1 2017-09-12 Richard Sandiford <richard.sandiford@linaro.org>
3 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Use
4 end_hard_regno instead of hard_regno_nregs.
5 * config/s390/s390.c (s390_reg_clobbered_rtx): Likewise.
6 * config/sparc/sparc.h (ASM_DECLARE_REGISTER_GLOBAL): Likewise.
7 * config/visium/visium.c (visium_hard_regno_mode_ok): Likewise.
8 * ira-color.c (improve_allocation): Likewise.
9 * lra-assigns.c (find_hard_regno_for_1): Likewise.
10 * lra-lives.c (mark_regno_live): Likewise.
11 (mark_regno_dead): Likewise.
12 * lra-remat.c (operand_to_remat): Likewise.
13 * lra.c (collect_non_operand_hard_regs): Likewise.
14 * postreload.c (reload_combine_note_store): Likewise.
15 (move2add_valid_value_p): Likewise.
16 * reload.c (regno_clobbered_p): Likewise.
18 2017-09-12 Richard Sandiford <richard.sandiford@linaro.org>
20 * config/frv/frv.c (FOR_EACH_REGNO): Use END_REGNO instead of
22 * config/v850/v850.c (v850_reorg): Likewise.
23 * reload.c (refers_to_regno_for_reload_p): Likewise.
24 (find_equiv_reg): Likewise.
25 * reload1.c (reload_reg_reaches_end_p): Likewise.
27 2017-09-12 Richard Sandiford <richard.sandiford@linaro.org>
29 * caller-save.c (add_used_regs): Use REG_NREGS instead of
31 * config/aarch64/aarch64.c (aarch64_split_combinev16qi): Likewise.
32 * config/arm/arm.c (output_move_neon): Likewise.
33 (arm_attr_length_move_neon): Likewise.
34 (neon_split_vcombine): Likewise.
35 * config/c6x/c6x.c (c6x_mark_reg_read): Likewise.
36 (c6x_mark_reg_written): Likewise.
37 (c6x_dwarf_register_span): Likewise.
38 * config/i386/i386.c (ix86_save_reg): Likewise.
39 * config/ia64/ia64.c (mark_reg_gr_used_mask): Likewise.
40 (rws_access_reg): Likewise.
41 * config/s390/s390.c (s390_call_saved_register_used): Likewise.
42 * mode-switching.c (create_pre_exit): Likewise.
43 * ree.c (combine_reaching_defs): Likewise.
44 (add_removable_extension): Likewise.
45 * regcprop.c (find_oldest_value_reg): Likewise.
46 (copyprop_hardreg_forward_1): Likewise.
47 * reload.c (reload_inner_reg_of_subreg): Likewise.
48 (push_reload): Likewise.
49 (combine_reloads): Likewise.
50 (find_dummy_reload): Likewise.
51 (reload_adjust_reg_for_mode): Likewise.
52 * reload1.c (find_reload_regs): Likewise.
53 (forget_old_reloads_1): Likewise.
54 (reload_reg_free_for_value_p): Likewise.
55 (reload_adjust_reg_for_temp): Likewise.
56 (emit_reload_insns): Likewise.
57 (delete_output_reload): Likewise.
58 * sel-sched.c (choose_best_reg_1): Likewise.
59 (choose_best_pseudo_reg): Likewise.
61 2017-09-12 Richard Sandiford <richard.sandiford@linaro.org>
62 Alan Hayward <alan.hayward@arm.com>
63 David Sherwood <david.sherwood@arm.com>
65 * defaults.h (SLOW_UNALIGNED_ACCESS): Delete.
66 * target.def (slow_unaligned_access): New hook.
67 * targhooks.h (default_slow_unaligned_access): Declare.
68 * targhooks.c (default_slow_unaligned_access): New function.
69 * doc/tm.texi.in (SLOW_UNALIGNED_ACCESS): Replace with...
70 (TARGET_SLOW_UNALIGNED_ACCESS): ...this.
71 * doc/tm.texi: Regenerate.
72 * config/alpha/alpha.h (SLOW_UNALIGNED_ACCESS): Delete.
73 * config/arm/arm.h (SLOW_UNALIGNED_ACCESS): Delete.
74 * config/i386/i386.h (SLOW_UNALIGNED_ACCESS): Delete commented-out
76 * config/powerpcspe/powerpcspe.h (SLOW_UNALIGNED_ACCESS): Delete.
77 * config/powerpcspe/powerpcspe.c (TARGET_SLOW_UNALIGNED_ACCESS):
79 (rs6000_slow_unaligned_access): New function.
80 (rs6000_emit_move): Use it instead of SLOW_UNALIGNED_ACCESS.
81 (expand_block_compare): Likewise.
82 (expand_strn_compare): Likewise.
83 (rs6000_rtx_costs): Likewise.
84 * config/riscv/riscv.h (SLOW_UNALIGNED_ACCESS): Delete.
85 (riscv_slow_unaligned_access): Likewise.
86 * config/riscv/riscv.c (riscv_slow_unaligned_access): Rename to...
87 (riscv_slow_unaligned_access_p): ...this and make static.
88 (riscv_option_override): Update accordingly.
89 (riscv_slow_unaligned_access): New function.
90 (TARGET_SLOW_UNALIGNED_ACCESS): Redefine.
91 * config/rs6000/rs6000.h (SLOW_UNALIGNED_ACCESS): Delete.
92 * config/rs6000/rs6000.c (TARGET_SLOW_UNALIGNED_ACCESS): Redefine.
93 (rs6000_slow_unaligned_access): New function.
94 (rs6000_emit_move): Use it instead of SLOW_UNALIGNED_ACCESS.
95 (rs6000_rtx_costs): Likewise.
96 * config/rs6000/rs6000-string.c (expand_block_compare)
97 (expand_strn_compare): Use targetm.slow_unaligned_access instead
98 of SLOW_UNALIGNED_ACCESS.
99 * config/tilegx/tilegx.h (SLOW_UNALIGNED_ACCESS): Delete.
100 * config/tilepro/tilepro.h (SLOW_UNALIGNED_ACCESS): Delete.
101 * calls.c (expand_call): Use targetm.slow_unaligned_access instead
102 of SLOW_UNALIGNED_ACCESS.
103 * expmed.c (simple_mem_bitfield_p): Likewise.
104 * expr.c (alignment_for_piecewise_move): Likewise.
105 (emit_group_load_1): Likewise.
106 (emit_group_store): Likewise.
107 (copy_blkmode_from_reg): Likewise.
108 (emit_push_insn): Likewise.
109 (expand_assignment): Likewise.
110 (store_field): Likewise.
111 (expand_expr_real_1): Likewise.
112 * gimple-fold.c (gimple_fold_builtin_memory_op): Likewise.
113 * lra-constraints.c (simplify_operand_subreg): Likewise.
114 * stor-layout.c (bit_field_mode_iterator::next_mode): Likewise.
115 * gimple-ssa-store-merging.c: Likewise in block comment at start
117 * tree-ssa-strlen.c: Include target.h.
118 (handle_builtin_memcmp): Use targetm.slow_unaligned_access instead
119 of SLOW_UNALIGNED_ACCESS.
120 * system.h (SLOW_UNALIGNED_ACCESS): Poison.
122 2017-09-12 Richard Sandiford <richard.sandiford@linaro.org>
124 PR rtl-optimization/82185
125 * expmed.c (emit_store_flag_int): Only test tem if it has been
128 2017-09-12 Richard Biener <rguenther@suse.de>
131 * match.pd ((FTYPE) N CMP CST): Fix typo.
133 2017-09-12 Simon Atanasyan <simon.atanasyan@imgtec.com>
135 * config/mips/mips.c (mips_attribute_table): Add 'short_call'
137 (mips_near_type_p): Add 'short_call' attribute as a synonym
139 * doc/extend.texi (short_call): Document new function attribute.
141 2017-09-12 Jakub Jelinek <jakub@redhat.com>
144 * c-common.c (sync_resolve_size): Instead of c_dialect_cxx ()
145 assertion check that in the condition.
146 (get_atomic_generic_size): Likewise. Before testing if parameter
147 has pointer type, if it has array type, call for C++
148 default_conversion to perform array-to-pointer conversion.
150 2017-09-12 Richard Biener <rguenther@suse.de>
152 * tree-vect-generic.c (expand_vector_operations_1): Do nothing
153 for operations we cannot scalarize.
155 2017-09-12 Aldy Hernandez <aldyh@redhat.com>
157 * tree-ssa-threadbackward.c (fsm_find_thread_path): Make GC
158 vectors heap vectors. Clean up comments.
159 Make visited_bbs a reference.
160 (profitable_jump_thread_path): Make GC
161 vectors heap vectors. Clean up comments.
163 (convert_and_register_jump_thread_path): Make GC vectors heap
165 (check_subpath_and_update_thread_path): Same. Clean up comments.
166 Make visited_bbs a reference.
167 (handle_phi): Abstract common code to to
168 register_jump_thread_path_if_profitable.
169 Rename VAR_BB to DEF_BB.
171 Make GC vectors heap vectors.
172 Make visited_bbs a reference.
173 (handle_assignment): Same.
174 (register_jump_thread_path_if_profitable): New.
175 (fsm_find_control_statement_thread_paths): Rename VAR_BB to
177 Make GC vectors heap vectors. Clean up comments.
178 Make visited_bbs a reference.
179 (find_jump_threads_backwards): Make visited_bbs live in the stack.
180 * tree-ssa-threadupdate.c (delete_jump_thread_path): Fix typo in
183 2017-09-11 Max Filippov <jcmvbkbc@gmail.com>
186 * config/xtensa/xtensa.c (xtensa_mem_offset): Check that both
187 words of E_DImode object are reachable by xtensa_uimm8x4 access.
189 2017-09-11 Vidya Praveen <vidyapraveen@arm.com>
191 Revert r251800 and r251799.
193 2017-09-11 Martin Jambor <mjambor@suse.cz>
196 * hsa-gen.c (gen_hsa_phi_from_gimple_phi): Process ADDR_EXPRs in
197 arguments in advance.
198 * hsa-regalloc.c (naive_process_phi): New parameter predecessors,
199 use it to find predecessor edges.
200 (naive_outof_ssa): Collect vector of predecessors.
202 2017-09-08 Jason Merrill <jason@redhat.com>
204 PR c++/70029 - ICE with ref-qualifier and -flto
205 * langhooks.h (struct lang_hooks_for_types): Add
206 copy_lang_qualifiers.
207 * attribs.c (build_type_attribute_qual_variant): Use it.
208 * langhooks-def.h (LANG_HOOKS_COPY_LANG_QUALIFIERS): Default to
210 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Use it.
211 * tree.c (verify_type): Re-enable TYPE_CANONICAL main variant check.
213 2017-09-08 Eric Botcazou <ebotcazou@adacore.com>
216 * config/sparc/sparc.md (mulsi3): Rename into *mulsi3_sp32.
217 (*mulsi3_sp64): New instruction.
218 (mulsi3): New expander.
220 2017-09-08 Uros Bizjak <ubizjak@gmail.com>
222 * config/alpha/alpha.c (alpha_print_operand) <case 'S'>: Remove.
224 2017-09-08 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
226 * sancov.c: Include memmodel.h.
228 2017-09-07 Eric Botcazou <ebotcazou@adacore.com>
231 * config/sparc/sparc.c (sparc_emit_set_symbolic_const64): Deal with too
234 2017-09-07 Carl Love <cel@us.ibm.com>
236 * config/rs6000/vsx.md (define_insn "*stxvl"): Add missing argument to
237 the sldi instruction.
239 2017-09-07 David Edelsohn <dje.gcc@gmail.com>
241 * sancov.c: Include tm_p.h.
243 2017-09-07 Jakub Jelinek <jakub@redhat.com>
246 * output.h (switch_to_other_text_partition): New declaration.
247 * varasm.c (switch_to_other_text_partition): New function.
248 * config/rs6000/rs6000.c (uses_TOC): Return 2 if
249 NOTE_INSN_SWITCH_TEXT_SECTIONS is seen before finding load_toc_* insn.
250 (rs6000_elf_declare_function_name): If uses_TOC returned 2, switch
251 to the other text partition before emitting LCL label and switch back
252 after emitting the word after it.
254 2017-09-07 Richard Biener <rguenther@suse.de>
256 * passes.def (pass_split_crit_edges): Remove instance before PRE.
257 * tree-ssa-pre.c (pass_pre::execute): Instead manually split
258 critical edges here, after loop init.
259 (pass_data_pre): Remove PROP_no_crit_edges flags.
260 * tree-ssa-sccvn.c (vn_reference_lookup_3): Use vn_valueize
261 for valueization of call args to avoid leaking VN_TOP.
262 (visit_use): Assert we do not visit default defs.
263 (init_scc_vn): Use build_decl for VN_TOP to make name nicer.
264 Use error_mark_node to more easily detect leaking VN_TOP.
265 All default-defs are varying, not VN_TOP. Mark them visited.
266 (run_scc_vn): Make code match comment.
268 2017-09-07 Michael Meissner <meissner@linux.vnet.ibm.com>
270 * config/rs6000/rs6000-cpus.def (OTHER_VSX_VECTOR_MASKS): Delete
271 OPTION_MASK_FLOAT128_KEYWORD.
272 (POWERPC_MASKS): Likewise.
273 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Delete
274 support for the -mfloat128-type option, and make -mfloat128
275 default on PowerPC Linux systems. Define or undefine
276 __FLOAT128__ and __FLOAT128_HARDWARE__ for the current options.
277 Define __float128 to be __ieee128 if IEEE 128-bit support is
278 enabled, or undefine it.
279 (rs6000_cpu_cpp_builtins): Delete defining __FLOAT128__ here.
280 Delete defining __FLOAT128_TYPE__.
281 * config/rs6000/rs6000.opt (x_TARGET_FLOAT128_TYPE): Delete the
282 -mfloat128-type option and make -mfloat128 default on PowerPC
284 (TARGET_FLOAT128_TYPE): Likewise.
285 (-mfloat128-type): Likewise.
286 * config/rs6000/rs6000.c (rs6000_option_override_internal):
287 Delete the -mfloat128-type option and make -mfloat128 default on
288 PowerPC Linux systems. Always use __ieee128 to be the keyword for
289 the IEEE 128-bit type, and map __float128 to __ieee128 if IEEE
290 128-bit floating point is enabled. Change tests from using
291 -mfloat128-type to -mfloat128.
292 (rs6000_mangle_type): Use the correct mangling for the __float128
293 type even if normal long double is restricted to 64-bits.
294 (floatn_mode): Enable the _Float128 type by default on VSX Linux
296 * config/rs6000/rs6000.h (MASK_FLOAT128_TYPE): Delete.
297 (MASK_FLOAT128_KEYWORD): Define new shortcut macro.
298 (RS6000BTM_FLOAT128): Define in terms of -mfloat128, not
300 * doc/invoke.texi (RS/6000 and PowerPC Options): Update
301 documentation for -mfloat128.
303 2017-09-06 Olivier Hainque <hainque@adacore.com>
305 * config.gcc (powerpc-wrs-vxworksspe): Now match as vxworks*spe.
