop.c (associate_plusminus): Handle
10883 pointer subtraction of the form (T)(P + A) - (T)P.
10885 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10887 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10890 2014-01-28 Richard Biener <rguenther@suse.de>
10893 2014-01-28 Richard Biener <rguenther@suse.de>
10895 PR rtl-optimization/45364
10896 PR rtl-optimization/59890
10897 * var-tracking.c (local_get_addr_clear_given_value): Handle
10898 already cleared slot.
10899 (val_reset): Handle not allocated local_get_addr_cache.
10900 (vt_find_locations): Use post-order on the inverted CFG.
10902 2014-01-28 Richard Biener <rguenther@suse.de>
10904 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10906 2014-01-28 Richard Biener <rguenther@suse.de>
10908 PR rtl-optimization/45364
10909 PR rtl-optimization/59890
10910 * var-tracking.c (local_get_addr_clear_given_value): Handle
10911 already cleared slot.
10912 (val_reset): Handle not allocated local_get_addr_cache.
10913 (vt_find_locations): Use post-order on the inverted CFG.
10915 2014-01-28 Alan Modra <amodra@gmail.com>
10917 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10918 * configure.ac <recursive call for build != host>: Define
10919 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10920 and LD_FOR_BUILD too.
10921 * configure: Regenerate.
10923 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10925 * config/i386/i386.c (get_builtin_code_for_version): Separate
10926 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10927 Broadwell from Haswell.
10929 2014-01-27 Steve Ellcey <sellcey@mips.com>
10931 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10932 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10933 * config/mips/mips.c (mips_option_override): Change setting
10935 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10936 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10937 Change from Mask to Var.
10939 2014-01-27 Jeff Law <law@redhat.com>
10941 * ipa-inline.c (inline_small_functions): Fix typo.
10943 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10945 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10946 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10947 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10948 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10949 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10950 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10951 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10952 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10953 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10954 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10955 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10956 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10957 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10958 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10959 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10960 (_mm512_storeu_epi64): Ditto.
10961 (_mm512_cmpge_epi32_mask): Ditto.
10962 (_mm512_cmpge_epu32_mask): Ditto.
10963 (_mm512_cmpge_epi64_mask): Ditto.
10964 (_mm512_cmpge_epu64_mask): Ditto.
10965 (_mm512_cmple_epi32_mask): Ditto.
10966 (_mm512_cmple_epu32_mask): Ditto.
10967 (_mm512_cmple_epi64_mask): Ditto.
10968 (_mm512_cmple_epu64_mask): Ditto.
10969 (_mm512_cmplt_epi32_mask): Ditto.
10970 (_mm512_cmplt_epu32_mask): Ditto.
10971 (_mm512_cmplt_epi64_mask): Ditto.
10972 (_mm512_cmplt_epu64_mask): Ditto.
10973 (_mm512_cmpneq_epi32_mask): Ditto.
10974 (_mm512_cmpneq_epu32_mask): Ditto.
10975 (_mm512_cmpneq_epi64_mask): Ditto.
10976 (_mm512_cmpneq_epu64_mask): Ditto.
10977 (_mm512_expand_pd): Ditto.
10978 (_mm512_expand_ps): Ditto.
10979 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10980 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10981 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10982 * config/i386/i386.c (ix86_builtins): Add
10983 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10984 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10985 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10986 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10987 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10988 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10989 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10990 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10991 IX86_BUILTIN_PMOVUSQW512_MEM.
10992 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10993 __builtin_ia32_pmovsqd512mem_mask,
10994 __builtin_ia32_pmovqd512mem_mask,
10995 __builtin_ia32_pmovusqw512mem_mask,
10996 __builtin_ia32_pmovsqw512mem_mask,
10997 __builtin_ia32_pmovqw512mem_mask,
10998 __builtin_ia32_pmovusdw512mem_mask,
10999 __builtin_ia32_pmovsdw512mem_mask,
11000 __builtin_ia32_pmovdw512mem_mask,
11001 __builtin_ia32_pmovqb512mem_mask,
11002 __builtin_ia32_pmovusqb512mem_mask,
11003 __builtin_ia32_pmovsqb512mem_mask,
11004 __builtin_ia32_pmovusdb512mem_mask,
11005 __builtin_ia32_pmovsdb512mem_mask,
11006 __builtin_ia32_pmovdb512mem_mask.
11007 (bdesc_args): Add __builtin_ia32_expanddf512,
11008 __builtin_ia32_expandsf512.
11009 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11010 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11011 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11012 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11013 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11014 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11015 (avx512f_<code>v8div16qi2_mask_store): This.
11016 (avx512f_expand<mode>): New.
11018 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11020 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11022 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11023 (_mm512_prefetch_i32scatter_pd): Ditto.
11024 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11025 (_mm512_prefetch_i64scatter_pd): Ditto.
11026 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11027 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11028 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11029 (_mm512_prefetch_i32scatter_ps): Ditto.
11030 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11031 (_mm512_prefetch_i64scatter_ps): Ditto.
11032 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11033 * config/i386/i386-builtin-types.def: Define
11034 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11035 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11036 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11037 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11038 IX86_BUILTIN_SCATTERPFQPD.
11039 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11040 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11041 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11042 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11043 __builtin_ia32_scatterpfqps.
11044 (ix86_expand_builtin): Expand new built-ins.
11045 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11046 fix memory access data type.
11047 (*avx512pf_gatherpf<mode>_mask): Ditto.
11048 (*avx512pf_gatherpf<mode>): Ditto.
11049 (avx512pf_scatterpf<mode>): Ditto.
11050 (*avx512pf_scatterpf<mode>_mask): Ditto.
11051 (*avx512pf_scatterpf<mode>): Ditto.
11052 (GATHER_SCATTER_SF_MEM_MODE): New.
11053 (avx512pf_gatherpf<mode>df): Ditto.
11054 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11055 (*avx512pf_scatterpf<mode>df): Ditto.
11057 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11060 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11061 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11064 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11066 * common/config/arm/arm-common.c
11067 (arm_rewrite_mcpu): Handle multiple names.
11069 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11071 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11073 * gimple-builder.h (create_gimple_tmp): Delete.
11075 2014-01-27 Christian Bruel <christian.bruel@st.com>
11077 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11080 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11082 * config/pa/pa.md (call): Generate indirect long calls to non-local
11083 functions when outputing 32-bit code.
11084 (call_value): Likewise except for special call to buggy powf function.
11086 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11087 portable runtime and PIC indirect calls.
11088 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11089 and PIC call sequences. Use ldo instead of blr to set return register
11090 in PIC call sequence.
11092 2014-01-25 Walter Lee <walt@tilera.com>
11094 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11095 avoid clobbering a live register.