307 2017-09-06 Wish Wu <wishwu007@gmail.com>
308 Jakub Jelinek <jakub@redhat.com>
310 * asan.c (initialize_sanitizer_builtins): Add
311 BT_FN_VOID_UINT8_UINT8, BT_FN_VOID_UINT16_UINT16,
312 BT_FN_VOID_UINT32_UINT32, BT_FN_VOID_UINT64_UINT64,
313 BT_FN_VOID_FLOAT_FLOAT, BT_FN_VOID_DOUBLE_DOUBLE and
314 BT_FN_VOID_UINT64_PTR variables.
315 * builtin-types.def (BT_FN_VOID_UINT8_UINT8): New fn type.
316 (BT_FN_VOID_UINT16_UINT16): Likewise.
317 (BT_FN_VOID_UINT32_UINT32): Likewise.
318 (BT_FN_VOID_FLOAT_FLOAT): Likewise.
319 (BT_FN_VOID_DOUBLE_DOUBLE): Likewise.
320 (BT_FN_VOID_UINT64_PTR): Likewise.
321 * common.opt (flag_sanitize_coverage): New variable.
322 (fsanitize-coverage=trace-pc): Remove.
323 (fsanitize-coverage=): Add.
324 * flag-types.h (enum sanitize_coverage_code): New enum.
325 * fold-const.c (fold_range_test): Disable non-short-circuit
326 optimization if flag_sanitize_coverage.
327 (fold_truth_andor): Likewise.
328 * tree-ssa-ifcombine.c (ifcombine_ifandif): Likewise.
329 * opts.c (COVERAGE_SANITIZER_OPT): Define.
330 (coverage_sanitizer_opts): New array.
331 (get_closest_sanitizer_option): Add OPTS argument, handle also
332 OPT_fsanitize_coverage_.
333 (parse_sanitizer_options): Adjusted to also handle
334 OPT_fsanitize_coverage_.
335 (common_handle_option): Add OPT_fsanitize_coverage_.
336 * sancov.c (instrument_comparison, instrument_switch): New function.
337 (sancov_pass): Add trace-cmp support.
338 * sanitizer.def (BUILT_IN_SANITIZER_COV_TRACE_CMP1,
339 BUILT_IN_SANITIZER_COV_TRACE_CMP2, BUILT_IN_SANITIZER_COV_TRACE_CMP4,
340 BUILT_IN_SANITIZER_COV_TRACE_CMP8,
341 BUILT_IN_SANITIZER_COV_TRACE_CONST_CMP1,
342 BUILT_IN_SANITIZER_COV_TRACE_CONST_CMP2,
343 BUILT_IN_SANITIZER_COV_TRACE_CONST_CMP4,
344 BUILT_IN_SANITIZER_COV_TRACE_CONST_CMP8,
345 BUILT_IN_SANITIZER_COV_TRACE_CMPF, BUILT_IN_SANITIZER_COV_TRACE_CMPD,
346 BUILT_IN_SANITIZER_COV_TRACE_SWITCH): New builtins.
347 * doc/invoke.texi: Document -fsanitize-coverage=trace-cmp.
349 2017-09-06 Richard Earnshaw <rearnsha@arm.com>
351 * config/arm/parsecpu.awk (fatal): Note that we've encountered an
352 error. Only quit immediately if parsing is complete.
353 (BEGIN): Initialize fatal_err and parse_done.
354 (begin fpu, end fpu): Check number of arguments.
355 (begin arch, end arch): Likewise.
356 (begin cpu, end cpu): Likewise.
357 (cname, tune for, tune flags, architecture, fpu, option): Likewise.
358 (optalias): Likewise.
360 2017-09-06 Richard Earnshaw <rearnsha@arm.com>
362 * config.gcc (arm*-*-*): Don't add arm-isa.h to tm_p_file.
363 * config/arm/arm-isa.h: Delete. Move definitions to ...
364 * arm-cpus.in: ... here. Use new feature and fgroup values.
365 * config/arm/arm.c (arm_option_override): Use lower case for feature
367 * config/arm/arm.h (TARGET_HARD_FLOAT): Likewise.
368 (TARGET_VFP3, TARGET_VFP5, TARGET_FMA): Likewise.
369 * config/arm/parsecpu.awk (END): Add new command 'isa'.
371 (print_isa_bits_for): New function.
372 (gen_isa): New function.
373 (gen_comm_data): Use print_isa_bits_for.
374 (define feature): New keyword.
375 (define fgroup): New keyword.
376 * config/arm/t-arm (OPTIONS_H_EXTRA): Add arm-isa.h
377 (arm-isa.h): Add rule to generate file.
378 * common/config/arm/arm-common.c: (arm_canon_arch_option): Use lower
379 case for feature bit names.
381 2017-09-06 Richard Biener <rguenther@suse.de>
383 * tree-ssa-pre.c (NECESSARY): Remove.
384 (create_expression_by_pieces): Do not touch pass-local flags.
385 (insert_into_preds_of_block): Likewise.
386 (do_pre_regular_insertion): Likewise.
387 (eliminate_insert): Likewise.
388 (eliminate_dom_walker::before_dom_children): Likewise.
389 (fini_eliminate): Do not look at inserted_exprs.
390 (mark_operand_necessary): Remove.
391 (remove_dead_inserted_code): Replace with simple work-list
392 algorithm based on inserted_exprs and SSA uses.
393 (pass_pre::execute): Re-order fini_eliminate and
394 remove_dead_inserted_code.
396 2017-09-06 Olivier Hainque <hainque@adacore.com>
398 * config/powerpcspe/vxworks.h (VXCPU_FOR_8548): Correct definition
399 for VxWorks 7. Adjust surrounding comments.
401 2017-09-06 Richard Biener <rguenther@suse.de>
403 * gimple-ssa-strength-reduction.c
404 (find_candidates_dom_walker::before_dom_children): Also allow
407 2017-09-06 Richard Biener <rguenther@suse.de>
409 PR tree-optimization/82108
410 * tree-vect-stmts.c (vectorizable_load): Fix pointer adjustment
411 for gap in the non-permutation SLP case.
413 2017-09-06 Martin Jambor <mjambor@suse.cz>
415 PR tree-optimization/82078
416 * tree-sra.c (sort_and_splice_var_accesses): Move call to
417 add_access_to_work_queue...
418 (build_accesses_from_assign): ...here.
419 (propagate_all_subaccesses): Make sure racc is the group
420 representative, if there is one.
422 2017-09-06 Jakub Jelinek <jakub@redhat.com>
425 * varasm.c (categorize_decl_for_section): Use SECCAT_TBSS for TLS vars with
428 2017-09-06 Richard Biener <rguenther@suse.de>
430 * gimple-ssa-strength-reduction.c
431 (find_candidates_dom_walker::before_doom_children): Use a
432 type and not a mode check.
434 2017-09-06 Bernd Edlinger <bernd.edlinger@hotmail.de>
437 * config/arm/predicates.md (arm_general_adddi_operand): Create new
439 * config/arm/arm.md (*arm_adddi3, *arm_subdi3): Use new predicates.
441 2017-09-05 Jeff Law <law@redhat.com>
443 PR tree-optimization/64910
444 * tree-ssa-reassoc.c (reassociate_bb): Restrict last change to
445 cases where we have 3 or more operands.
447 2017-09-05 Jakub Jelinek <jakub@redhat.com>
450 * omp-low.c (lower_omp_for): Recompute tree invariant if
451 gimple_omp_for_initial/final is ADDR_EXPR.
454 * omp-expand.c (expand_omp_simd): Force second operands of COND_EXPR
455 into gimple val before gimplification fo the COND_EXPR.
457 2017-09-05 Aldy Hernandez <aldyh@redhat.com>
459 * tree-ssa-threadupdate.c (duplicate_thread_path): Remove unused
460 REGION_COPY argument.
461 (thread_through_all_blocks): Remove unused argument to
462 duplicate_thread_path.
464 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
465 Alan Hayward <alan.hayward@arm.com>
466 David Sherwood <david.sherwood@arm.com>
468 * config/aarch64/aarch64-protos.h (aarch64_gen_adjusted_ldpstp):
469 Take a scalar_mode rather than a machine_mode.
470 (aarch64_operands_adjust_ok_for_ldpstp): Likewise.
471 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Likewise.
472 (aarch64_operands_adjust_ok_for_ldpstp): Likewise.
473 (aarch64_gen_adjusted_ldpstp): Likewise.
474 (aarch64_expand_vector_init): Use scalar_mode instead of machine_mode.
476 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
477 Alan Hayward <alan.hayward@arm.com>
478 David Sherwood <david.sherwood@arm.com>
480 * config/aarch64/aarch64-protos.h (aarch64_is_extend_from_extract):
481 Take a scalar_int_mode instead of a machine_mode.
482 (aarch64_mask_and_shift_for_ubfiz_p): Likewise.
483 (aarch64_output_scalar_simd_mov_immediate): Likewise.
484 (aarch64_simd_scalar_immediate_valid_for_move): Likewise.
485 (aarch64_simd_attr_length_rglist): Delete.
486 * config/aarch64/aarch64.c (aarch64_is_extend_from_extract): Take
487 a scalar_int_mode instead of a machine_mode.
488 (aarch64_add_offset): Likewise.
489 (aarch64_internal_mov_immediate): Likewise
490 (aarch64_add_constant_internal): Likewise.
491 (aarch64_add_constant): Likewise.
492 (aarch64_movw_imm): Likewise.
493 (aarch64_rtx_arith_op_extract_p): Likewise.
494 (aarch64_mask_and_shift_for_ubfiz_p): Likewise.
495 (aarch64_simd_scalar_immediate_valid_for_move): Likewise.
496 Remove assert that the mode isn't a vector.
497 (aarch64_output_scalar_simd_mov_immediate): Likewise.
498 (aarch64_expand_mov_immediate): Update calls after above changes.
499 (aarch64_output_casesi): Use as_a <scalar_int_mode>.
500 (aarch64_and_bitmask_imm): Check for scalar integer modes.
501 (aarch64_move_imm): Likewise.
502 (aarch64_can_const_movi_rtx_p): Likewise.
503 (aarch64_strip_extend): Likewise.
504 (aarch64_extr_rtx_p): Likewise.
505 (aarch64_rtx_costs): Likewise, using wode_mode as the mode of
506 a CONST_INT when the mode parameter is VOIDmode.
507 (aarch64_float_const_rtx_p): Use scalar_int_mode for a temporary.
509 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
511 * machmode.h (bitwise_mode_for_mode): Return opt_mode.
512 * stor-layout.c (bitwise_mode_for_mode): Likewise.
513 (bitwise_type_for_mode): Update accordingly.
515 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
517 * stor-layout.h (mode_for_size_tree): Return an opt_mode.
518 * stor-layout.c (mode_for_size_tree): Likewise.
519 (mode_for_array): Update accordingly.
520 (layout_decl): Likewise.
521 (compute_record_mode): Likewise. Only set the mode once.
523 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
525 * target.def (get_mask_mode): Change return type to opt_mode.
527 * doc/tm.texi: Regenerate.
528 * targhooks.h (default_get_mask_mode): Return an opt_mode.
529 * targhooks.c (default_get_mask_mode): Likewise.
530 * config/i386/i386.c (ix86_get_mask_mode): Likewise.
531 * optabs-query.c (can_vec_mask_load_store_p): Update use of
532 targetm.get_mask_mode.
533 * tree.c (build_truth_vector_type): Likewise.
535 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
537 * machmode.h (mode_for_vector): Return an opt_mode.
538 * stor-layout.c (mode_for_vector): Likewise.
539 (mode_for_int_vector): Update accordingly.
540 (layout_type): Likewise.
541 * config/i386/i386.c (emit_memmov): Likewise.
542 (ix86_expand_set_or_movmem): Likewise.
543 (ix86_expand_vector_init): Likewise.
544 (ix86_get_mask_mode): Likewise.
545 * config/powerpcspe/powerpcspe.c (rs6000_expand_vec_perm_const_1):
547 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Likewise.
548 * expmed.c (extract_bit_field_1): Likewise.
549 * expr.c (expand_expr_real_2): Likewise.
550 * optabs-query.c (can_vec_perm_p): Likewise.
551 (can_vec_mask_load_store_p): Likewise.
552 * optabs.c (expand_vec_perm): Likewise.
553 * targhooks.c (default_get_mask_mode): Likewise.
554 * tree-vect-stmts.c (vectorizable_store): Likewise.
555 (vectorizable_load): Likewise.
556 (get_vectype_for_scalar_type_and_size): Likewise.
558 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
560 * machmode.h (mode_for_int_vector): New function.
561 * stor-layout.c (mode_for_int_vector): Likewise.
562 * config/aarch64/aarch64.c (aarch64_emit_approx_sqrt): Use it.
563 * config/powerpcspe/powerpcspe.c (rs6000_do_expand_vec_perm): Likewise.
564 * config/rs6000/rs6000.c (rs6000_do_expand_vec_perm): Likewise.
565 * config/s390/s390.c (s390_expand_vec_compare_cc): Likewise.
566 (s390_expand_vcond): Likewise.
568 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
570 * machmode.h (opt_machine_mode): New type.
571 (opt_mode<T>): Allow construction from anything that can be
573 (is_a, as_a, dyn_cast): Add overloads for opt_mode.
574 (mode_for_size): Return an opt_machine_mode.
575 * stor-layout.c (mode_for_size): Likewise.
576 (mode_for_size_tree): Update call accordingly.
577 (bitwise_mode_for_mode): Likewise.
578 (make_fract_type): Likewise.
579 (make_accum_type): Likewise.
580 * caller-save.c (replace_reg_with_saved_mem): Update call
582 * config/alpha/alpha.h (SECONDARY_MEMORY_NEEDED_MODE): Likewise.
583 * config/i386/i386.h (SECONDARY_MEMORY_NEEDED_MODE): Likewise.
584 * config/s390/s390.h (SECONDARY_MEMORY_NEEDED_MODE): Likewise.
585 * config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED_MODE): Likewise.
586 * expmed.c (extract_bit_field_1): Likewise.
587 * reload.c (get_secondary_mem): Likewise.
588 * varasm.c (assemble_integer): Likewise.
589 * lower-subreg.c (simplify_subreg_concatn): Likewise. Move
592 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
594 * machmode.h (decimal_float_mode_for_size): New function.
595 * real.h (REAL_VALUE_TO_TARGET_LONG_DOUBLE): Use float_mode_for_size.
596 (REAL_VALUE_TO_TARGET_DOUBLE): Likewise.
597 (REAL_VALUE_TO_TARGET_SINGLE): Likewise.
598 (REAL_VALUE_TO_TARGET_DECIMAL128): Use decimal_float_mode_for_size.
599 (REAL_VALUE_TO_TARGET_DECIMAL64): Likewise.
600 (REAL_VALUE_TO_TARGET_DECIMAL32): Likewise.
602 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
604 * builtins.c (expand_builtin_powi): Use int_mode_for_size.
605 (get_builtin_sync_mode): Likewise.
606 (expand_ifn_atomic_compare_exchange): Likewise.
607 (expand_builtin_atomic_clear): Likewise.
608 (expand_builtin_atomic_test_and_set): Likewise.
609 (fold_builtin_atomic_always_lock_free): Likewise.
610 * calls.c (compute_argument_addresses): Likewise.
611 (emit_library_call_value_1): Likewise.
612 (store_one_arg): Likewise.
613 * combine.c (combine_instructions): Likewise.
614 * config/aarch64/aarch64.c (aarch64_function_value): Likewise.
615 * config/arm/arm.c (arm_function_value): Likewise.
616 (aapcs_allocate_return_reg): Likewise.
617 * config/c6x/c6x.c (c6x_expand_movmem): Likewise.
618 * config/i386/i386.c (construct_container): Likewise.
619 (ix86_gimplify_va_arg): Likewise.
620 (ix86_expand_sse_cmp): Likewise.
621 (emit_memmov): Likewise.
622 (emit_memset): Likewise.
623 (expand_small_movmem_or_setmem): Likewise.
624 (ix86_expand_pextr): Likewise.
625 (ix86_expand_pinsr): Likewise.
626 * config/lm32/lm32.c (lm32_block_move_inline): Likewise.
627 * config/microblaze/microblaze.c (microblaze_block_move_straight):
629 * config/mips/mips.c (mips_function_value_1) Likewise.
630 (mips_block_move_straight): Likewise.
631 (mips_expand_ins_as_unaligned_store): Likewise.
632 * config/powerpcspe/powerpcspe.c
633 (rs6000_darwin64_record_arg_advance_flush): Likewise.
634 (rs6000_darwin64_record_arg_flush): Likewise.
635 * config/rs6000/rs6000.c
636 (rs6000_darwin64_record_arg_advance_flush): Likewise.
637 (rs6000_darwin64_record_arg_flush): Likewise.
638 * config/sparc/sparc.c (sparc_function_arg_1): Likewise.
639 (sparc_function_value_1): Likewise.
640 * config/spu/spu.c (adjust_operand): Likewise.
641 (spu_emit_branch_or_set): Likewise.
642 (arith_immediate_p): Likewise.
643 * emit-rtl.c (gen_lowpart_common): Likewise.
644 * expr.c (expand_expr_real_1): Likewise.
645 * function.c (assign_parm_setup_block): Likewise.
646 * gimple-ssa-store-merging.c (encode_tree_to_bitpos): Likewise.
647 * reload1.c (alter_reg): Likewise.
648 * stor-layout.c (mode_for_vector): Likewise.
649 (layout_type): Likewise.
651 2017-09-05 Richard Sandiford <richard.sandiford@linaro.org>
653 * config/spu/spu.c (exp2_immediate_p): Use int_mode_for_mode.
654 (spu_convert_move): Likewise.
655 * lower-subreg.c (resolve_simple_move): Likewise.
657 2017-09-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
660 * config/rs6000/altivec.md (altivec_vsum2sws): Convert from a
661 define_insn to a define_expand.
662 (altivec_vsum2sws_direct): New define_insn.
663 (altivec_vsumsws): Convert from a define_insn to a define_expand.
665 2017-09-05 Wilco Dijkstra <wdijkstr@arm.com>
667 * config/arm/arm.c (arm_option_params_internal): Improve setting of
670 2017-09-05 H.J. Lu <hongjiu.lu@intel.com>
675 * config/i386/i386.c (ix86_finalize_stack_frame_flags): Don't
676 realign stack if stack alignment needed is less than incoming
679 2017-09-05 Marek Polacek <polacek@redhat.com>
682 * convert.c (convert_to_integer_1) <case NEGATE_EXPR>: Move the ubsan
685 2017-09-05 Wilco Dijkstra <wdijkstr@arm.com>
687 * explow.c (get_dynamic_stack_size): Improve dynamic alignment.
689 2017-09-05 Richard Biener <rguenther@suse.de>
691 PR tree-optimization/82084
692 * fold-const.c (can_native_encode_string_p): Handle wide characters.
694 2017-09-05 Richard Biener <rguenther@suse.de>
696 PR tree-optimization/82102
697 * tree-ssa-pre.c (fini_eliminate): Check if lhs is NULL.
699 2017-09-05 Martin Liska <mliska@suse.cz>
701 PR tree-optimization/82032
702 * tree-cfg.c (generate_range_test): New function.
703 * tree-cfg.h (generate_range_test): Declared here.
704 * tree-cfgcleanup.c (convert_single_case_switch): New function.
705 (cleanup_control_expr_graph): Use it.
706 * tree-switch-conversion.c (try_switch_expansion): Remove
708 (emit_case_nodes): Use generate_range_test.
710 2017-09-04 Uros Bizjak <ubizjak@gmail.com>
713 * config/i386/i386.md (*<btsc><mode>_mask): Add
714 TARGET_USE_BT to insn constraint.
715 (*btr<mode>_mask): Ditto.
717 2017-09-04 Wilco Dijkstra <wdijkstr@arm.com>
719 * config/arm/arm.c (arm_legitimate_index_p): Add comment.
720 (thumb2_legitimate_index_p): Use correct range for DI/DF mode.
722 2017-09-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
725 * config/arm/arm.md (*arm_adddi3, *arm_subdi3): Split early except for
726 TARGET_NEON and TARGET_IWMMXT.
727 (anddi3, iordi3, xordi3, one_cmpldi2): Split while expanding except for
728 TARGET_NEON and TARGET_IWMMXT.
729 (*one_cmpldi2_insn): Moved the body of one_cmpldi2 here.
731 2017-09-04 Uros Bizjak <ubizjak@gmail.com>
733 * config/i386/i386-protos.h (ix86_tls_address_pattern_p) New prototype.
734 (ix86_rewrite_tls_address): Ditto.
735 * config/i386/i386.c (ix86_tls_address_pattern_p) New function.
736 (ix86_rewrite_tls_address_1): Ditto.
737 (ix86_rewrite_tls_address): Ditto.
738 * config/i386/predicates.md (tls_address_pattern): New predicate.
739 * config/i386/i386.md (TLS address splitter): New splitter.
741 2017-09-04 Richard Biener <rguenther@suse.de>
743 PR tree-optimization/82084
744 * fold-const.h (can_native_encode_string_p): Declare.
745 * fold-const.c (can_native_encode_string_p): Factor out from ...
746 (native_encode_string): ... here.
747 * tree-vect-stmts.c (vectorizable_store): Call it to avoid
748 vectorizing stores from constants we later cannot handle.
750 2017-09-04 Marek Polacek <polacek@redhat.com>
753 * doc/invoke.texi: Update -Wtautological-compare documentation.
755 2017-09-04 Jeff Law <law@redhat.com>
757 PR tree-optimization/64910
758 * tree-ssa-reassoc.c (reassociate_bb): For bitwise binary ops,
759 swap the first and last operand if the last is a constant.
761 2017-09-04 Marek Polacek <polacek@redhat.com>
764 * convert.c (do_narrow): When sanitizing signed integer overflows,
765 bail out for signed types.
766 (convert_to_integer_1) <case NEGATE_EXPR>: Likewise.
768 2017-09-04 Richard Biener <rguenther@suse.de>
770 PR tree-optimization/82060
771 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
772 Move devirtualization after stmt folding and before EH/AB/noreturn
773 cleanup to get the stmt refs canonicalized. Use a bool instead
774 of gimple_modified_p since that doesn't work for NOPs. Schedule
775 NOPs generated by folding for removal.
777 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
778 Alan Hayward <alan.hayward@arm.com>
779 David Sherwood <david.sherwood@arm.com>
781 * coretypes.h (pad_direction): New enum.
782 * defaults.h (DEFAULT_FUNCTION_ARG_PADDING): Delete.
783 (FUNCTION_ARG_PADDING): Likewise.
784 * target.def (function_arg_padding): New hook.
785 * targhooks.h (default_function_arg_padding): Declare.
786 * targhooks.c (default_function_arg_padding): New function.
787 * doc/tm.texi.in (FUNCTION_ARG_PADDING): Replace with...
788 (TARGET_FUNCTION_ARG_PADDING): ...this.
789 * doc/tm.texi: Regenerate.
790 * calls.c (store_unaligned_arguments_into_pseudos): Use pad_direction
791 instead of direction.
792 (compute_argument_addresses): Likewise.
793 (load_register_parameters): Likewise.
794 (emit_library_call_value_1): Likewise.
795 (store_one_arg): Use targetm.calls.function_arg_padding instead
796 of FUNCTION_ARG_PADDING.
797 (must_pass_in_stack_var_size_or_pad): Likewise.
798 * expr.c (emit_group_load_1): Use pad_direction instead of direction.
799 (emit_group_store): Likewise.
800 (emit_single_push_insn_1): Use targetm.calls.function_arg_padding
801 instead of FUNCTION_ARG_PADDING.
802 (emit_push_insn): Likewise, and propagate enum change throughout
804 * function.h (direction): Delete.
805 (locate_and_pad_arg_data::where_pad): Use pad_direction instead
807 * function.c (assign_parm_find_stack_rtl): Likewise.
808 (assign_parm_setup_block_p): Likewise.
809 (assign_parm_setup_block): Likewise.
810 (gimplify_parameters): Likewise.
811 (locate_and_pad_parm): Use targetm.calls.function_arg_padding
812 instead of FUNCTION_ARG_PADDING, and propagate enum change throughout
814 * config/aarch64/aarch64.h (FUNCTION_ARG_PADDING): Delete.
815 (BLOCK_REG_PADDING): Use pad_direction instead of direction.
816 * config/aarch64/aarch64-protos.h (aarch64_pad_arg_upward): Delete.
817 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): Replace with...
818 (aarch64_function_arg_padding): ...this new function.
819 (aarch64_gimplify_va_arg_expr): Use pad_direction instead of direction.
820 (TARGET_FUNCTION_ARG_PADDING): Redefine.
821 * config/arm/arm.h (FUNCTION_ARG_PADDING): Delete.
822 (BLOCK_REG_PADDING): Use pad_direction instead of direction.
823 * config/arm/arm-protos.h (arm_pad_arg_upward): Delete.
824 * config/arm/arm.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
825 (arm_pad_arg_upward): Replace with...
826 (arm_function_arg_padding): ...this new function.
827 * config/c6x/c6x.h (BLOCK_REG_PADDING): Use pad_direction instead
829 * config/ia64/hpux.h (FUNCTION_ARG_PADDING): Delete.
830 * config/ia64/ia64-protos.h (ia64_hpux_function_arg_padding): Delete.
831 * config/ia64/ia64.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
832 (ia64_hpux_function_arg_padding): Replace with...
833 (ia64_function_arg_padding): ...this new function. Use pad_direction
834 instead of direction. Check for TARGET_HPUX.
835 * config/iq2000/iq2000.h (FUNCTION_ARG_PADDING): Delete.
836 * config/iq2000/iq2000.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
837 (iq2000_function_arg_padding): New function.
838 * config/mips/mips-protos.h (mips_pad_arg_upward): Delete.
839 * config/mips/mips.c (mips_pad_arg_upward): Replace with...
840 (mips_function_arg_padding): ...this new function.
841 (mips_pad_reg_upward): Update accordingly.
842 (TARGET_FUNCTION_ARG_PADDING): Redefine.
843 * config/mips/mips.h (PAD_VARARGS_DOWN): Use
844 targetm.calls.function_arg_padding.
845 (FUNCTION_ARG_PADDING): Delete.
846 (BLOCK_REG_PADDING): Use pad_direction instead of direction.
847 * config/nios2/nios2.h (FUNCTION_ARG_PADDING): Delete.
848 (PAD_VARARGS_DOWN): Use targetm.calls.function_arg_padding.
849 * config/nios2/nios2-protos.h (nios2_function_arg_padding): Delete.
850 (nios2_block_reg_padding): Return pad_direction instead of direction.
851 * config/nios2/nios2.c (nios2_block_reg_padding): Return pad_direction
852 instead of direction.
853 (nios2_function_arg_padding): Likewise. Make static.
854 (TARGET_FUNCTION_ARG_PADDING): Redefine.
855 * config/pa/pa.h (FUNCTION_ARG_PADDING): Delete.
856 (BLOCK_REG_PADDING): Use targetm.calls.function_arg_padding.
857 * config/pa/pa-protos.h (pa_function_arg_padding): Delete.
858 * config/pa/pa.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
859 (pa_function_arg_padding): Make static. Return pad_direction instead
861 * config/powerpcspe/powerpcspe.h (FUNCTION_ARG_PADDING): Delete.
862 (PAD_VARARGS_DOWN): Use targetm.calls.function_arg_padding.
863 * config/powerpcspe/aix.h (BLOCK_REG_PADDING): Use pad_direction
864 instead of direction. Use targetm.calls.function_arg_padding.
865 * config/powerpcspe/darwin.h (BLOCK_REG_PADDING): Likewise.
866 * config/powerpcspe/freebsd64.h (BLOCK_REG_PADDING): Likewise.
867 * config/powerpcspe/linux64.h (BLOCK_REG_PADDING): Likewise.
868 * config/powerpcspe/powerpcspe-protos.h (function_arg_padding): Delete.
869 * config/powerpcspe/powerpcspe.c (TARGET_FUNCTION_ARG_PADDING):
871 (function_arg_padding): Rename to...
872 (rs6000_function_arg_padding): ...this. Make static. Return
873 pad_direction instead of direction.
874 (rs6000_return_in_msb): Use rs6000_function_arg_padding.
875 * config/rs6000/rs6000.h (FUNCTION_ARG_PADDING): Delete.
876 (PAD_VARARGS_DOWN): Use targetm.calls.function_arg_padding.
877 * config/rs6000/aix.h (BLOCK_REG_PADDING): Use pad_direction
878 instead of direction. Use targetm.calls.function_arg_padding.
879 * config/rs6000/darwin.h (BLOCK_REG_PADDING): Likewise.
880 * config/rs6000/freebsd64.h (BLOCK_REG_PADDING): Likewise.
881 * config/rs6000/linux64.h (BLOCK_REG_PADDING): Likewise.
882 * config/rs6000/rs6000-protos.h (function_arg_padding): Delete.
883 * config/rs6000/rs6000.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
884 (function_arg_padding): Rename to...
885 (rs6000_function_arg_padding): ...this. Make static. Return
886 pad_direction instead of direction.
887 (rs6000_return_in_msb): Use rs6000_function_arg_padding.
888 * config/s390/s390.h (FUNCTION_ARG_PADDING): Delete.
889 * config/s390/s390.c (s390_function_arg_padding): New function.
890 (TARGET_FUNCTION_ARG_PADDING): Redefine.
891 * config/sparc/sparc.h (FUNCTION_ARG_PADDING): Delete.
892 * config/sparc/sparc-protos.h (function_arg_padding): Delete.
893 * config/sparc/sparc.c (TARGET_FUNCTION_ARG_PADDING): Redefine.
894 (function_arg_padding): Rename to...
895 (sparc_function_arg_padding): ...this. Make static. Return
896 pad_direction instead of direction.
897 * config/spu/spu.h (FUNCTION_ARG_PADDING): Delete.
898 * config/spu/spu.c (spu_function_arg_padding): New function.
899 (TARGET_FUNCTION_ARG_PADDING): Redefine.
900 * system.h (FUNCTION_ARG_PADDING): Poison.
902 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
903 Alan Hayward <alan.hayward@arm.com>
904 David Sherwood <david.sherwood@arm.com>
906 * target.def (modes_tieable_p): New hook.