11097 2014-01-25 Walter Lee <walt@tilera.com>
11099 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11100 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11101 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11102 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11104 2014-01-25 Walter Lee <walt@tilera.com>
11106 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11107 arguments on even registers.
11108 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11110 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11111 (BIGGEST_ALIGNMENT): Ditto.
11112 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11114 2014-01-25 Walter Lee <walt@tilera.com>
11116 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11117 insns before bundling.
11118 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11120 2014-01-25 Walter Lee <walt@tilera.com>
11122 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11123 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11124 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11126 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11128 * config/mips/constraints.md (kl): Delete.
11129 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11130 define expands, using...
11131 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11132 instructions for MIPS16.
11133 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11134 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11136 2014-01-25 Walter Lee <walt@tilera.com>
11138 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11142 2014-01-25 Walter Lee <walt@tilera.com>
11144 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11145 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11147 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11149 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11152 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11154 * print-rtl.c (in_call_function_usage): New var.
11155 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11156 EXPR_LIST mode as mode and not as reg note name.
11158 PR middle-end/59561
11159 * cfgloopmanip.c (copy_loop_info): If
11160 loop->warned_aggressive_loop_optimizations, make sure
11161 the flag is set in target loop too.
11163 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11165 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11167 * builtins.def: Likewise.
11168 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11169 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11170 * ira.c (ira_setup_eliminable_regset): Likewise.
11171 * omp-low.c (gate_expand_omp): Likewise.
11172 (execute_lower_omp): Likewise.
11173 (diagnose_sb_0): Likewise.
11174 (gate_diagnose_omp_blocks): Likewise.
11175 (simd_clone_clauses_extract): Likewise.
11178 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11180 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11181 correction for little endian...
11182 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11185 2014-01-24 Jeff Law <law@redhat.com>
11187 PR tree-optimization/59919
11188 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11189 for non-returning calls.
11191 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11193 * common/config/aarch64/aarch64-common.c
11194 (aarch64_rewrite_mcpu): Handle multiple names.
11195 * config/aarch64/aarch64.h
11196 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11198 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11200 * input.c (add_file_to_cache_tab): Handle the case where fopen
11203 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11206 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11207 from push operand if code of push isn't PRE_DEC.
11209 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11212 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11213 -mquad-memory-atomic. Update -mquad-memory documentation to say
11214 it is only used for non-atomic loads/stores.
11216 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11217 -mquad-memory or -mquad-memory-atomic switches.
11219 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11220 -mquad-memory-atomic to ISA 2.07 support.
11222 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11223 to separate support of normal quad word memory operations (ldq, stq)
11224 from the atomic quad word memory operations.
11226 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11227 support to separate non-atomic quad word operations from atomic
11228 quad word operations. Disable non-atomic quad word operations in
11229 little endian mode so that we don't have to swap words after the
11230 load and before the store.
11231 (quad_load_store_p): Add comment about atomic quad word support.
11232 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11233 options printed with -mdebug=reg.
11235 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11236 -mquad-memory-atomic as the test for whether we have quad word
11237 atomic instructions.
11238 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11239 or -mp8-vector are used, allow byte/half-word atomic operations.
11241 * config/rs6000/sync.md (load_lockedti): Insure that the address
11242 is a proper indexed or indirect address for the lqarx instruction.
11243 On little endian systems, swap the hi/lo registers after the lqarx
11245 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11246 insure the address is valid for the lqarx instruction.
11247 (store_conditionalti): Insure that the address is a proper indexed
11248 or indirect address for the stqcrx. instruction. On little endian
11249 systems, swap the hi/lo registers before doing the stqcrx.
11251 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11252 insure the address is valid for the stqcrx. instruction.
11254 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11255 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11256 type of quad memory support is available.
11258 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11260 PR regression/59915
11261 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11262 there is a danger of looping.
11264 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11266 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11267 force flag_ira_loop_pressure if set via command line.
11269 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11271 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11272 (ashr_simd): New builtin handling DI mode.
11273 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11274 (aarch64_sshr_simddi): New match pattern.
11275 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11276 (vshrd_n_s64): Likewise.
11277 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11279 2014-01-23 Nick Clifton <nickc@redhat.com>
11281 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11282 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11283 favour of mcu specific scripts.
11284 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11287 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11288 Alex Velenko <Alex.Velenko@arm.com>
11290 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11291 (vaddv_s16): Likewise.
11292 (vaddv_s32): Likewise.
11293 (vaddv_u8): Likewise.
11294 (vaddv_u16): Likewise.
11295 (vaddv_u32): Likewise.
11296 (vaddvq_s8): Likewise.
11297 (vaddvq_s16): Likewise.
11298 (vaddvq_s32): Likewise.
11299 (vaddvq_s64): Likewise.
11300 (vaddvq_u8): Likewise.
11301 (vaddvq_u16): Likewise.
11302 (vaddvq_u32): Likewise.
11303 (vaddvq_u64): Likewise.
11304 (vaddv_f32): Likewise.
11305 (vaddvq_f32): Likewise.
11306 (vaddvq_f64): Likewise.
11307 (vmaxv_f32): Likewise.
11308 (vmaxv_s8): Likewise.
11309 (vmaxv_s16): Likewise.
11310 (vmaxv_s32): Likewise.
11311 (vmaxv_u8): Likewise.
11312 (vmaxv_u16): Likewise.
11313 (vmaxv_u32): Likewise.
11314 (vmaxvq_f32): Likewise.
11315 (vmaxvq_f64): Likewise.
11316 (vmaxvq_s8): Likewise.
11317 (vmaxvq_s16): Likewise.
11318 (vmaxvq_s32): Likewise.
11319 (vmaxvq_u8): Likewise.
11320 (vmaxvq_u16): Likewise.
11321 (vmaxvq_u32): Likewise.
11322 (vmaxnmv_f32): Likewise.
11323 (vmaxnmvq_f32): Likewise.
11324 (vmaxnmvq_f64): Likewise.
11325 (vminv_f32): Likewise.
11326 (vminv_s8): Likewise.
11327 (vminv_s16): Likewise.
11328 (vminv_s32): Likewise.
11329 (vminv_u8): Likewise.
11330 (vminv_u16): Likewise.
11331 (vminv_u32): Likewise.
11332 (vminvq_f32): Likewise.
11333 (vminvq_f64): Likewise.
11334 (vminvq_s8): Likewise.
11335 (vminvq_s16): Likewise.
11336 (vminvq_s32): Likewise.
11337 (vminvq_u8): Likewise.
11338 (vminvq_u16): Likewise.
11339 (vminvq_u32): Likewise.
11340 (vminnmv_f32): Likewise.
11341 (vminnmvq_f32): Likewise.
11342 (vminnmvq_f64): Likewise.
11344 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11346 * config/aarch64/aarch64-simd.md
11347 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11348 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11349 (*aarch64_mul3_elt<mode>): Likewise.