907 * doc/tm.texi (MODES_TIEABLE_P): Replace with...
908 (TARGET_MODES_TIEABLE_P): ...this.
909 * doc/tm.texi.in: Regenerate.
910 * hooks.h (hook_bool_mode_mode_true): Declare.
911 * hooks.c (hook_bool_mode_mode_true): New function.
912 * combine.c (subst): Use targetm.modes_tieable_p instead of
914 * dse.c (find_shift_sequence): Likewise.
915 * expmed.c (extract_low_bits): Likewise.
916 * lower-subreg.c: Include target.h.
917 (find_decomposable_subregs): Use targetm.modes_tieable_p instead of
919 * rtlanal.c (rtx_cost): Likewise.
920 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Delete.
921 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): Delete.
922 * config/aarch64/aarch64.c (aarch64_modes_tieable_p): Make static.
923 (TARGET_MODES_TIEABLE_P): Redefine.
924 * config/alpha/alpha.h (MODES_TIEABLE_P): Delete.
925 * config/alpha/alpha.c (alpha_modes_tieable_p): New function.
926 (TARGET_MODES_TIEABLE_P): Redefine.
927 * config/arc/arc.h (MODES_TIEABLE_P): Delete.
928 * config/arc/arc.c (TARGET_MODES_TIEABLE_P): Redefine.
929 (arc_modes_tieable_p): New function.
930 * config/arm/arm.h (MODES_TIEABLE_P): Delete.
931 * config/arm/arm-protos.h (arm_modes_tieable_p): Delete.
932 * config/arm/arm.c (TARGET_MODES_TIEABLE_P): Redefine.
933 (arm_modes_tieable_p): Make static.
934 * config/avr/avr.h (MODES_TIEABLE_P): Delete.
935 * config/bfin/bfin.h (MODES_TIEABLE_P): Delete.
936 * config/bfin/bfin.c (bfin_modes_tieable_p): New function.
937 (TARGET_MODES_TIEABLE_P): Redefine.
938 * config/c6x/c6x.h (MODES_TIEABLE_P): Delete.
939 * config/c6x/c6x.c (c6x_modes_tieable_p): New function.
940 (TARGET_MODES_TIEABLE_P): Redefine.
941 * config/cr16/cr16.h (MODES_TIEABLE_P): Delete.
942 * config/cr16/cr16.c (TARGET_MODES_TIEABLE_P): Redefine.
943 (cr16_modes_tieable_p): New function.
944 * config/cris/cris.h (MODES_TIEABLE_P): Delete.
945 * config/epiphany/epiphany.h (MODES_TIEABLE_P): Delete.
946 * config/fr30/fr30.h (MODES_TIEABLE_P): Delete.
947 (TRULY_NOOP_TRUNCATION): Update comment.
948 * config/frv/frv.h (MODES_TIEABLE_P): Delete.
949 (TRULY_NOOP_TRUNCATION): Update comment.
950 * config/frv/frv.c (TARGET_MODES_TIEABLE_P): Redefine.
951 (frv_modes_tieable_p): New function.
952 * config/ft32/ft32.h (MODES_TIEABLE_P): Delete.
953 * config/h8300/h8300.h (MODES_TIEABLE_P): Delete.
954 * config/h8300/h8300.c (h8300_modes_tieable_p): New function.
955 (TARGET_MODES_TIEABLE_P): Redefine.
956 * config/i386/i386.h (MODES_TIEABLE_P): Delete.
957 * config/i386/i386-protos.h (ix86_modes_tieable_p): Delete.
958 * config/i386/i386.c (ix86_modes_tieable_p): Make static.
959 (TARGET_MODES_TIEABLE_P): Redefine.
960 * config/ia64/ia64.h (MODES_TIEABLE_P): Delete.
961 * config/ia64/ia64.c (TARGET_MODES_TIEABLE_P): Redefine.
962 (ia64_modes_tieable_p): New function.
963 * config/iq2000/iq2000.h (MODES_TIEABLE_P): Delete.
964 * config/iq2000/iq2000.c (TARGET_MODES_TIEABLE_P): Redefine.
965 (iq2000_modes_tieable_p): New function.
966 * config/lm32/lm32.h (MODES_TIEABLE_P): Delete.
967 * config/lm32/lm32.c (TARGET_MODES_TIEABLE_P): Redefine.
968 (lm32_modes_tieable_p): New function.
969 * config/m32c/m32c.h (MODES_TIEABLE_P): Delete.
970 * config/m32c/m32c-protos.h (m32c_modes_tieable_p): Delete.
971 * config/m32c/m32c.c (m32c_modes_tieable_p): Make static.
972 (TARGET_MODES_TIEABLE_P): Redefine.
973 * config/m32r/m32r.h (MODES_TIEABLE_P): Delete.
974 * config/m32r/m32r.c (TARGET_MODES_TIEABLE_P): Redefine.
975 (m32r_modes_tieable_p): New function.
976 * config/m68k/m68k.h (MODES_TIEABLE_P): Delete.
977 * config/m68k/m68k.c (TARGET_MODES_TIEABLE_P): Redefine.
978 (m68k_modes_tieable_p): New function.
979 * config/mcore/mcore.h (MODES_TIEABLE_P): Delete.
980 * config/mcore/mcore.c (TARGET_MODES_TIEABLE_P): Redefine.
981 (mcore_modes_tieable_p): New function.
982 * config/microblaze/microblaze.h (MODES_TIEABLE_P): Delete.
983 * config/microblaze/microblaze.c (microblaze_modes_tieable_p): New
985 (TARGET_MODES_TIEABLE_P): Redefine.
986 * config/mips/mips.h (MODES_TIEABLE_P): Delete.
987 * config/mips/mips-protos.h (mips_modes_tieable_p): Delete.
988 * config/mips/mips.c (mips_modes_tieable_p): Make static.
989 (TARGET_MODES_TIEABLE_P): Redefine.
990 * config/mmix/mmix.h (MODES_TIEABLE_P): Delete.
991 * config/mn10300/mn10300.h (MODES_TIEABLE_P): Delete.
992 * config/mn10300/mn10300-protos.h (mn10300_modes_tieable): Delete.
993 * config/mn10300/mn10300.c (mn10300_modes_tieable): Rename to...
994 (mn10300_modes_tieable_p): ...this and make static.
995 (TARGET_MODES_TIEABLE_P): Redefine.
996 * config/moxie/moxie.h (MODES_TIEABLE_P): Delete.
997 * config/msp430/msp430.h (MODES_TIEABLE_P): Delete.
998 * config/msp430/msp430-protos.h (msp430_modes_tieable_p): Delete.
999 * config/msp430/msp430.c (TARGET_MODES_TIEABLE_P): Redefine.
1000 (msp430_modes_tieable_p): Make static.
1001 * config/nds32/nds32.h (MODES_TIEABLE_P): Delete.
1002 * config/nds32/nds32.c (nds32_modes_tieable_p): New function.
1003 (TARGET_MODES_TIEABLE_P): Redefine.
1004 * config/nios2/nios2.h (MODES_TIEABLE_P): Delete.
1005 * config/nvptx/nvptx.h (MODES_TIEABLE_P): Delete.
1006 * config/nvptx/nvptx.c (nvptx_modes_tieable_p): New function.
1007 (TARGET_MODES_TIEABLE_P): Redefine.
1008 * config/pa/pa.h (MODES_TIEABLE_P): Delete.
1009 * config/pa/pa-protos.h (pa_modes_tieable_p): Delete.
1010 * config/pa/pa.c (pa_modes_tieable_p): Make static.
1011 (TARGET_MODES_TIEABLE_P): Redefine.
1012 * config/pdp11/pdp11.h (MODES_TIEABLE_P): Delete.
1013 * config/pdp11/pdp11.c (TARGET_MODES_TIEABLE_P): Redefine.
1014 (pdp11_modes_tieable_p): New function.
1015 * config/powerpcspe/powerpcspe.h (MODES_TIEABLE_P): Delete.
1016 * config/powerpcspe/powerpcspe.c (TARGET_MODES_TIEABLE_P): Redefine.
1017 (rs6000_modes_tieable_p): New function.
1018 (rs6000_debug_reg_global): Use it instead of MODES_TIEABLE_P.
1019 * config/powerpcspe/powerpcspe.md: Update comment.
1020 * config/riscv/riscv.h (MODES_TIEABLE_P): Delete.
1021 * config/riscv/riscv.c (riscv_modes_tieable_p): New function.
1022 (TARGET_MODES_TIEABLE_P): Redefine.
1023 * config/rl78/rl78.h (MODES_TIEABLE_P): Delete.
1024 * config/rl78/rl78.c (TARGET_MODES_TIEABLE_P): Redefine.
1025 (rl78_modes_tieable_p): New function.
1026 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Delete.
1027 * config/rs6000/rs6000.c (TARGET_MODES_TIEABLE_P): Redefine.
1028 (rs6000_modes_tieable_p): New function.
1029 (rs6000_debug_reg_global): Use it instead of MODES_TIEABLE_P.
1030 * config/rs6000/rs6000.md: Update comment.
1031 * config/rx/rx.h (MODES_TIEABLE_P): Delete.
1032 * config/rx/rx.c (rx_modes_tieable_p): New function.
1033 (TARGET_MODES_TIEABLE_P): Redefine.
1034 * config/s390/s390.h (MODES_TIEABLE_P): Delete.
1035 * config/s390/s390.c (s390_modes_tieable_p): New function.
1036 (TARGET_MODES_TIEABLE_P): Redefine.
1037 * config/sh/sh.h (MODES_TIEABLE_P): Delete.
1038 * config/sh/sh.c (TARGET_MODES_TIEABLE_P): Redefine.
1039 (sh_modes_tieable_p): New function.
1040 * config/sparc/sparc.h (MODES_TIEABLE_P): Delete.
1041 * config/sparc/sparc-protos.h (sparc_modes_tieable_p): Delete.
1042 * config/sparc/sparc.c (TARGET_MODES_TIEABLE_P): Redefine.
1043 (sparc_modes_tieable_p): Make static.
1044 * config/spu/spu.h (MODES_TIEABLE_P): Delete.
1045 * config/spu/spu.c (spu_modes_tieable_p): New function.
1046 (TARGET_MODES_TIEABLE_P): Redefine.
1047 * config/stormy16/stormy16.h (MODES_TIEABLE_P): Delete.
1048 * config/stormy16/stormy16.c (xstormy16_modes_tieable_p): New function.
1049 (TARGET_MODES_TIEABLE_P): Redefine.
1050 * config/tilegx/tilegx.h (MODES_TIEABLE_P): Delete.
1051 * config/tilepro/tilepro.h (MODES_TIEABLE_P): Delete.
1052 * config/v850/v850.h (MODES_TIEABLE_P): Delete.
1053 * config/v850/v850.c (v850_modes_tieable_p): New function.
1054 (TARGET_MODES_TIEABLE_P): Redefine.
1055 * config/vax/vax.h (MODES_TIEABLE_P): Delete.
1056 * config/visium/visium.h (MODES_TIEABLE_P): Delete.
1057 * config/visium/visium.c (TARGET_MODES_TIEABLE_P): Redefine.
1058 (visium_modes_tieable_p): New function.
1059 * config/xtensa/xtensa.h (MODES_TIEABLE_P): Delete.
1060 * config/xtensa/xtensa.c (TARGET_MODES_TIEABLE_P): Redefine.
1061 (xtensa_modes_tieable_p): New function.
1062 * system.h (MODES_TIEABLE_P): Poison.
1064 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
1065 Alan Hayward <alan.hayward@arm.com>
1066 David Sherwood <david.sherwood@arm.com>
1068 * target.def (hard_regno_mode_ok): New hook.
1069 * doc/tm.texi (HARD_REGNO_MODE_OK): Replace with...
1070 (TARGET_HARD_REGNO_MODE_OK): ...this.
1071 * doc/tm.texi.in: Regenerate.
1072 * hooks.h (hook_bool_uint_mode_true): Declare.
1073 * hooks.c (hook_bool_uint_mode_true): New function.
1074 * doc/md.texi: Refer to targetm.hard_regno_mode_ok instead of
1076 * genpreds.c (write_insn_preds_c): Add an include of target.h.
1077 * alias.c (init_alias_target): Use targetm.hard_regno_mode_ok
1078 instead of HARD_REGNO_MODE_OK.
1079 * caller-save.c: Include target.h.
1080 (reg_save_code): Use targetm.hard_regno_mode_ok instead of
1082 * combine.c (can_combine_p): Likewise.
1083 (combinable_i3pat): Likewise.
1084 (can_change_dest_mode): Likewise.
1085 * expr.c (init_expr_target): Likewise.
1086 (convert_move): Likewise.
1087 (convert_modes): Likewise.
1088 * ira.c (setup_prohibited_class_mode_regs): Likewise.
1089 (setup_prohibited_mode_move_regs): Likewise.
1090 * ira.h (target_ira): Likewise.
1091 * lra-assigns.c (find_hard_regno_for_1): Likewise.
1092 * lra-constraints.c (process_alt_operands): Likewise.
1093 (split_reg): Likewise.
1094 * recog.c (peep2_find_free_register): Likewise.
1095 * ree.c (combine_reaching_defs): Likewise.
1096 * regcprop.c (maybe_mode_change): Likewise.
1097 * reginfo.c (init_reg_sets_1): Likewise.
1098 (choose_hard_reg_mode): Likewise.
1099 (simplifiable_subregs): Likewise.
1100 * regrename.c (check_new_reg_p): Likewise.
1101 * reload.c (find_valid_class): Likewise.
1102 (find_valid_class_1): Likewise.
1103 (reload_inner_reg_of_subreg): Likewise.
1104 (push_reload): Likewise.
1105 (combine_reloads): Likewise.
1106 (find_dummy_reload): Likewise.
1107 (find_reloads): Likewise.
1108 * reload1.c (find_reg): Likewise.
1109 (set_reload_reg): Likewise.
1110 (allocate_reload_reg): Likewise.
1111 (choose_reload_regs): Likewise.
1112 (reload_adjust_reg_for_temp): Likewise.
1113 * rtlanal.c (subreg_size_offset_from_lsb): Likewise.
1114 (simplify_subreg_regno): Likewise.
1115 * sel-sched.c (init_regs_for_mode): Likewise.
1116 * varasm.c (make_decl_rtl): Likewise.
1117 * config/aarch64/aarch64.h (HARD_REGNO_MODE_OK): Delete.
1118 (MODES_TIEABLE_P): Use targetm.hard_regno_mode_ok instead of
1120 * config/aarch64/aarch64-protos.h (aarch64_hard_regno_mode_ok): Delete.
1121 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Make static.
1122 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1123 * config/alpha/alpha.h (HARD_REGNO_MODE_OK): Delete.
1124 * config/alpha/alpha.c (alpha_hard_regno_mode_ok): New function.
1125 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1126 * config/arc/arc.h (arc_hard_regno_mode_ok): Delete.
1127 (arc_mode_class): Delete.
1128 (HARD_REGNO_MODE_OK): Delete.
1129 * config/arc/arc.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1130 (arc_hard_regno_mode_ok): Rename old array to...
1131 (arc_hard_regno_mode_ok_modes): ...this.
1132 (arc_conditional_register_usage): Update accordingly.