11350 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11351 (*aarch64_mul3_elt_to_64v2df): Likewise.
11352 (*aarch64_mla_elt<mode>): Likewise.
11353 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11354 (*aarch64_mls_elt<mode>): Likewise.
11355 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11356 (*aarch64_fma4_elt<mode>): Likewise.
11357 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11358 (*aarch64_fma4_elt_to_64v2df): Likewise.
11359 (*aarch64_fnma4_elt<mode>): Likewise.
11360 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11361 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11362 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11363 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11364 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11365 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11366 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11367 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11368 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11370 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11372 * config/aarch64/aarch64-simd.md
11373 (aarch64_be_checked_get_lane<mode>): New define_expand.
11374 * config/aarch64/aarch64-simd-builtins.def
11375 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11376 New builtin definition.
11377 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11378 Use new safe be builtin.
11380 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11382 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11384 (aarch64_be_st1<mode>): Likewise.
11385 (aarch_ld1<VALL:mode>): Define_expand modified.
11386 (aarch_st1<VALL:mode>): Likewise.
11387 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11388 (UNSPEC_ST1): Likewise.
11390 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11392 * config/microblaze/microblaze.md: Add trap insn and attribute
11394 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11396 PR preprocessor/58580
11397 * input.h (location_get_source_line): Take an additional line_size
11399 (void diagnostics_file_cache_fini): Declare new function.
11400 * input.c (struct fcache): New type.
11401 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11402 New static constants.
11403 (diagnostic_file_cache_init, total_lines_num)
11404 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11405 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11406 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11407 (get_next_line, read_next_line, goto_next_line, read_line_num):
11408 New static function definitions.
11409 (diagnostic_file_cache_fini): New function.
11410 (location_get_source_line): Take an additional output line_len
11411 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11413 * diagnostic.c (diagnostic_finish): Call
11414 diagnostic_file_cache_fini.
11415 (adjust_line): Take an additional input parameter for the length
11416 of the line, rather than calculating it with strlen.
11417 (diagnostic_show_locus): Adjust the use of
11418 location_get_source_line and adjust_line with respect to their new
11419 signature. While displaying a line now, do not stop at the first
11420 null byte. Rather, display the zero byte as a space and keep
11421 going until we reach the size of the line.
11422 * Makefile.in: Add vec.o to OBJS-libcommon
11424 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11425 Ilya Tocar <ilya.tocar@intel.com>
11427 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11428 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11429 (__builtin_ia32_kmov16): Ditto.
11430 * config/i386/i386.md (UNSPEC_KMOV): New.
11433 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11435 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11436 (_mm512_storeu_si512): Ditto.
11438 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11441 * rtl.h (get_referenced_operands): Declare.
11442 * recog.c (get_referenced_operands): New function.
11443 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11444 operands have been referenced when recording LO_SUM references.
11446 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11448 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11450 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11452 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11453 Enable for generic and recent AMD targets.
11455 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11457 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11458 ARG_SIZE note when adjustment was eliminated.
11460 2014-01-22 Jeff Law <law@redhat.com>
11462 PR tree-optimization/59597
11463 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11464 in file. Accept new argument REGISTERING and use it to modify
11465 dump output appropriately.
11466 (register_jump_thread): Corresponding changes.
11467 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11468 thread paths involving joiner blocks. Add code to dump cancelled
11469 jump threading paths.
11471 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11473 PR rtl-optimization/59477
11474 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11475 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11477 2014-01-22 Tom Tromey <tromey@redhat.com>
11479 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11481 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11483 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11485 PR rtl-optimization/59896
11486 * lra-constraints.c (process_alt_operands): Check unused note for
11487 matched operands of insn with no output reloads.
11489 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11491 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11492 (mips_move_from_gpr_cost): Likewise.
11494 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11496 PR rtl-optimization/59858
11497 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11498 ira_class_hard_regs_num.
11499 (process_alt_operands): Increase reject for dying matched operand.
11501 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11504 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11505 smaller than size, perform several stores or loads and stores
11506 at dst + count - size to store or copy all of size bytes, rather
11507 than just last modesize bytes.
11509 2014-01-20 DJ Delorie <dj@redhat.com>
11511 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11512 that CLOBBERs are REGs before propogating their values.
11514 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11516 PR middle-end/59789
11517 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11518 (cgraph_inline_failed_type): New function.
11519 * cgraph.h (DEFCIFCODE): Add type.
11520 (cgraph_inline_failed_type_t): New enum.
11521 (cgraph_inline_failed_type): New prototype.
11522 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11523 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11524 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11525 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11526 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11527 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11528 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11529 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11530 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11531 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11532 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11533 OPTIMIZATION_MISMATCH.
11534 * tree-inline.c (expand_call_inline): Emit errors during
11535 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11537 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11540 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11541 mode attribute in insn output.
11543 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11545 * output.h (output_constant): Delete.
11546 * varasm.c (output_constant): Make private.
11548 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11550 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11552 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11554 PR middle-end/59860
11555 * tree.h (fold_builtin_strcat): New prototype.
11556 * builtins.c (fold_builtin_strcat): No longer static. Add len
11557 argument, if non-NULL, don't call c_strlen. Optimize
11558 directly into __builtin_memcpy instead of __builtin_strcpy.
11559 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11560 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11562 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11564 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11565 for SImode_address_operand operands, having only a REG argument.
11567 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11569 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11570 loader name using mbig-endian.
11571 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11573 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11575 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11576 (-mtune): Likewise.
11579 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11581 * config/aarch64/aarch64-protos.h
11582 (aarch64_cannot_change_mode_class_ptr): Declare.
11583 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11584 aarch64_cannot_change_mode_class_ptr): New.
11585 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11586 backend hook aarch64_cannot_change_mode_class.
11588 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11590 * common/config/aarch64/aarch64-common.c
11591 (aarch64_handle_option): Don't handle any option order logic here.
11592 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11593 selected_cpu, warn on architecture version mismatch.
11594 (aarch64_override_options): Fix parsing order for option strings.
11596 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11597 Iain Sandoe <iain@codesourcery.com>
11600 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11601 warning. Amend comment to reflect current functionality.
11603 2014-01-20 Richard Biener <rguenther@suse.de>
11605 PR middle-end/59860
11606 * builtins.c (fold_builtin_strcat): Remove case better handled
11607 by tree-ssa-strlen.c.
11609 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11611 * config/aarch64/aarch64.opt
11612 (mcpu, march, mtune): Make case-insensitive.
11614 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11617 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11618 if operands[1] is a REG or ZERO_EXTEND of a REG.
11620 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11622 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11624 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11626 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11627 long non-pic millicode calls.