1133 (arc_mode_class): Make static.
1134 (arc_hard_regno_mode_ok): New function.
1135 * config/arm/arm.h (HARD_REGNO_MODE_OK): Delete.
1136 * config/arm/arm-protos.h (arm_hard_regno_mode_ok): Delete.
1137 * config/arm/arm.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1138 (arm_hard_regno_mode_ok): Make static.
1139 * config/arm/arm.md (movdi): Use targetm.hard_regno_mode_ok instead of
1141 * config/avr/avr-protos.h (avr_hard_regno_mode_ok): Delete.
1142 * config/avr/avr.h (HARD_REGNO_MODE_OK): Delete.
1143 * config/avr/avr.c (avr_hard_regno_mode_ok): Make static and
1145 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1146 * config/bfin/bfin-protos.h (hard_regno_mode_ok): Delete.
1147 * config/bfin/bfin.h (HARD_REGNO_MODE_OK): Delete.
1148 * config/bfin/bfin.c (hard_regno_mode_ok): Rename to...
1149 (bfin_hard_regno_mode_ok): ...this. Make static and return a bool.
1150 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1151 * config/bfin/predicates.md (valid_reg_operand): Use
1152 targetm.hard_regno_mode_ok instead of HARD_REGNO_MODE_OK.
1153 * config/c6x/c6x.h (HARD_REGNO_MODE_OK): Delete.
1154 * config/c6x/c6x.c (c6x_hard_regno_mode_ok): New function.
1155 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1156 * config/cr16/cr16.h (HARD_REGNO_MODE_OK): Delete.
1157 * config/cr16/cr16-protos.h (cr16_hard_regno_mode_ok): Delete.
1158 * config/cr16/cr16.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1159 (cr16_hard_regno_mode_ok): Make static and return a bool.
1160 * config/cris/cris.h (HARD_REGNO_MODE_OK): Delete.
1161 * config/cris/cris.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1162 (cris_hard_regno_mode_ok): New function.
1163 * config/epiphany/epiphany.h (epiphany_hard_regno_mode_ok): Delete.
1164 (epiphany_mode_class): Delete.
1165 (HARD_REGNO_MODE_OK): Delete.
1166 * config/epiphany/epiphany-protos.h (hard_regno_mode_ok): Delete.
1167 * config/epiphany/epiphany.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1168 (hard_regno_mode_ok): Rename to...
1169 (epiphany_hard_regno_mode_ok): ...this. Make static and return a bool.
1170 * config/fr30/fr30.h (HARD_REGNO_MODE_OK): Delete.
1171 * config/fr30/fr30.md: Refer to targetm.hard_regno_mode_ok instead of
1173 * config/frv/frv.h (HARD_REGNO_MODE_OK): Delete.
1174 * config/frv/frv-protos.h (frv_hard_regno_mode_ok): Delete.
1175 * config/frv/frv.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1176 (frv_hard_regno_mode_ok): Make static and return a bool.
1177 * config/frv/frv.md: Refer to targetm.hard_regno_mode_ok instead of
1179 * config/ft32/ft32.h (HARD_REGNO_MODE_OK): Delete.
1180 * config/h8300/h8300.h (HARD_REGNO_MODE_OK): Delete.
1181 * config/h8300/h8300-protos.h (h8300_hard_regno_mode_ok): Delete.
1182 * config/h8300/h8300.c (h8300_hard_regno_mode_ok): Make static
1184 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1185 * config/i386/i386.h (HARD_REGNO_MODE_OK): Delete.
1186 * config/i386/i386-protos.h (ix86_hard_regno_mode_ok): Delete.
1187 * config/i386/i386.c (ix86_hard_regno_mode_ok): Make static and
1189 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1190 * config/ia64/ia64.h (HARD_REGNO_MODE_OK): Delete.
1191 * config/ia64/ia64.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1192 (ia64_hard_regno_mode_ok): New function.
1193 * config/iq2000/iq2000.h (HARD_REGNO_MODE_OK): Delete.
1194 * config/iq2000/iq2000.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1195 (iq2000_hard_regno_mode_ok): New function.
1196 * config/lm32/lm32.h (HARD_REGNO_MODE_OK): Delete.
1197 * config/lm32/lm32.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1198 (lm32_hard_regno_mode_ok): New function.
1199 * config/m32c/m32c.h (HARD_REGNO_MODE_OK): Delete.
1200 * config/m32c/m32c-protos.h (m32c_hard_regno_ok): Delete.
1201 * config/m32c/m32c.c (class_can_hold_mode): Use m32c_hard_regno_mode_ok
1202 instead of HARD_REGNO_MODE_OK.
1203 (m32c_hard_regno_ok): Rename to...
1204 (m32c_hard_regno_mode_ok): ...this. Make static and return a bool.
1205 (m32c_cannot_change_mode_class): Update accordingly.
1206 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1207 * config/m32r/m32r.h (m32r_hard_regno_mode_ok): Delete.
1208 (m32r_mode_class): Delete.
1209 (HARD_REGNO_MODE_OK): Delete.
1210 * config/m32r/m32r.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1211 (m32r_hard_regno_mode_ok): Rename to...
1212 (m32r_hard_regno_modes): ...this.
1213 (m32r_mode_class): Make static.
1214 (m32r_hard_regno_mode_ok): New function.
1215 * config/m68k/m68k.h (HARD_REGNO_MODE_OK): Delete.
1216 * config/m68k/m68k-protos.h (m68k_regno_mode_ok): Delete.
1217 * config/m68k/m68k.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1218 (m68k_hard_regno_mode_ok): Make static.
1219 * config/mcore/mcore.h (HARD_REGNO_MODE_OK): Delete.
1220 * config/mcore/mcore.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1221 (mcore_hard_regno_mode_ok): New function.
1222 * config/microblaze/microblaze.h (microblaze_hard_regno_mode_ok)
1223 (HARD_REGNO_MODE_OK): Delete.
1224 * config/microblaze/microblaze.c (microblaze_hard_regno_mode_ok):
1226 (microblaze_hard_regno_mode_ok_p): ...this and make static.
1227 (microblaze_hard_regno_mode_ok): New function.
1228 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1229 * config/mips/mips.h (HARD_REGNO_MODE_OK): Delete.
1230 (mips_hard_regno_mode_ok): Delete.
1231 * config/mips/mips.c (mips_hard_regno_mode_ok): Rename to...
1232 (mips_hard_regno_mode_ok_p): ...this and make static.
1233 (mips_hard_regno_mode_ok_p): Rename to...
1234 (mips_hard_regno_mode_ok_uncached): ...this.
1235 (mips_hard_regno_mode_ok): New function.
1236 (mips_class_max_nregs): Use mips_hard_regno_mode_ok instead
1237 of HARD_REGNO_MODE_OK.
1238 (mips_option_override): Update after above name changes.
1239 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1240 * config/mmix/mmix.h (HARD_REGNO_MODE_OK): Delete.
1241 * config/mn10300/mn10300.h (HARD_REGNO_MODE_OK): Delete.
1242 * config/mn10300/mn10300-protos.h (mn10300_hard_regno_mode_ok): Delete.
1243 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Make static.
1244 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1245 * config/moxie/moxie.h (HARD_REGNO_MODE_OK): Delete.
1246 * config/msp430/msp430.h (HARD_REGNO_MODE_OK): Delete.
1247 * config/msp430/msp430-protos.h (msp430_hard_regno_mode_ok): Delete.
1248 * config/msp430/msp430.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1249 (msp430_hard_regno_mode_ok): Make static and return a bool.
1250 * config/nds32/nds32.h (HARD_REGNO_MODE_OK): Delete.
1251 * config/nds32/nds32-protos.h (nds32_hard_regno_mode_ok): Delete.
1252 * config/nds32/nds32.c (nds32_hard_regno_mode_ok): Make static
1254 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1255 * config/nios2/nios2.h (HARD_REGNO_MODE_OK): Delete.
1256 * config/nvptx/nvptx.h (HARD_REGNO_MODE_OK): Delete.
1257 * config/pa/pa.h (MODES_TIEABLE_P): Update commentary.
1258 * config/pa/pa32-regs.h (HARD_REGNO_MODE_OK): Rename to...
1259 (PA_HARD_REGNO_MODE_OK): ...this
1260 * config/pa/pa64-regs.h (HARD_REGNO_MODE_OK): Rename to...
1261 (PA_HARD_REGNO_MODE_OK): ...this.
1262 * config/pa/pa.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1263 (pa_hard_regno_mode_ok): New function.
1264 * config/pdp11/pdp11.h (HARD_REGNO_MODE_OK): Delete.
1265 * config/pdp11/pdp11.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1266 (pdp11_hard_regno_mode_ok): New function.
1267 * config/powerpcspe/powerpcspe.h (HARD_REGNO_MODE_OK): Delete.
1268 * config/powerpcspe/powerpcspe-protos.h (rs6000_hard_regno_mode_ok_p):
1270 * config/powerpcspe/powerpcspe.c (rs6000_hard_regno_mode_ok_p):
1272 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1273 (rs6000_hard_regno_mode_ok): Rename to...
1274 (rs6000_hard_regno_mode_ok_uncached): ...this.
1275 (rs6000_init_hard_regno_mode_ok): Update accordingly.
1276 (rs6000_hard_regno_mode_ok): New function.
1277 * config/riscv/riscv.h (HARD_REGNO_MODE_OK): Delete.
1278 * config/riscv/riscv-protos.h (riscv_hard_regno_mode_ok_p): Delete.
1279 * config/riscv/riscv.c (riscv_hard_regno_mode_ok_p): Rename to...
1280 (riscv_hard_regno_mode_ok): ...this and make static.
1281 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1282 * config/rl78/rl78.h (HARD_REGNO_MODE_OK): Delete.
1283 * config/rl78/rl78-protos.h (rl78_hard_regno_mode_ok): Delete.
1284 * config/rl78/rl78.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1285 (rl78_hard_regno_mode_ok): Make static and return bool.
1286 * config/rs6000/rs6000.h (HARD_REGNO_MODE_OK): Delete.
1287 * config/rs6000/rs6000-protos.h (rs6000_hard_regno_mode_ok_p):
1289 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok_p): Make static.
1290 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1291 (rs6000_hard_regno_mode_ok): Rename to...
1292 (rs6000_hard_regno_mode_ok_uncached): ...this.
1293 (rs6000_init_hard_regno_mode_ok): Update accordingly.
1294 (rs6000_hard_regno_mode_ok): New function.
1295 * config/rx/rx.h (HARD_REGNO_MODE_OK): Delete.
1296 * config/rx/rx.c (rx_hard_regno_mode_ok): New function.
1297 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1298 * config/s390/s390.h (HARD_REGNO_MODE_OK): Delete.
1299 * config/s390/s390-protos.h (s390_hard_regno_mode_ok): Delete.
1300 * config/s390/s390.c (s390_hard_regno_mode_ok): Make static.
1301 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1302 * config/sh/sh.h (HARD_REGNO_MODE_OK): Delete.
1303 * config/sh/sh-protos.h (sh_hard_regno_mode_ok): Delete.
1304 * config/sh/sh.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1305 (sh_hard_regno_mode_ok): Make static.
1306 * config/sparc/constraints.md: Refer to targetm.hard_regno_mode_ok
1307 instead of HARD_REGNO_MODE_OK.
1308 * config/sparc/sparc.h (hard_regno_mode_classes): Delete.
1309 (sparc_mode_class): Delete.
1310 (HARD_REGNO_MODE_OK): Delete.
1311 * config/sparc/sparc.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1312 (hard_regno_mode_classes): Make static.
1313 (sparc_mode_class): Likewise.
1314 (sparc_hard_regno_mode_ok): New function.
1315 * config/spu/spu.h (HARD_REGNO_MODE_OK): Delete.
1316 * config/stormy16/stormy16.h (HARD_REGNO_MODE_OK): Delete.
1317 * config/stormy16/stormy16.c (xstormy16_hard_regno_mode_ok): New
1319 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1320 * config/tilegx/tilegx.h (HARD_REGNO_MODE_OK): Delete.
1321 * config/tilepro/tilepro.h (HARD_REGNO_MODE_OK): Delete.
1322 * config/v850/v850.h (HARD_REGNO_MODE_OK): Delete.
1323 * config/v850/v850.c (v850_hard_regno_mode_ok): New function.
1324 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1325 * config/vax/vax.h (HARD_REGNO_MODE_OK): Delete.
1326 * config/visium/visium.h (HARD_REGNO_MODE_OK): Delete.
1327 * config/visium/visium.c (TARGET_HARD_REGNO_MODE_OK): Redefine.
1328 (visium_hard_regno_mode_ok): New function.
1329 * config/visium/visium.md: Refer to targetm.hard_regno_mode_ok
1330 instead of HARD_REGNO_MODE_OK.
1331 * config/xtensa/xtensa.h (xtensa_hard_regno_mode_ok): Delete.
1332 (HARD_REGNO_MODE_OK): Delete.
1333 * config/xtensa/xtensa.c (xtensa_hard_regno_mode_ok): Rename to...
1334 (xtensa_hard_regno_mode_ok_p): ...this and make static.
1335 (xtensa_option_override): Update accordingly.
1336 (TARGET_HARD_REGNO_MODE_OK): Redefine.
1337 (xtensa_hard_regno_mode_ok): New function.
1338 * system.h (HARD_REGNO_MODE_OK): Poison.
1340 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
1341 Alan Hayward <alan.hayward@arm.com>
1342 David Sherwood <david.sherwood@arm.com>
1344 * target.def (hard_regno_call_part_clobbered): New hook.
1345 * doc/tm.texi.in (HARD_REGNO_CALL_PART_CLOBBERED): Replace with...
1346 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): ...this hook.
1347 * doc/tm.texi: Regenerate.
1348 * hooks.h (hook_bool_uint_mode_false): Declare.
1349 * hooks.c (hook_bool_uint_mode_false): New function.
1350 * regs.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1351 * cselib.c (cselib_process_insn): Use
1352 targetm.hard_regno_call_part_clobbered instead of
1353 HARD_REGNO_CALL_PART_CLOBBERED.
1354 * ira-conflicts.c (ira_build_conflicts): Likewise.
1355 * ira-costs.c (ira_tune_allocno_costs): Likewise.
1356 * lra-constraints.c (need_for_call_save_p): Likewise.
1357 * lra-lives.c: Include target.h.
1358 (check_pseudos_live_through_calls): Use
1359 targetm.hard_regno_call_part_clobbered instead of
1360 HARD_REGNO_CALL_PART_CLOBBERED.
1361 * regcprop.c: Include target.h.
1362 (copyprop_hardreg_forward_1): Use
1363 targetm.hard_regno_call_part_clobbered instead of
1364 HARD_REGNO_CALL_PART_CLOBBERED.
1365 * reginfo.c (choose_hard_reg_mode): Likewise.
1366 * regrename.c (check_new_reg_p): Likewise.
1367 * reload.c (find_equiv_reg): Likewise.
1368 * reload1.c (emit_reload_insns): Likewise.
1369 * sched-deps.c (deps_analyze_insn): Likewise.
1370 * sel-sched.c (init_regs_for_mode): Likewise.
1371 (mark_unavailable_hard_regs): Likewise.
1372 * targhooks.c (default_dwarf_frame_reg_mode): Likewise.
1373 * config/aarch64/aarch64.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1374 * config/aarch64/aarch64.c (aarch64_hard_regno_call_part_clobbered):
1376 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1377 * config/avr/avr.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1378 * config/avr/avr-protos.h (avr_hard_regno_call_part_clobbered):
1380 * config/avr/avr.c (avr_hard_regno_call_part_clobbered): Make static
1382 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1383 * config/i386/i386.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1384 * config/i386/i386.c (ix86_hard_regno_call_part_clobbered): New
1386 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1387 * config/mips/mips.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1388 * config/mips/mips.c (mips_hard_regno_call_part_clobbered): New
1390 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1391 * config/powerpcspe/powerpcspe.h (HARD_REGNO_CALL_PART_CLOBBERED):
1393 * config/powerpcspe/powerpcspe.c
1394 (rs6000_hard_regno_call_part_clobbered): New function.
1395 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1396 * config/rs6000/rs6000.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1397 * config/rs6000/rs6000.c (rs6000_hard_regno_call_part_clobbered):
1399 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1400 * config/s390/s390.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1401 * config/s390/s390.c (s390_hard_regno_call_part_clobbered): New
1403 (TARGET_HARD_REGNO_CALL_PART_CLOBBERED): Redefine.
1404 * config/sh/sh.h (HARD_REGNO_CALL_PART_CLOBBERED): Delete.
1405 * system.h (HARD_REGNO_CALL_PART_CLOBBERED): Poison.
1407 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
1408 Alan Hayward <alan.hayward@arm.com>
1409 David Sherwood <david.sherwood@arm.com>
1411 * rtl.h (subreg_memory_offset): Declare.
1412 * emit-rtl.c (subreg_memory_offset): New function.
1413 * expmed.c (store_bit_field_1): Use it.
1414 * expr.c (undefined_operand_subword_p): Likewise.
1415 * simplify-rtx.c (simplify_subreg): Likewise.
1417 2017-09-04 Alexander Monakov <amonakov@ispras.ru>
1419 PR rtl-optimization/57448
1422 * optabs.c (expand_atomic_load): Place compiler memory barriers if
1423 using atomic_load pattern.
1424 (expand_atomic_store): Likewise.
1426 2017-09-04 Jakub Jelinek <jakub@redhat.com>
1429 * gimple-fold.c (gimple_fold_call): Optimize away useless UBSAN_PTR
1430 and UBSAN_BOUNDS internal calls. Clean up IFN_UBSAN_OBJECT_SIZE
1431 handling. Use replace_call_with_value with NULL instead of
1432 gsi_replace, unlink_stmt_vdef and release_defs.
1434 * gdbhooks.py (OptMachineModePrinter.to_string): Use 8 spaces
1437 * lra-remat.c (reg_overlap_for_remat_p): Fix a pasto.
1439 2017-09-04 Richard Sandiford <richard.sandiford@linaro.org>
1442 * rtl.h (emit_library_call_value_1): Declare.
1443 (emit_library_call): Replace declaration with a series of overloads.
1444 Remove the parameter count argument.
1445 (emit_library_call_value): Likewise.
1446 * calls.c (emit_library_call_value_1): Make global. Replace varargs
1447 with an "rtx_mode_t *".
1448 (emit_library_call_value): Delete.
1449 (emit_library_call): Likewise.
1450 * asan.c (asan_emit_stack_protection): Update calls accordingly.
1451 (asan_emit_allocas_unpoison): Likewise.
1452 * builtins.c (expand_builtin_powi): Likewise.
1453 (expand_asan_emit_allocas_unpoison): Likewise.
1454 * cfgexpand.c (expand_main_function): Likewise.
1455 * config/aarch64/aarch64.c (aarch64_trampoline_init): Likewise.
1456 * config/aarch64/aarch64.h (PROFILE_HOOK): Likewise.
1457 * config/alpha/alpha.c (alpha_trampoline_init): Likewise.
1458 * config/arm/arm.c (arm_trampoline_init): Likewise.
1459 (arm_call_tls_get_addr): Likewise.
1460 (arm_expand_divmod_libfunc): Likewise.
1461 * config/bfin/bfin.md (umulsi3_highpart): Likewise.
1462 (smulsi3_highpart): Likewise.
1463 * config/c6x/c6x.c (c6x_initialize_trampoline): Likewise.
1464 (c6x_expand_compare): Likewise.
1465 (c6x_expand_movmem): Likewise.
1466 * config/frv/frv.c (frv_trampoline_init): Likewise.
1467 * config/i386/i386.c (ix86_trampoline_init): Likewise.
1468 (ix86_expand_divmod_libfunc): Likewise.
1469 * config/ia64/ia64.c (ia64_expand_tls_address): Likewise.
1470 (ia64_expand_compare): Likewise.
1471 (ia64_profile_hook): Likewise.
1472 * config/ia64/ia64.md (save_stack_nonlocal): Likewise.
1473 (nonlocal_goto): Likewise.
1474 (restore_stack_nonlocal): Likewise.
1475 * config/m32r/m32r.c (block_move_call): Likewise.
1476 (m32r_trampoline_init): Likewise.
1477 * config/m68k/linux.h (FINALIZE_TRAMPOLINE): Likewise.
1478 * config/m68k/m68k.c (m68k_call_tls_get_addr): Likewise.
1479 (m68k_call_m68k_read_tp): Likewise.
1480 * config/microblaze/microblaze.c (microblaze_call_tls_get_addr)
1481 (microblaze_expand_divide): Likewise.
1482 * config/mips/mips.h (mips_args): Likewise.
1483 * config/mips/sdemtk.h (mips_sync_icache): Likewise.
1484 (MIPS_ICACHE_SYNC): Likewise.
1485 * config/nios2/nios2.c (nios2_emit_expensive_div): Likewise.
1486 (nios2_trampoline_init): Likewise.
1487 * config/pa/pa.c (hppa_tls_call): Likewise.
1488 (pa_trampoline_init): Likewise.
1489 * config/pa/pa.md (canonicalize_funcptr_for_compare): Likewise.
1490 * config/powerpcspe/powerpcspe.c (rs6000_legitimize_tls_address)
1491 (expand_strn_compare): Likewise.
1492 (rs6000_generate_compare): Likewise.
1493 (rs6000_expand_float128_convert): Likewise.
1494 (output_profile_hook): Likewise.
1495 (rs6000_trampoline_init): Likewise.
1496 * config/powerpcspe/powerpcspe.md (neg<mode>2): Likewise.
1497 * config/riscv/riscv.h (PROFILE_HOOK): Likewise.
1498 * config/rs6000/rs6000-string.c (expand_strn_compare): Likewise.
1499 * config/rs6000/rs6000.c (rs6000_legitimize_tls_address): Likewise.
1500 (rs6000_generate_compare): Likewise.
1501 (rs6000_expand_float128_convert): Likewise.
1502 (output_profile_hook): Likewise.
1503 (rs6000_trampoline_init): Likewise.
1504 * config/rs6000/rs6000.md (neg<mode>2): Likewise.
1505 * config/sh/sh.c (sh_trampoline_init): Likewise.
1506 * config/sparc/sparc.c (emit_soft_tfmode_libcall): Likewise.
1507 (sparc_emit_float_lib_cmp): Likewise.
1508 (sparc32_initialize_trampoline): Likewise.
1509 (sparc64_initialize_trampoline): Likewise.
1510 (sparc_profile_hook): Likewise.
1511 * config/spu/spu.c (ea_load_store): Likewise.
1512 * config/spu/spu.md (floatunssidf2): Likewise.
1513 * config/tilegx/tilegx.c (tilegx_trampoline_init): Likewise.
1514 * config/tilepro/tilepro.c (tilepro_trampoline_init): Likewise.
1515 * config/visium/visium.c (expand_block_move_4): Likewise.
1516 (expand_block_move_2): Likewise.
1517 (expand_block_move_1): Likewise.
1518 (expand_block_set_4): Likewise.
1519 (expand_block_set_2): Likewise.
1520 (expand_block_set_1): Likewise.
1521 (visium_trampoline_init): Likewise.
1522 (visium_profile_hook): Likewise.
1523 * config/xtensa/xtensa.c (xtensa_expand_nonlocal_goto): Likewise.
1524 (xtensa_setup_frame_addresses): Likewise.
1525 (xtensa_trampoline_init): Likewise.
1526 * except.c (sjlj_emit_function_enter): Likewise.
1527 (sjlj_emit_function_exit): Likewise.
1528 * explow.c (allocate_dynamic_stack_space): Likewise.
1529 (probe_stack_range): Likewise.
1530 * expr.c (convert_mode_scalar): Likewise.
1531 * optabs.c (expand_binop): Likewise.
1532 (expand_twoval_binop_libfunc): Likewise.
1533 (expand_unop): Likewise.
1534 (prepare_cmp_insn): Likewise.
1535 (prepare_float_lib_cmp): Likewise.
1536 (expand_float): Likewise.
1537 (expand_fix): Likewise.
1538 (expand_fixed_convert): Likewise.
1539 (maybe_emit_sync_lock_test_and_set): Likewise.
1540 (expand_atomic_compare_and_swap): Likewise.
1541 (expand_mem_thread_fence): Likewise.
1542 (expand_atomic_fetch_op): Likewise.
1544 2017-09-03 Gerald Pfeifer <gerald@pfeifer.com>
1546 * doc/generic.texi (OpenACC): Adjust URL.
1547 * doc/invoke.texi (C Dialect Options): Ditto.
1549 2017-09-03 Uros Bizjak <ubizjak@gmail.com>
1551 * config/i386/i386.md (*bt<mode>): Use nonimmediate_operand
1552 predicate for operand 1. Add (m,<S>) constraint.
1553 (*jcc_bt<mode>): Use nonimmediate_operand predicate for operand 1.
1554 Prevent memory operand 1 with register operand 2.
1556 2017-09-01 Segher Boessenkool <segher@kernel.crashing.org>
1558 PR rtl-optimization/82024
1559 * combine.c (try_combine): If the combination result is a PARALLEL,
1560 and we only need to retain the SET in there that would be placed
1561 at I2, check that we can place that at I3 instead, before doing so.
1563 2017-09-01 Jakub Jelinek <jakub@redhat.com>
1566 * config/i386/i386.c (ix86_init_large_pic_reg): Return label
1568 (ix86_init_pic_reg): Remember label from ix86_init_large_pic_reg,
1569 if non-NULL and preceded by NOTE_INSN_BASIC_BLOCK, swap the note
1572 2017-09-01 Joerg Sonnenberger <joerg@bec.de>
1573 Jeff Law <law@redhat.com>
1575 * varasm.c (bss_initializer_p): Do not put constants into .bss
1576 (categorize_decl_for_section): Handle bss_initializer_p returning
1577 false when DECL_INITIAL is NULL.
1579 2017-09-01 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1582 * config/s390/s390.c (s390_can_inline_p): New function.
1584 2017-09-01 Jeff Law <law@redhat.com>
1586 PR tree-optimization/82052
1587 * tree-ssa-scopedtables.c (avail_exprs_stack::lookup_avail_expr):
1588 Always initialize the returned slot after a hash table miss
1589 when INSERT is true.
1591 2017-09-01 Alexander Monakov <amonakov@ispras.ru>
1593 * config/s390/s390.md (mem_signal_fence): Remove.
1594 * doc/md.texi (mem_signal_fence): Remove.
1595 * optabs.c (expand_mem_signal_fence): Remove uses of mem_signal_fence.
1597 * target-insns.def (mem_signal_fence): Remove.
1599 2017-09-01 Jakub Jelinek <jakub@redhat.com>
1602 * doc/invoke.texi: Document -fsanitize=pointer-overflow.
1605 * asan.c (create_odr_indicator): Strip name encoding from assembler
1606 name before appending it after __odr_asan_.
1608 2017-09-01 Martin Liska <mliska@suse.cz>
1610 PR tree-optimization/82059
1611 * gimple-ssa-isolate-paths.c (isolate_path): Add profile and
1612 frequency only when an edge is redirected.
1614 2017-09-01 Claudiu Zissulescu <claziss@synopsys.com>
1616 * config/arc/arc-c.c (__ARC_LPC_WIDTH__): Add builtin define.
1617 * config/arc/arc.c (ARC_MAX_LOOP_LENGTH): Define.
1618 (arc_conditional_register_usage): Remove ARC600 lp_count
1620 (arc_file_start): Emit Tag_ARC_CPU_variation.
1621 (arc_can_use_doloop_p): New conditions to use ZOLs.
1622 (hwloop_fail): New function.
1623 (hwloop_optimize): Likewise.
1624 (hwloop_pattern_reg): Likewise.
1625 (arc_doloop_hooks): New struct, to be used with reorg_loops.
1626 (arc_reorg_loops): New function, calls reorg_loops.
1627 (arc_reorg): Call arc_reorg_loops. Remove old ZOL handling.
1628 (arc600_corereg_hazard): Remove ZOL checking, case handled by
1630 (arc_loop_hazard): Remove function, functionality moved into
1632 (arc_hazard): Remove arc_loop_hazard call.
1633 (arc_adjust_insn_length): Remove ZOL handling, functionality moved
1634 into hwloop_optimize.
1635 (arc_label_align): Remove ZOL handling.
1636 * config/arc/arc.h (LOOP_ALIGN): Changed to 0.
1637 * config/arc/arc.md (doloop_begin): Remove pattern.
1638 (doloop_begin_i): Likewise.
1639 (doloop_end_i): Likewise.
1640 (doloop_fallback): Likewise.
1641 (doloop_fallback_m): Likewise.
1642 (doloop_end): Reimplement expand.
1643 (arc_lp): New pattern for LP instruction.
1644 (loop_end): New pattern.
1645 (loop_fail): Likewise.
1646 (decrement_and_branch_until_zero): Likewise.
1647 * config/arc/arc.opt (mlpc-width): New option.
1648 * doc/invoke.texi (mlpc-width): Document option.
1650 2017-09-01 Claudiu Zissulescu <claziss@synopsys.com>
1652 * config/arc/arc.c (arc_ifcvt): Remove use of merge_blocks call.
1653 (arc_ccfsm_advance): Fix checking for delay slots.
1654 (arc_reorg): Add rtl dump after each call to arc_ifcvt.
1656 2017-09-01 Claudiu Zissulescu <claziss@synopsys.com>
1658 * config/arc/arc.md (movqi_insn): Add stores to save constant long
1660 (movhi_insn): Update store instruction constraint which are saving
1661 6-bit short immediates.
1662 (movsi_insn): Consider also short scaled load operations.
1663 (zero_extendhisi2_i): Use Usd constraint instead of T.
1664 (extendhisi2_i): Add q constraint.
1665 (arc_clzsi2): Add type and length attributes.
1666 (arc_ctzsi2): Likewise.
1667 * config/arc/constraints.md (Usc): Update constraint, the
1668 assembler can parse two relocations for a single instruction.
1670 2017-09-01 Claudiu Zissulescu <claziss@synopsys.com>
1672 * config/arc/arc.c (arc_use_anchors_for_symbol_p): New function.
1673 (TARGET_USE_ANCHORS_FOR_SYMBOL_P): Define.