11629 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11631 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11633 2014-01-19 Kito Cheng <kito@0xlab.org>
11635 * builtins.c (expand_movstr): Check movstr expand done or fail.
11637 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11638 H.J. Lu <hongjiu.lu@intel.com>
11641 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11642 to DImode for zero-extended addresses.
11644 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11646 PR rtl-optimization/57763
11647 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11648 on the new indirect jump_insn and increment LABEL_NUSES (label).
11650 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11654 * config.gcc (x86_archs): New variable.
11655 (x86_64_archs): Likewise.
11656 (x86_cpus): Likewise.
11657 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11658 --with-arch/--with-cpu= options.
11659 Support --with-arch=/--with-cpu={nehalem,westmere,
11660 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11662 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11664 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11665 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11667 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11669 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11671 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11674 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11675 clear cpp_get_options (parse_in)->warn_unused_macros for
11676 ix86_target_macros_internal with cpp_define.
11678 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11680 * jump.c (delete_related_insns): Keep (use (insn))s.
11681 * reorg.c (redundant_insn): Check for barriers too.
11683 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11685 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11687 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11689 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11690 call to $$dyncall when TARGET_LONG_CALLS is true.
11692 2014-01-17 Jeff Law <law@redhat.com>
11694 * ree.c (combine_set_extension): Temporarily disable test for
11695 changing number of hard registers.
11697 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11699 PR middle-end/58125
11700 * ipa-inline-analysis.c (inline_free_summary):
11701 Do not free summary of aliases.
11703 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11705 PR middle-end/59706
11706 * gimplify.c (gimplify_expr): Use create_tmp_var
11707 instead of create_tmp_var_raw. If cond doesn't have
11708 integral type, don't add the IFN_ANNOTATE builtin at all.
11710 2014-01-17 Martin Jambor <mjambor@suse.cz>
11713 * ipa-cp.c (prev_edge_clone): New variable.
11714 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11715 Also resize prev_edge_clone vector.
11716 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11717 (ipcp_edge_removal_hook): New function.
11718 (ipcp_driver): Register ipcp_edge_removal_hook.
11720 2014-01-17 Andrew Pinski <apinski@cavium.com>
11721 Steve Ellcey <sellcey@mips.com>
11724 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11727 2014-01-17 Jeff Law <law@redhat.com>
11729 PR middle-end/57904
11730 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11731 so that pass_ccp runs first.
11733 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11735 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11736 (ix86_adjust_cost): Use !TARGET_XXX.
11737 (do_reorder_for_imul): Likewise.
11738 (swap_top_of_ready_list): Likewise.
11739 (ix86_sched_reorder): Likewise.
11741 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11743 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11744 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11745 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11746 (intel_memset): New. Duplicate slm_memset.
11747 (intel_cost): New. Duplicate slm_cost.
11748 (m_INTEL): New macro.
11749 (processor_target_table): Add "intel".
11750 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11751 with PROCESSOR_INTEL for "intel".
11752 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11753 PROCESSOR_SILVERMONT.
11754 (ix86_issue_rate): Likewise.
11755 (ix86_adjust_cost): Likewise.
11756 (ia32_multipass_dfa_lookahead): Likewise.
11757 (swap_top_of_ready_list): Likewise.
11758 (ix86_sched_reorder): Likewise.
11759 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11760 instead of TARGET_OPT_AGU.
11761 * config/i386/i386.h (TARGET_INTEL): New.
11762 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11763 (processor_type): Add PROCESSOR_INTEL.
11764 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11765 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11767 2014-01-17 Marek Polacek <polacek@redhat.com>
11770 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11773 2014-01-17 Richard Biener <rguenther@suse.de>
11775 PR tree-optimization/46590
11776 * opts.c (default_options_table): Add entries for
11777 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11778 all enabled at -O1 but not for -Og.
11779 * common.opt (fbranch-count-reg): Remove Init(1).
11780 (fmove-loop-invariants): Likewise.
11781 (ftree-pta): Likewise.
11783 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11785 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11786 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11787 decls to at least the GCC 4.8 used alignments.
11790 * tree-nested.c (convert_nonlocal_reference_stmt,
11791 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11792 of GIMPLE_BIND stmts, adjust associated decls.
11794 2014-01-17 Richard Biener <rguenther@suse.de>
11796 PR tree-optimization/46590
11797 * vec.h (vec<>::bseach): New member function implementing
11798 binary search according to C89 bsearch.
11799 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11800 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11801 bitmap pointer again. Make accesses_in_loop a flat array.
11802 (mem_ref_obstack): New global.
11803 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11804 (mark_ref_stored): Likewise.
11805 (ref_indep_loop_p_2): Likewise.
11806 (set_ref_stored_in_loop): New helper function.
11807 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11808 (memref_free): Adjust.
11809 (record_mem_ref_loc): Simplify.
11810 (gather_mem_refs_stmt): Adjust.
11811 (sort_locs_in_loop_postorder_cmp): New function.
11812 (analyze_memory_references): Sort accesses_in_loop after
11813 loop postorder number.
11814 (find_ref_loc_in_loop_cmp): New function.
11815 (for_all_locs_in_loop): Find relevant cluster of locs in
11816 accesses_in_loop and iterate without recursion.
11817 (execute_sm): Avoid uninit warning.
11818 (struct ref_always_accessed): Simplify.
11819 (ref_always_accessed::operator ()): Likewise.
11820 (ref_always_accessed_p): Likewise.
11821 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11822 loop postorder numbers here.
11823 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11826 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11829 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11830 on decls for which assemble_alias has been called.
11832 2014-01-17 Nick Clifton <nickc@redhat.com>
11834 * config/msp430/msp430.opt: (mcpu): New option.
11835 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11836 (msp430_option_override): Parse target_cpu. If the MCU name
11837 matches a generic string, clear target_mcu.
11838 (msp430_attr): Allow numeric interrupt values up to 63.
11839 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11840 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11842 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11844 * config/msp430/msp430.md (popm): Use %J rather than %I.
11845 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11846 (addhi_cy_i): Use immediate_operand for operand 2.
11847 * doc/invoke.texi: Document -mcpu option.
11849 2014-01-17 Richard Biener <rguenther@suse.de>
11851 PR rtl-optimization/38518
11852 * df.h (df_analyze_loop): Declare.
11853 * df-core.c: Include cfgloop.h.
11854 (df_analyze_1): Split out main part of df_analyze.
11855 (df_analyze): Adjust.
11856 (loop_inverted_post_order_compute): New function.
11857 (loop_post_order_compute): Likewise.
11858 (df_analyze_loop): New function avoiding whole-function
11859 postorder computes.
11860 * loop-invariant.c (find_defs): Use df_analyze_loop.
11861 (find_invariants): Adjust.
11862 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11864 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11866 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11867 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11869 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11871 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11872 traversal when removing references.