1675 2017-08-31 Olivier Hainque <hainque@adacore.com>
1677 * config.gcc (powerpc-wrs-vxworks|vxworksae|vxworksmils): Now
1678 match as powerpc-wrs-vxworks*.
1680 2017-08-31 James Greenhalgh <james.greenhalgh@arm.com>
1682 * config/aarch64/aarch64-simd.md (aarch64_mla_elt_merge<mode>): Fix
1683 register constraint for by-element operand.
1684 (aarch64_mls_elt_merge<mode>): Likewise.
1686 2017-08-31 Claudiu Zissulescu <claziss@synopsys.com>
1688 * config/arc/arc.c (arc_can_follow_jump): Check for short
1691 2017-08-31 Claudiu Zissulescu <claziss@synopsys.com>
1693 * config.gcc: Use g.opt for arc.
1694 * config/arc/arc.c (LEGITIMATE_SCALED_ADDRESS_P): Deleted,
1695 functionality moved to ...
1696 (legitimate_scaled_address_p): New function, ...here.
1697 (LEGITIMATE_SMALL_DATA_OFFSET_P): New define.
1698 (LEGITIMATE_SMALL_DATA_ADDRESS_P): Use the above define.
1699 (legitimate_offset_address_p): Delete TARGET_NO_SDATA_SET
1701 (arc_override_options): Handle G option.
1702 (arc_output_pic_addr_const): Correct function definition.
1703 (arc_legitimate_address_p): Use legitimate_scaled_address_p.
1704 (arc_decl_anon_ns_mem_p): Delete.
1705 (arc_in_small_data_p): Overhaul this function to take into
1706 consideration the value given via G option.
1707 (arc_rewrite_small_data_1): Renamed and corrected old
1708 arc_rewrite_small_data function.
1709 (arc_rewrite_small_data): New function.
1710 (small_data_pattern): Don't use pic_offset_table_rtx.
1711 * config/arc/arc.h (CC1_SPEC): Recognize G option.
1712 * config/arc/simdext.md (movmisalignv2hi): Use
1713 prepare_move_operands function.
1715 (movmisalign*): Likewise.
1716 * doc/invoke.texi (ARC options): Document -G option.
1718 2017-08-31 Claudiu Zissulescu <claziss@synopsys.com>
1720 * config/arc/arc-protos.h (compact_sda_memory_operand): Update
1722 * config/arc/arc.c (arc_print_operand): Output scalled address for
1723 sdata whenever is possible.
1724 (arc_in_small_data_p): Allow sdata for 64bit datum when double
1725 load/stores are available.
1726 (compact_sda_memory_operand): Check for the alignment required by
1727 code density instructions.
1728 * config/arc/arc.md (movsi_insn): Use newly introduced Us0
1730 * config/arc/constraints.md (Usd): Update constraint.
1731 (Us0): New constraint.
1732 (Usc): Update constraint.
1734 2017-08-31 Richard Biener <rguenther@suse.de>
1737 * dwarf2out.c (dwarf2out_early_global_decl): Process each
1740 2017-08-31 Tamar Christina <tamar.christina@arm.com>
1742 * config/aarch64/aarch64-builtins.c (aarch64_init_simd_builtins):
1743 Resize type_signature.
1745 2017-08-31 Richard Sandiford <richard.sandiford@linaro.org>
1746 Alan Hayward <alan.hayward@arm.com>
1747 David Sherwood <david.sherwood@arm.com>
1749 * config/aarch64/aarch64.c (aarch64_base_register_rtx_p): Only allow
1750 subregs whose inner modes can be stored in GPRs.
1751 (aarch64_classify_index): Likewise.
1753 2017-08-31 Richard Sandiford <richard.sandiford@linaro.org>
1754 Alan Hayward <alan.hayward@arm.com>
1755 David Sherwood <david.sherwood@arm.com>
1757 * config/aarch64/iterators.md (V_cmp_result): Rename to...
1758 (V_INT_EQUIV): ...this.
1759 (v_cmp_result): Rename to...
1760 (v_int_equiv): ...this.
1761 * config/aarch64/aarch64.md (xorsign<mode>3): Update accordingly.
1762 * config/aarch64/aarch64-simd.md (xorsign<mode>3): Likewise.
1763 (copysign<mode>3): Likewise.
1764 (aarch64_simd_bsl<mode>_internal): Likewise.
1765 (aarch64_simd_bsl<mode>): Likewise.
1766 (vec_cmp<mode><mode>): Likewise.
1767 (vcond<mode><mode>): Likewise.
1768 (vcond<v_cmp_mixed><mode>): Likewise.
1769 (vcondu<mode><v_cmp_mixed>): Likewise.
1770 (aarch64_cm<optab><mode>): Likewise.
1771 (aarch64_cmtst<mode>): Likewise.
1772 (aarch64_fac<optab><mode>): Likewise.
1773 (vec_perm_const<mode>): Likewise.
1774 (vcond_mask_<mode><v_cmp_result>): Rename to...
1775 (vcond_mask_<mode><v_int_equiv>): ...this.
1776 (vec_cmp<mode><v_cmp_result>): Rename to...
1777 (vec_cmp<mode><v_int_equiv>): ...this.
1779 2017-08-31 Richard Sandiford <richard.sandiford@linaro.org>
1780 Alan Hayward <alan.hayward@arm.com>
1781 David Sherwood <david.sherwood@arm.com>
1783 * config/aarch64/aarch64-modes.def: Remove 32-, 48- and 64-byte
1785 * config/aarch64/iterators.md (VRL2, VRL3, VRL4): Delete.
1786 * config/aarch64/aarch64.md (UNSPEC_LD2_DREG, UNSPEC_LD3_DREG)
1787 (UNSPEC_LD4_DREG): New unspecs.
1788 * config/aarch64/aarch64-simd.md (aarch64_ld2<mode>_dreg_le)
1789 (aarch64_ld2<mode>_dreg_be): Replace with...
1790 (aarch64_ld2<mode>_dreg): ...this pattern and use the new DREG
1792 (aarch64_ld3<mode>_dreg_le)
1793 (aarch64_ld3<mode>_dreg_be): Replace with...
1794 (aarch64_ld3<mode>_dreg): ...this pattern and use the new DREG
1796 (aarch64_ld4<mode>_dreg_le)
1797 (aarch64_ld4<mode>_dreg_be): Replace with...
1798 (aarch64_ld4<mode>_dreg): ...this pattern and use the new DREG
1801 2017-08-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1803 PR tree-optimization/81987
1804 * gimple-ssa-strength-reduction.c (insert_initializers): Don't
1805 insert an initializer in a location not dominated by the stride
1808 2017-08-30 Eric Botcazou <ebotcazou@adacore.com>
1810 * tree-eh.c (lower_try_finally_switch): Set the location of the finally
1811 on the entire header of the finally block in the fallthru case.
1813 2017-08-30 Eric Botcazou <ebotcazou@adacore.com>
1815 * varasm.c (decode_addr_const): Deal with INDIRECT_REF <INTEGER_CST>.
1817 2017-08-30 Pat Haugen <pthaugen@us.ibm.com>
1819 * config/rs6000/rs6000.c (rs6000_emit_prologue_move_from_cr): Rename from
1820 rs6000_emit_move_from_cr and call renamed function.
1821 (rs6000_emit_prologue): Call renamed functions.
1822 * config/rs6000/rs6000.md (prologue_movesi_from_cr): Rename from
1823 movesi_from_cr, remove volatile CRs.
1825 2017-08-30 Jon Beniston <jon@beniston.com>
1826 Richard Biener <rguenther@suse.de>
1828 * tree-vect-patterns.c (vect_pattern_recog_1): Use VECTOR_TYPE_P instead
1829 of VECTOR_MODE_P check.
1830 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Allow single
1831 element vector types.
1833 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1835 * df.h (df_read_modify_subreg_p): Remove in favor of...
1836 * rtl.h (read_modify_subreg_p): ...this new function. Take a
1837 const_rtx instead of an rtx.
1838 * cprop.c (local_cprop_find_used_regs): Update accordingly.
1839 * df-problems.c (df_word_lr_mark_ref): Likewise.
1840 * ira-lives.c (mark_pseudo_reg_live): Likewise.
1841 (mark_pseudo_reg_dead): Likewise.
1842 (mark_ref_dead): Likewise.
1843 * reginfo.c (init_subregs_of_mode): Likewise.
1844 * sched-deps.c (sched_analyze_1): Likewise.
1845 * df-scan.c (df_def_record_1): Likewise.
1846 (df_uses_record): Likewise.
1847 (df_read_modify_subreg_p): Remove in favor of...
1848 * rtlanal.c (read_modify_subreg_p): ...this new function. Take a
1849 const_rtx instead of an rtx.
1851 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1852 Alan Hayward <alan.hayward@arm.com>
1853 David Sherwood <david.sherwood@arm.com>
1855 * rtl.h (partial_subreg_p): New function.
1856 * caller-save.c (save_call_clobbered_regs): Use it.
1857 * calls.c (expand_call): Likewise.
1858 * combine.c (combinable_i3pat): Likewise.
1859 (simplify_set): Likewise.
1860 (make_extraction): Likewise.
1861 (make_compound_operation_int): Likewise.
1862 (gen_lowpart_or_truncate): Likewise.
1863 (force_to_mode): Likewise.
1864 (make_field_assignment): Likewise.
1865 (reg_truncated_to_mode): Likewise.
1866 (record_truncated_value): Likewise.
1867 (move_deaths): Likewise.
1868 * cse.c (record_jump_cond): Likewise.
1869 (cse_insn): Likewise.
1870 * cselib.c (cselib_lookup_1): Likewise.
1871 * expmed.c (extract_bit_field_using_extv): Likewise.
1872 * function.c (assign_parm_setup_reg): Likewise.
1873 * ifcvt.c (noce_convert_multiple_sets): Likewise.
1874 * ira-build.c (create_insn_allocnos): Likewise.
1875 * lra-coalesce.c (merge_pseudos): Likewise.
1876 * lra-constraints.c (match_reload): Likewise.
1877 (simplify_operand_subreg): Likewise.
1878 (curr_insn_transform): Likewise.
1879 * lra-lives.c (process_bb_lives): Likewise.
1880 * lra.c (new_insn_reg): Likewise.
1881 (lra_substitute_pseudo): Likewise.
1882 * regcprop.c (mode_change_ok): Likewise.
1883 (maybe_mode_change): Likewise.
1884 (copyprop_hardreg_forward_1): Likewise.
1885 * reload.c (push_reload): Likewise.
1886 (find_reloads): Likewise.
1887 (find_reloads_subreg_address): Likewise.
1888 * reload1.c (alter_reg): Likewise.
1889 (eliminate_regs_1): Likewise.
1890 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
1892 2017-08-30 David Edelsohn <dje.gcc@gmail.com>
1894 * config/rs6000/rs6000.c (rs6000_expand_binop_builtin): Revert
1895 back to if statements, including unpack.
1897 2017-08-30 Martin Liska <mliska@suse.cz>
1900 * ipa-icf-gimple.c (func_checker::compare_tree_list_operand):
1902 (func_checker::compare_asm_inputs_outputs): ... this function.
1903 (func_checker::compare_gimple_asm): Use the function to compare
1904 also ASM constrains.
1905 * ipa-icf-gimple.h: Rename the function.
1907 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1908 Alan Hayward <alan.hayward@arm.com>
1909 David Sherwood <david.sherwood@arm.com>
1911 * coretypes.h (complex_mode): New type.
1912 * gdbhooks.py (build_pretty_printer): Handle it.
1913 * machmode.h (complex_mode): New class.
1914 (complex_mode::includes_p): New function.
1915 (is_complex_int_mode): Likewise.
1916 (is_complex_float_mode): Likewise.
1917 * genmodes.c (get_mode_class): Handle complex mode classes.
1918 * function.c (expand_function_end): Use is_complex_int_mode.
1920 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1921 Alan Hayward <alan.hayward@arm.com>
1922 David Sherwood <david.sherwood@arm.com>
1924 * coretypes.h (scalar_mode_pod): New typedef.
1925 * gdbhooks.py (build_pretty_printer): Handle it.
1926 * machmode.h (gt_ggc_mx, gt_pch_nx): New functions.
1927 * fixed-value.h (fixed_value::mode): Change type to scalar_mode_pod.
1928 * fold-const.c (fold_convert_const_int_from_fixed): Use scalar_mode.
1929 * tree-streamer-in.c (unpack_ts_fixed_cst_value_fields): Use
1932 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1933 Alan Hayward <alan.hayward@arm.com>
1934 David Sherwood <david.sherwood@arm.com>
1936 * machmode.h (mode_for_vector): Take a scalar_mode instead
1938 * stor-layout.c (mode_for_vector): Likewise.
1939 * explow.c (promote_mode): Use as_a <scalar_mode>.
1940 * sdbout.c (sdbout_parms): Use is_a <scalar_mode>.
1942 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1943 Alan Hayward <alan.hayward@arm.com>
1944 David Sherwood <david.sherwood@arm.com>
1946 * target.def (preferred_simd_mode): Take a scalar_mode
1947 instead of a machine_mode.
1948 * targhooks.h (default_preferred_simd_mode): Likewise.
1949 * targhooks.c (default_preferred_simd_mode): Likewise.
1950 * config/arc/arc.c (arc_preferred_simd_mode): Likewise.
1951 * config/arm/arm.c (arm_preferred_simd_mode): Likewise.
1952 * config/c6x/c6x.c (c6x_preferred_simd_mode): Likewise.
1953 * config/epiphany/epiphany.c (epiphany_preferred_simd_mode): Likewise.
1954 * config/i386/i386.c (ix86_preferred_simd_mode): Likewise.
1955 * config/mips/mips.c (mips_preferred_simd_mode): Likewise.
1956 * config/nvptx/nvptx.c (nvptx_preferred_simd_mode): Likewise.
1957 * config/powerpcspe/powerpcspe.c (rs6000_preferred_simd_mode):
1959 * config/rs6000/rs6000.c (rs6000_preferred_simd_mode): Likewise.
1960 * config/s390/s390.c (s390_preferred_simd_mode): Likewise.
1961 * config/sparc/sparc.c (sparc_preferred_simd_mode): Likewise.
1962 * config/aarch64/aarch64.c (aarch64_preferred_simd_mode): Likewise.
1963 (aarch64_simd_scalar_immediate_valid_for_move): Update accordingly.
1964 * doc/tm.texi: Regenerate.
1965 * optabs-query.c (can_vec_mask_load_store_p): Return false for
1968 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1969 Alan Hayward <alan.hayward@arm.com>
1970 David Sherwood <david.sherwood@arm.com>
1972 * target.def (scalar_mode_supported_p): Take a scalar_mode
1973 instead of a machine_mode.
1974 * targhooks.h (default_scalar_mode_supported_p): Likewise.
1975 * targhooks.c (default_scalar_mode_supported_p): Likewise.
1976 * config/aarch64/aarch64.c (aarch64_scalar_mode_supported_p): Likewise.
1977 * config/alpha/alpha.c (alpha_scalar_mode_supported_p): Likewise.
1978 * config/arm/arm.c (arm_scalar_mode_supported_p): Likewise.
1979 * config/avr/avr.c (avr_scalar_mode_supported_p): Likewise.