11874 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11877 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11879 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11881 PR middle-end/56791
11882 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11883 pushing a reload for an autoinc when we had previously reloaded an
11884 inner part of the address.
11886 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11888 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11890 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11891 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11892 when not giving up or versioning for alias only because of
11894 (vect_analyze_data_ref_dependences): Set to true.
11895 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11897 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11898 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11901 PR middle-end/58344
11902 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11905 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11906 operand 0 predicate for gathers, use a new pseudo as subtarget.
11908 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11910 PR middle-end/59609
11911 * lra-constraints.c (process_alt_operands): Add printing debug info.
11912 Check absence of input/output reloads for matched operands too.
11914 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11916 PR rtl-optimization/59835
11917 * ira.c (ira_init_register_move_cost): Increase cost for
11920 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11922 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11924 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11927 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11928 non-register objects. Use gen_(high/low)part more consistently.
11931 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11934 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11935 endian support, remove tests for WORDS_BIG_ENDIAN.
11936 (p8_mfvsrd_3_<mode>): Likewise.
11937 (reload_gpr_from_vsx<mode>): Likewise.
11938 (reload_gpr_from_vsxsf): Likewise.
11939 (p8_mfvsrd_4_disf): Likewise.
11941 2014-01-16 Richard Biener <rguenther@suse.de>
11943 PR rtl-optimization/46590
11944 * lcm.c (compute_antinout_edge): Use postorder iteration.
11945 (compute_laterin): Use inverted postorder iteration.
11947 2014-01-16 Nick Clifton <nickc@redhat.com>
11949 PR middle-end/28865
11950 * varasm.c (output_constant): Return the number of bytes actually
11952 (output_constructor_array_range): Update the field size with the
11953 number of bytes emitted by output_constant.
11954 (output_constructor_regular_field): Likewise. Also do not
11955 complain if the total number of bytes emitted is now greater
11956 than the expected fieldpos.
11957 * output.h (output_constant): Update prototype and descriptive comment.
11959 2014-01-16 Marek Polacek <polacek@redhat.com>
11961 PR middle-end/59827
11962 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11963 it is error_mark_node.
11965 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11967 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11968 VALID_AVX256_REG_OR_OI_MODE.
11970 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11972 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11973 current procedure should be profiled.
11975 2014-01-15 Andrew Pinski <apinski@cavium.com>
11977 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11978 of moving from/to the STACK_REG register class.
11980 2014-01-15 Richard Henderson <rth@redhat.com>
11983 * reginfo.c (global_regs_decl): Globalize.
11984 * rtl.h (global_regs_decl): Declare.
11985 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11986 reserved via global_regs.
11988 2014-01-15 Teresa Johnson <tejohnson@google.com>
11990 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11992 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11994 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11995 and vmulosh rather than call gen_vec_widen_smult_*.
11996 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11997 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11998 (vec_widen_smult_even_v16qi): Likewise.
11999 (vec_widen_umult_even_v8hi): Likewise.
12000 (vec_widen_smult_even_v8hi): Likewise.
12001 (vec_widen_umult_odd_v16qi): Likewise.
12002 (vec_widen_smult_odd_v16qi): Likewise.
12003 (vec_widen_umult_odd_v8hi): Likewise.
12004 (vec_widen_smult_odd_v8hi): Likewise.
12005 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12006 vmuloub rather than call gen_vec_widen_umult_*.
12007 (vec_widen_umult_lo_v16qi): Likewise.
12008 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12009 vmulosb rather than call gen_vec_widen_smult_*.
12010 (vec_widen_smult_lo_v16qi): Likewise.
12011 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12012 rather than call gen_vec_widen_umult_*.
12013 (vec_widen_umult_lo_v8hi): Likewise.
12014 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12015 rather than call gen_vec_widen_smult_*.
12016 (vec_widen_smult_lo_v8hi): Likewise.
12018 2014-01-15 Jeff Law <law@redhat.com>
12020 PR tree-optimization/59747
12021 * ree.c (find_and_remove_re): Properly handle case where a second
12022 eliminated extension requires widening a copy created for elimination
12023 of a prior extension.
12024 (combine_set_extension): Ensure that the number of hard regs needed
12025 for a destination register does not change when we widen it.
12027 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12029 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12030 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12031 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12032 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12033 (avr-*-rtems*): Likewise.
12034 (bfin*-rtems*): Likewise.
12035 (moxie-*-rtems*): Likewise.
12036 (h8300-*-rtems*): Likewise.
12037 (i[34567]86-*-rtems*): Likewise.
12038 (lm32-*-rtems*): Likewise.
12039 (m32r-*-rtems*): Likewise.
12040 (m68k-*-rtems*): Likewise.
12041 (microblaze*-*-rtems*): Likewise.
12042 (mips*-*-rtems*): Likewise.
12043 (powerpc-*-rtems*): Likewise.
12044 (sh-*-rtems*): Likewise.
12045 (sparc-*-rtems*): Likewise.
12046 (sparc64-*-rtems*): Likewise.
12047 (v850-*-rtems*): Likewise.
12048 (m32c-*-rtems*): Likewise.
12050 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12052 PR rtl-optimization/59511
12053 * ira.c (ira_init_register_move_cost): Use memory costs for some
12054 cases of register move cost calculations.
12055 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12056 instead of BB frequency.
12057 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12058 * lra-assigns.c (find_hard_regno_for): Ditto.
12060 2014-01-15 Richard Biener <rguenther@suse.de>
12062 PR tree-optimization/59822
12063 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12064 (vectorizable_load): Use it to hoist defs of uses of invariant
12065 loads out of the loop.
12067 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12068 Kugan Vivekanandarajah <kuganv@linaro.org>
12071 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12074 2014-01-15 Richard Biener <rguenther@suse.de>
12076 PR rtl-optimization/59802
12077 * lcm.c (compute_available): Use inverted postorder to seed
12078 the initial worklist.
12080 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12083 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12084 ADDR_REGS for invalid symrefs in non-PIC code.
12086 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12089 * builtins.c (determine_block_size): Initialize *probable_max_size
12090 even if len_rtx is CONST_INT.
12092 2014-01-14 Andrew Pinski <apinski@cavium.com>
12094 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12095 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12096 (cortexa53_tunings): Likewise.
12097 (aarch64_sched_issue_rate): New function.
12098 (TARGET_SCHED_ISSUE_RATE): Define.
12100 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12102 * ira-costs.c (find_costs_and_classes): Add missed
12103 ira_init_register_move_cost_if_necessary.
12105 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12108 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12110 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12113 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12114 to indicate if type is used for function return value. Warn ABI
12115 change if the vector mode isn't available for function return value.
12116 (ix86_function_arg_advance): Pass false to type_natural_mode.
12117 (ix86_function_arg): Likewise.
12118 (ix86_gimplify_va_arg): Likewise.
12119 (function_arg_32): Don't warn ABI change.
12120 (ix86_function_value): Pass true to type_natural_mode.
12121 (ix86_return_in_memory): Likewise.
12122 (ix86_struct_value_rtx): Removed.
12123 (TARGET_STRUCT_VALUE_RTX): Likewise.
12125 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12127 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12128 converting a conditional jump into a conditional return.
12130 2014-01-14 Richard Biener <rguenther@suse.de>
12132 PR tree-optimization/58921
12133 PR tree-optimization/59006
12134 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12135 hoisting invariant stmts.
12136 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12137 invariant loads on the preheader edge if possible.
12139 2014-01-14 Joey Ye <joey.ye@arm.com>
12141 * doc/plugin.texi (Building GCC plugins): Update to C++.
12143 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12145 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12146 (_mm_rcp28_round_ss): Ditto.
12147 (_mm_rsqrt28_round_sd): Ditto.
12148 (_mm_rsqrt28_round_ss): Ditto.
12149 (_mm_rcp28_sd): Ditto.
12150 (_mm_rcp28_ss): Ditto.
12151 (_mm_rsqrt28_sd): Ditto.
12152 (_mm_rsqrt28_ss): Ditto.
12153 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12154 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12155 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12156 (IX86_BUILTIN_RCP28SD): Ditto.
12157 (IX86_BUILTIN_RCP28SS): Ditto.
12158 (IX86_BUILTIN_RSQRT28SD): Ditto.
12159 (IX86_BUILTIN_RSQRT28SS): Ditto.
12160 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12161 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12162 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12163 (ix86_expand_special_args_builtin): Expand new FTYPE.
12164 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12165 (srcp14<mode>): Make insn unary.
12166 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12167 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12168 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12169 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12170 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12171 Fix rounding: make it SAE only.
12172 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12174 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12176 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12177 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12178 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12179 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12180 (round_saeonly_mask_scalar_operand4): Ditto.
12181 (round_saeonly_mask_scalar_op3): Ditto.
12182 (round_saeonly_mask_scalar_op4): Ditto.
12184 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12186 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12187 Implement -maltivec=be for vec_insert and vec_extract.
12189 2014-01-10 DJ Delorie <dj@redhat.com>
12191 * config/msp430/msp430.md (call_internal): Don't allow memory
12192 references with SP as the base register.
12193 (call_value_internal): Likewise.
12194 * config/msp430/constraints.md (Yc): New. For memory references
12195 that don't use SP as a base register.
12197 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12198 "an integer without a # prefix"
12199 * config/msp430/msp430.md (epilogue_helper): Use it.
12201 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12204 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12205 AVX512F gather builtins.
12206 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12207 on gather decls with INTEGER_TYPE masktype.
12208 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12209 directly into the builtin rather than hoisting it before loop.
12211 PR tree-optimization/59387
12212 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12213 (scev_const_prop): If folded_casts and type has undefined overflow,
12214 use force_gimple_operand instead of force_gimple_operand_gsi and
12215 for each added stmt if it is assign with
12216 arith_code_with_undefined_signed_overflow, call
12217 rewrite_to_defined_overflow.
12218 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12219 gimple-fold.h instead.
12220 (arith_code_with_undefined_signed_overflow,
12221 rewrite_to_defined_overflow): Moved to ...
12222 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12223 rewrite_to_defined_overflow): ... here. No longer static.
12224 Include gimplify-me.h.
12225 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12226 rewrite_to_defined_overflow): New prototypes.
12228 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12230 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12232 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12234 * builtins.c (get_object_alignment_2): Minor tweak.
12235 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12237 2014-01-13 Christian Bruel <christian.bruel@st.com>
12239 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12240 optimized non constant lengths.
12242 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12245 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12246 load as __atomic_load_N if possible.
12248 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12250 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12252 (rs6000_expand_builtin): Adjust call.
12254 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12257 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12258 * config/rs6000/rs6000.c: Include target-globals.h.
12259 (rs6000_set_current_function): Instead of doing target_reinit
12260 unconditionally, use save_target_globals_default_opts and
12261 restore_target_globals.
12263 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12265 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12266 (rs6000_expand_builtin): Handle mffs and mtfsf.
12267 (rs6000_init_builtins): Define mffs and mtfsf.
12268 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12269 (rs6000_mffs): New pattern.
12270 (rs6000_mtfsf): New pattern.
12272 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12274 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12275 Start narrowing with START. Apply candidate-use pair
12276 and check overall cost in narrowing.
12277 (iv_ca_prune): Pass new argument.
12279 2014-01-10 Jeff Law <law@redhat.com>
12281 PR middle-end/59743
12282 * ree.c (combine_reaching_defs): Ensure the defining statement
12283 occurs before the extension when optimizing extensions with
12284 different source and destination hard registers.
12286 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12289 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12290 vtables into the type inheritance graph.
12292 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12294 PR rtl-optimization/59754
12295 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12296 modes in the REGNO != REGNO case.
12298 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12300 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12302 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12304 PR tree-optimization/59745
12305 * tree-predcom.c (tree_predictive_commoning_loop): Call
12306 free_affine_expand_cache if giving up because components is NULL.
12308 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12309 GC in payload of target_globals struct instead of allocating them on
12310 the heap and the larger structs separately using GC.
12311 * target-globals.h (struct target_globals): Make regs, hard_regs,
12312 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12313 of GTY((skip)) and change type to void *.
12314 (reset_target_globals): Cast loads from those fields to corresponding
12317 2014-01-10 Steve Ellcey <sellcey@mips.com>
12320 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12321 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12322 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12324 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12327 * aarch64-modes.def (CC_Zmode): New flags mode.
12328 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12329 represents an equality.
12330 (aarch64_get_condition_code): Handle CC_Zmode.
12331 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12333 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12335 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12336 extraction in good case.
12338 2014-01-10 Richard Biener <rguenther@suse.de>
12340 PR tree-optimization/59374
12341 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12342 checking after SLP discovery. Mark stmts not participating
12343 in any SLP instance properly.
12345 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12347 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12348 when handling a SET rtx.
12350 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12352 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12353 (cortex-a57): Likewise.
12354 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12356 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12358 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12359 non-iwmmxt builtins.
12361 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12365 * ipa-devirt.c record_target_from_binfo): Take as argument
12366 stack of binfos and lookup matching one for virtual inheritance.
12367 (possible_polymorphic_call_targets_1): Update.
12369 2014-01-10 Huacai Chen <chenhc@lemote.com>
12371 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12372 kernel strings for Loongson-2E/2F/3A.
12374 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12376 PR middle-end/59670
12377 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12378 is_gimple_call before calling gimple_call_internal_p.
12380 2014-01-09 Steve Ellcey <sellcey@mips.com>
12382 * Makefile.in (TREE_FLOW_H): Remove.
12383 (TREE_SSA_H): Add file names from tree-flow.h.
12384 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12385 * tree.h: Remove tree-flow.h reference.
12386 * hash-table.h: Remove tree-flow.h reference.
12387 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12388 reference with tree-ssa-loop.h.
12390 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12392 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12393 default element-order behavior for -maltivec.
12394 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12395 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12396 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12397 when targeting big endian, at least for now.
12398 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12400 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12402 PR middle-end/47735
12403 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12404 var satisfies use_register_for_decl, just take into account type
12405 alignment, rather than decl alignment.
12407 PR tree-optimization/59622
12408 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12409 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12410 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12411 Don't devirtualize for inplace at all. For targets.length () == 1,
12412 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12414 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12416 * config/i386/i386.md (cpu): Remove the unused btver1.
12418 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12420 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12422 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12425 * tree-core.h (struct target_globals): New forward declaration.
12426 (struct tree_target_option): Add globals field.
12427 * tree.h (TREE_TARGET_GLOBALS): Define.
12428 (prepare_target_option_nodes_for_pch): New prototype.
12429 * target-globals.h (struct target_globals): Define even if
12430 !SWITCHABLE_TARGET.
12431 * tree.c (prepare_target_option_node_for_pch,
12432 prepare_target_option_nodes_for_pch): New functions.
12433 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12434 * config/i386/i386.c: Include target-globals.h.
12435 (ix86_set_current_function): Instead of doing target_reinit
12436 unconditionally, use save_target_globals_default_opts and
12437 restore_target_globals.
12439 2014-01-09 Richard Biener <rguenther@suse.de>
12441 PR tree-optimization/59715
12442 * tree-cfg.h (split_critical_edges): Declare.
12443 * tree-cfg.c (split_critical_edges): Export.
12444 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12446 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12448 * cfgexpand.c (expand_stack_vars): Optionally disable
12449 asan stack protection.
12450 (expand_used_vars): Likewise.
12451 (partition_stack_vars): Likewise.
12452 * asan.c (asan_emit_stack_protection): Optionally disable
12453 after return stack usage.
12454 (instrument_derefs): Optionally disable memory access instrumentation.
12455 (instrument_builtin_call): Likewise.
12456 (instrument_strlen_call): Likewise.
12457 (asan_protect_global): Optionally disable global variables protection.
12458 * doc/invoke.texi: Added doc for new options.
12459 * params.def: Added new options.
12460 * params.h: Likewise.
12462 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12464 PR rtl-optimization/59724
12465 * ifcvt.c (cond_exec_process_if_block): Don't call
12466 flow_find_head_matching_sequence with 0 longest_match.
12467 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12468 non-active insns if !stop_after.
12469 (try_head_merge_bb): Revert 2014-01-07 changes.
12471 2014-01-08 Jeff Law <law@redhat.com>
12473 * ree.c (get_sub_rtx): New function, extracted from...
12474 (merge_def_and_ext): Here.
12475 (combine_reaching_defs): Use get_sub_rtx.
12477 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12479 * cgraph.h (varpool_variable_node): Do not choke on null node.
12481 2014-01-08 Catherine Moore <clm@codesourcery.com>
12483 * config/mips/mips.md (simple_return): Attempt to use JRC
12485 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12487 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12489 PR rtl-optimization/59137
12490 * reorg.c (steal_delay_list_from_target): Call update_block for
12492 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12494 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12496 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12497 two duplicate entries.
12499 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12502 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12504 * config/mips/mips.c (mips_truncated_op_cost): New function.
12505 (mips_rtx_costs): Adjust test for BADDU.
12506 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12508 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12510 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12511 (*baddu_si): ...this new pattern.
12513 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12516 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12518 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12520 PR middle-end/57748
12521 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12523 (expand_expr, expand_normal): Adjust.
12524 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12525 inner_reference_p. Use inner_reference_p to expand inner references.
12526 (store_expr): Adjust.
12527 * cfgexpand.c (expand_call_stmt): Adjust.
12529 2014-01-08 Rong Xu <xur@google.com>
12531 * gcov-io.c (gcov_var): Move from gcov-io.h.
12532 (gcov_position): Ditto.
12533 (gcov_is_error): Ditto.
12534 (gcov_rewrite): Ditto.
12535 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12536 only part to libgcc/libgcov.h.
12538 2014-01-08 Marek Polacek <polacek@redhat.com>
12540 PR middle-end/59669
12541 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12543 2014-01-08 Marek Polacek <polacek@redhat.com>
12546 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12548 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12550 PR rtl-optimization/59649
12551 * stor-layout.c (get_mode_bounds): For BImode return
12552 0 and STORE_FLAG_VALUE.
12554 2014-01-08 Richard Biener <rguenther@suse.de>
12556 PR middle-end/59630
12557 * gimple.h (is_gimple_builtin_call): Remove.
12558 (gimple_builtin_call_types_compatible_p): New.
12559 (gimple_call_builtin_p): New overload.
12560 * gimple.c (is_gimple_builtin_call): Remove.
12561 (validate_call): Rename to ...
12562 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12563 check return types.
12564 (validate_type): New static function.
12565 (gimple_call_builtin_p): New overload and adjust.
12566 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12567 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12568 (gimple_fold_stmt_to_constant_1): Likewise.
12569 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12571 2014-01-08 Richard Biener <rguenther@suse.de>
12573 PR middle-end/59471
12574 * gimplify.c (gimplify_expr): Gimplify register-register type
12575 VIEW_CONVERT_EXPRs to separate stmts.
12577 2014-01-07 Jeff Law <law@redhat.com>
12579 PR middle-end/53623
12580 * ree.c (combine_set_extension): Handle case where source
12581 and destination registers in an extension insn are different.
12582 (combine_reaching_defs): Allow source and destination registers
12583 in extension to be different under limited circumstances.
12584 (add_removable_extension): Remove restriction that the
12585 source and destination registers in the extension are the same.
12586 (find_and_remove_re): Emit a copy from the extension's
12587 destination to its source after the defining insn if
12588 the source and destination registers are different.
12590 PR middle-end/59285
12591 * ifcvt.c (merge_if_block): If we are merging a block with more than
12592 one successor with a block with no successors, remove any BARRIER
12593 after the second block.
12595 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12597 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12599 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12602 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12603 for 14-bit register offsets when INT14_OK_STRICT is false.
12605 2014-01-07 Roland Stigge <stigge@antcom.de>
12606 Michael Meissner <meissner@linux.vnet.ibm.com>
12609 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12610 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12612 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12614 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12617 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12619 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12620 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12621 rtx is const0_rtx or not.
12623 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12626 * target-globals.c (save_target_globals): Remove this_fn_optab
12628 * toplev.c: Include optabs.h.
12629 (target_reinit): Temporarily restore the global options if another
12630 set of options are in force.
12632 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12634 PR rtl-optimization/58668
12635 * cfgcleanup.c (flow_find_cross_jump): Don't count
12636 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12637 to determine what is counted.
12638 (flow_find_head_matching_sequence): Use active_insn_p to determine
12640 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12642 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12643 determine what is counted.
12645 PR tree-optimization/59643
12646 * tree-predcom.c (split_data_refs_to_components): If one dr is
12647 read and one write, determine_offset fails and the write isn't
12648 in the bad component, just put the read into the bad component.
12650 2014-01-07 Mike Stump <mikestump@comcast.net>
12651 Jakub Jelinek <jakub@redhat.com>
12654 * tree-core.h (struct tree_optimization_option): Change optabs
12655 type from unsigned char * to void *.
12656 * optabs.c (init_tree_optimization_optabs): Adjust
12657 TREE_OPTIMIZATION_OPTABS initialization.
12659 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12662 * config/i386/i386.h (struct machine_function): Add
12663 no_drap_save_restore field.
12664 * config/i386/i386.c (ix86_save_reg): Use
12665 !cfun->machine->no_drap_save_restore instead of
12666 crtl->stack_realign_needed.
12667 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12668 this function clears frame_pointer_needed. Set
12669 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12670 and DRAP reg is needed.
12672 2014-01-06 Marek Polacek <polacek@redhat.com>
12675 * doc/implement-c.texi: Mention that other integer types are
12676 permitted as bit-field types in strictly conforming mode.
12678 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12680 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12681 is newly allocated.
12683 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12685 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12687 2014-01-06 Martin Jambor <mjambor@suse.cz>
12690 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12692 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12694 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12698 * var-tracking.c (add_stores): Preserve the value of the source even if
12699 we don't record the store.
12701 2014-01-06 Terry Guo <terry.guo@arm.com>
12703 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12705 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12708 * config/darwin.c (darwin_function_section): Adjust return values to
12709 correspond to optimisation changes made in r206070.
12711 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12713 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12714 from prefetch_block tune setting.
12715 (nocona_cost): Correct size of prefetch block to 64.
12717 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12719 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12720 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12721 used to save the static chain register in the computation of the offset
12722 from which the FP registers need to be restored.
12724 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12726 PR tree-optimization/59519
12727 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12728 ICE if get_current_def (current_new_name) is already non-NULL, as long
12729 as it is a phi result of some other phi in *new_exit_bb that has
12732 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12733 or vmovdqu* for misaligned_operand.
12734 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12735 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12736 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12737 aligned_mem for AVX512F masked aligned load and store builtins and for
12738 non-temporal moves.
12740 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12742 PR tree-optimization/59651
12743 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12744 Address range for negative step should be added by TYPE_SIZE_UNIT.
12746 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12748 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12749 overlapping registers also for registers other than the stack pointer.
12751 2014-01-03 Marek Polacek <polacek@redhat.com>
12754 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12757 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12760 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12763 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12764 (push splitter): Use <P:MODE_SIZE> instead of
12765 GET_MODE_SIZE (<P:MODE>mode).
12766 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12767 (mov -1, reg peephole2): Likewise.
12768 * config/i386/sse.md (*mov<mode>_internal,
12769 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12770 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12771 *<code><mode>3, *andnot<mode>3<mask_name>,
12772 <mask_codefor><code><mode>3<mask_name>): Likewise.
12773 * config/i386/subst.md (mask_mode512bit_condition,
12774 sd_mask_mode512bit_condition): Likewise.
12776 2014-01-02 Xinliang David Li <davidxl@google.com>
12778 PR tree-optimization/59303
12779 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12780 (dump_predicates): Better output format.
12781 (pred_equal_p): New function.
12782 (is_neq_relop_p): Ditto.
12783 (is_neq_zero_form_p): Ditto.
12784 (pred_expr_equal_p): Ditto.
12785 (pred_neg_p): Ditto.
12786 (simplify_pred): Ditto.
12787 (simplify_preds_2): Ditto.
12788 (simplify_preds_3): Ditto.
12789 (simplify_preds_4): Ditto.
12790 (simplify_preds): Ditto.
12791 (push_pred): Ditto.
12792 (push_to_worklist): Ditto.
12793 (get_pred_info_from_cmp): Ditto.
12794 (is_degenerated_phi): Ditto.
12795 (normalize_one_pred_1): Ditto.
12796 (normalize_one_pred): Ditto.
12797 (normalize_one_pred_chain): Ditto.
12798 (normalize_preds): Ditto.
12799 (normalize_cond_1): Remove function.
12800 (normalize_cond): Ditto.
12801 (is_gcond_subset_of): Ditto.
12802 (is_subset_of_any): Ditto.
12803 (is_or_set_subset_of): Ditto.
12804 (is_and_set_subset_of): Ditto.
12805 (is_norm_cond_subset_of): Ditto.
12806 (pred_chain_length_cmp): Ditto.
12807 (convert_control_dep_chain_into_preds): Type change.
12808 (find_predicates): Ditto.
12809 (find_def_preds): Ditto.
12810 (destroy_predicates_vecs): Ditto.
12811 (find_matching_predicates_in_rest_chains): Ditto.
12812 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12813 (is_pred_expr_subset): Ditto.
12814 (is_pred_chain_subset_of): Ditto.
12815 (is_included_in): Ditto.
12816 (is_superset_of): Ditto.
12818 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12820 Update copyright years.
12822 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12824 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12825 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12826 config/arc/arc.md, config/arc/arc.opt,
12827 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12828 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12829 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12830 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12831 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12832 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12833 vtable-verify.c, vtable-verify.h: Use the standard form for the
12836 2014-01-02 Tobias Burnus <burnus@net-b.de>
12838 * gcc.c (process_command): Update copyright notice dates.
12839 * gcov-dump.c: Ditto.
12841 * doc/cpp.texi: Bump @copying's copyright year.
12842 * doc/cppinternals.texi: Ditto.
12843 * doc/gcc.texi: Ditto.
12844 * doc/gccint.texi: Ditto.
12845 * doc/gcov.texi: Ditto.
12846 * doc/install.texi: Ditto.
12847 * doc/invoke.texi: Ditto.
12849 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12851 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12853 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12855 * config/i386/sse.md (*mov<mode>_internal): Guard
12856 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12858 PR rtl-optimization/59647
12859 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12860 new_rtx into UNSIGNED_FLOAT rtxes.
12862 Copyright (C) 2014 Free Software Foundation, Inc.
12864 Copying and distribution of this file, with or without modification,
12865 are permitted in any medium without royalty provided the copyright
12866 notice and this notice are preserved.