1980 * config/c6x/c6x.c (c6x_scalar_mode_supported_p): Likewise.
1981 * config/i386/i386.c (ix86_scalar_mode_supported_p): Likewise.
1982 * config/ia64/ia64.c (ia64_scalar_mode_supported_p): Likewise.
1983 * config/mips/mips.c (mips_scalar_mode_supported_p): Likewise.
1984 * config/msp430/msp430.c (msp430_scalar_mode_supported_p): Likewise.
1985 * config/pa/pa.c (pa_scalar_mode_supported_p): Likewise.
1986 * config/pdp11/pdp11.c (pdp11_scalar_mode_supported_p): Likewise.
1987 * config/powerpcspe/powerpcspe.c (rs6000_scalar_mode_supported_p):
1989 * config/rs6000/rs6000.c (rs6000_scalar_mode_supported_p): Likewise.
1990 * config/s390/s390.c (s390_scalar_mode_supported_p): Likewise.
1991 * config/spu/spu.c (spu_scalar_mode_supported_p): Likewise.
1992 * config/tilegx/tilegx.c (tilegx_scalar_mode_supported_p): Likewise.
1993 * config/tilepro/tilepro.c (tilepro_scalar_mode_supported_p):
1995 * doc/tm.texi: Regenerate.
1997 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
1998 Alan Hayward <alan.hayward@arm.com>
1999 David Sherwood <david.sherwood@arm.com>
2001 * coretypes.h (opt_scalar_mode): New typedef.
2002 * gdbhooks.py (build_pretty_printers): Handle it.
2003 * machmode.h (mode_iterator::get_2xwider): Add overload for
2005 * emit-rtl.c (init_emit_once): Use opt_scalar_mode when iterating
2007 * expr.c (convert_mode_scalar): Likewise.
2008 * omp-low.c (omp_clause_aligned_alignment): Likewise.
2009 * optabs.c (expand_float): Likewise.
2010 (expand_fix): Likewise.
2011 * tree-vect-stmts.c (vectorizable_conversion): Likewise.
2013 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2014 Alan Hayward <alan.hayward@arm.com>
2015 David Sherwood <david.sherwood@arm.com>
2017 * optabs.c (expand_float): Explicitly check for scalars before
2018 using a branching expansion.
2019 (expand_fix): Likewise.
2021 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2022 Alan Hayward <alan.hayward@arm.com>
2023 David Sherwood <david.sherwood@arm.com>
2025 * expr.c (convert_mode): Split scalar handling out into...
2026 (convert_mode_scalar): ...this new function. Treat the modes
2029 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2030 Alan Hayward <alan.hayward@arm.com>
2031 David Sherwood <david.sherwood@arm.com>
2033 * omp-expand.c (expand_omp_atomic): Use is_int_mode, is_float_mode
2035 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Likewise.
2037 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2038 Alan Hayward <alan.hayward@arm.com>
2039 David Sherwood <david.sherwood@arm.com>
2041 * fixed-value.h (fixed_from_double_int): Take a scalar_mode
2042 rather than a machine_mode.
2043 (fixed_from_string): Likewise.
2044 (fixed_convert): Likewise.
2045 (fixed_convert_from_int): Likewise.
2046 (fixed_convert_from_real): Likewise.
2047 (real_convert_from_fixed): Likewise.
2048 * fixed-value.c (fixed_from_double_int): Likewise.
2049 (fixed_from_string): Likewise.
2050 (fixed_convert): Likewise.
2051 (fixed_convert_from_int): Likewise.
2052 (fixed_convert_from_real): Likewise.
2053 (real_convert_from_fixed): Likewise.
2054 * config/avr/avr.c (avr_out_round): Use as_a <scalar_mode>.
2056 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2057 Alan Hayward <alan.hayward@arm.com>
2058 David Sherwood <david.sherwood@arm.com>
2060 * emit-rtl.c (immed_double_const): Use is_a <scalar_mode> instead
2061 of separate mode class checks. Do not allow vector modes here.
2062 (immed_wide_int_const): Use as_a <scalar_mode>.
2063 * explow.c (trunc_int_for_mode): Likewise.
2064 * rtl.h (wi::int_traits<rtx_mode_t>::get_precision): Likewise.
2065 (wi::shwi): Likewise.
2066 (wi::min_value): Likewise.
2067 (wi::max_value): Likewise.
2068 * dwarf2out.c (loc_descriptor): Likewise.
2069 * simplify-rtx.c (simplify_immed_subreg): Fix rtx_mode_t argument
2072 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2073 Alan Hayward <alan.hayward@arm.com>
2074 David Sherwood <david.sherwood@arm.com>
2076 * tree.h (SCALAR_TYPE_MODE): New macro.
2077 * expr.c (expand_expr_addr_expr_1): Use it.
2078 (expand_expr_real_2): Likewise.
2079 * fold-const.c (fold_convert_const_fixed_from_fixed): Likeise.
2080 (fold_convert_const_fixed_from_int): Likewise.
2081 (fold_convert_const_fixed_from_real): Likewise.
2082 (native_encode_fixed): Likewise
2083 (native_encode_complex): Likewise
2084 (native_encode_vector): Likewise.
2085 (native_interpret_fixed): Likewise.
2086 (native_interpret_real): Likewise.
2087 (native_interpret_complex): Likewise.
2088 (native_interpret_vector): Likewise.
2089 * omp-simd-clone.c (simd_clone_adjust_return_type): Likewise.
2090 (simd_clone_adjust_argument_types): Likewise.
2091 (simd_clone_init_simd_arrays): Likewise.
2092 (simd_clone_adjust): Likewise.
2093 * stor-layout.c (layout_type): Likewise.
2094 * tree.c (build_minus_one_cst): Likewise.
2095 * tree-cfg.c (verify_gimple_assign_ternary): Likewise.
2096 * tree-inline.c (estimate_move_cost): Likewise.
2097 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Likewise.
2098 * tree-vect-loop.c (vect_create_epilog_for_reduction): Likewise.
2099 (vectorizable_reduction): Likewise.
2100 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Likewise.
2101 (vect_recog_mixed_size_cond_pattern): Likewise.
2102 (check_bool_pattern): Likewise.
2103 (adjust_bool_pattern): Likewise.
2104 (search_type_for_mask_1): Likewise.
2105 * tree-vect-slp.c (vect_schedule_slp_instance): Likewise.
2106 * tree-vect-stmts.c (vectorizable_conversion): Likewise.
2107 (vectorizable_load): Likewise.
2108 (vectorizable_store): Likewise.
2109 * ubsan.c (ubsan_encode_value): Likewise.
2110 * varasm.c (output_constant): Likewise.
2112 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2113 Alan Hayward <alan.hayward@arm.com>
2114 David Sherwood <david.sherwood@arm.com>
2116 * coretypes.h (scalar_mode): New class.
2117 * machmode.h (scalar_mode): Likewise.
2118 (scalar_mode::includes_p): New function.
2119 (mode_to_inner): Return a scalar_mode rather than a machine_mode.
2120 * gdbhooks.py (build_pretty_printers): Handle scalar_mode.
2121 * genmodes.c (get_mode_class): Handle remaining scalar modes.
2122 * cfgexpand.c (expand_debug_expr): Use scalar_mode.
2123 * expmed.c (store_bit_field_1): Likewise.
2124 (extract_bit_field_1): Likewise.
2125 * expr.c (write_complex_part): Likewise.
2126 (read_complex_part): Likewise.
2127 (emit_move_complex_push): Likewise.
2128 (expand_expr_real_2): Likewise.
2129 * function.c (assign_parm_setup_reg): Likewise.
2130 (assign_parms_unsplit_complex): Likewise.
2131 * optabs.c (expand_binop): Likewise.
2132 * rtlanal.c (subreg_get_info): Likewise.
2133 * simplify-rtx.c (simplify_immed_subreg): Likewise.
2134 * varasm.c (output_constant_pool_2): Likewise.
2136 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2137 Alan Hayward <alan.hayward@arm.com>
2138 David Sherwood <david.sherwood@arm.com>
2140 * expmed.c (extract_high_half): Use scalar_int_mode and remove
2142 (expmed_mult_highpart_optab): Likewise.
2143 (expmed_mult_highpart): Likewise.
2145 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2146 Alan Hayward <alan.hayward@arm.com>
2147 David Sherwood <david.sherwood@arm.com>
2149 * builtins.h (builtin_strncpy_read_str): Take a scalar_int_mode
2150 instead of a machine_mode.
2151 (builtin_memset_read_str): Likewise.
2152 * builtins.c (c_readstr): Likewise.
2153 (builtin_memcpy_read_str): Likewise.
2154 (builtin_strncpy_read_str): Likewise.
2155 (builtin_memset_read_str): Likewise.
2156 (builtin_memset_gen_str): Likewise.
2157 (expand_builtin_signbit): Use scalar_int_mode for local variables.
2158 * cfgexpand.c (convert_debug_memory_address): Take a scalar_int_mode
2159 instead of a machine_mode.
2160 * combine.c (simplify_if_then_else): Use scalar_int_mode for local
2162 (make_extraction): Likewise.
2163 (try_widen_shift_mode): Take and return scalar_int_modes instead
2165 * config/aarch64/aarch64.c (aarch64_libgcc_cmp_return_mode): Return
2166 a scalar_int_mode instead of a machine_mode.
2167 * config/avr/avr.c (avr_addr_space_address_mode): Likewise.
2168 (avr_addr_space_pointer_mode): Likewise.
2169 * config/cr16/cr16.c (cr16_unwind_word_mode): Likewise.
2170 * config/msp430/msp430.c (msp430_addr_space_pointer_mode): Likewise.
2171 (msp430_unwind_word_mode): Likewise.
2172 * config/spu/spu.c (spu_unwind_word_mode): Likewise.
2173 (spu_addr_space_pointer_mode): Likewise.
2174 (spu_addr_space_address_mode): Likewise.
2175 (spu_libgcc_cmp_return_mode): Likewise.
2176 (spu_libgcc_shift_count_mode): Likewise.
2177 * config/rl78/rl78.c (rl78_addr_space_address_mode): Likewise.
2178 (rl78_addr_space_pointer_mode): Likewise.
2179 (fl78_unwind_word_mode): Likewise.
2180 (rl78_valid_pointer_mode): Take a scalar_int_mode instead of a
2182 * config/alpha/alpha.c (vms_valid_pointer_mode): Likewise.
2183 * config/ia64/ia64.c (ia64_vms_valid_pointer_mode): Likewise.
2184 * config/mips/mips.c (mips_mode_rep_extended): Likewise.
2185 (mips_valid_pointer_mode): Likewise.
2186 * config/tilegx/tilegx.c (tilegx_mode_rep_extended): Likewise.
2187 * config/ft32/ft32.c (ft32_valid_pointer_mode): Likewise.
2188 (ft32_addr_space_pointer_mode): Return a scalar_int_mode instead
2190 (ft32_addr_space_address_mode): Likewise.
2191 * config/m32c/m32c.c (m32c_valid_pointer_mode): Take a
2192 scalar_int_mode instead of a machine_mode.
2193 (m32c_addr_space_pointer_mode): Return a scalar_int_mode instead
2195 (m32c_addr_space_address_mode): Likewise.
2196 * config/powerpcspe/powerpcspe.c (rs6000_abi_word_mode): Likewise.
2197 (rs6000_eh_return_filter_mode): Likewise.
2198 * config/rs6000/rs6000.c (rs6000_abi_word_mode): Likewise.
2199 (rs6000_eh_return_filter_mode): Likewise.
2200 * config/s390/s390.c (s390_libgcc_cmp_return_mode): Likewise.
2201 (s390_libgcc_shift_count_mode): Likewise.
2202 (s390_unwind_word_mode): Likewise.
2203 (s390_valid_pointer_mode): Take a scalar_int_mode rather than a
2205 * target.def (mode_rep_extended): Likewise.
2206 (valid_pointer_mode): Likewise.
2207 (addr_space.valid_pointer_mode): Likewise.
2208 (eh_return_filter_mode): Return a scalar_int_mode rather than
2210 (libgcc_cmp_return_mode): Likewise.
2211 (libgcc_shift_count_mode): Likewise.
2212 (unwind_word_mode): Likewise.
2213 (addr_space.pointer_mode): Likewise.
2214 (addr_space.address_mode): Likewise.
2215 * doc/tm.texi: Regenerate.
2216 * dojump.c (prefer_and_bit_test): Take a scalar_int_mode rather than
2218 (do_jump): Use scalar_int_mode for local variables.
2219 * dwarf2cfi.c (init_return_column_size): Take a scalar_int_mode
2220 rather than a machine_mode.
2221 * dwarf2out.c (convert_descriptor_to_mode): Likewise.
2222 (scompare_loc_descriptor_wide): Likewise.
2223 (scompare_loc_descriptor_narrow): Likewise.
2224 * emit-rtl.c (adjust_address_1): Use scalar_int_mode for local
2226 * except.c (sjlj_emit_dispatch_table): Likewise.
2227 (expand_builtin_eh_copy_values): Likewise.
2228 * explow.c (convert_memory_address_addr_space_1): Likewise.
2229 Take a scalar_int_mode rather than a machine_mode.
2230 (convert_memory_address_addr_space): Take a scalar_int_mode rather
2231 than a machine_mode.
2232 (memory_address_addr_space): Use scalar_int_mode for local variables.
2233 * expmed.h (expand_mult_highpart_adjust): Take a scalar_int_mode
2234 rather than a machine_mode.
2235 * expmed.c (mask_rtx): Likewise.
2236 (init_expmed_one_conv): Likewise.
2237 (expand_mult_highpart_adjust): Likewise.
2238 (extract_high_half): Likewise.
2239 (expmed_mult_highpart_optab): Likewise.
2240 (expmed_mult_highpart): Likewise.
2241 (expand_smod_pow2): Likewise.
2242 (expand_sdiv_pow2): Likewise.
2243 (emit_store_flag_int): Likewise.
2244 (adjust_bit_field_mem_for_reg): Use scalar_int_mode for local
2246 (extract_low_bits): Likewise.
2247 * expr.h (by_pieces_constfn): Take a scalar_int_mode rather than
2249 * expr.c (pieces_addr::adjust): Likewise.
2250 (can_store_by_pieces): Likewise.
2251 (store_by_pieces): Likewise.
2252 (clear_by_pieces_1): Likewise.
2253 (expand_expr_addr_expr_1): Likewise.
2254 (expand_expr_addr_expr): Use scalar_int_mode for local variables.
2255 (expand_expr_real_1): Likewise.
2256 (try_casesi): Likewise.
2257 * final.c (shorten_branches): Likewise.
2258 * fold-const.c (fold_convert_const_int_from_fixed): Change the
2259 type of "mode" to machine_mode.
2260 * internal-fn.c (expand_arith_overflow_result_store): Take a
2261 scalar_int_mode rather than a machine_mode.
2262 (expand_mul_overflow): Use scalar_int_mode for local variables.
2263 * loop-doloop.c (doloop_modify): Likewise.
2264 (doloop_optimize): Likewise.
2265 * optabs.c (expand_subword_shift): Take a scalar_int_mode rather
2266 than a machine_mode.
2267 (expand_doubleword_shift_condmove): Likewise.
2268 (expand_doubleword_shift): Likewise.
2269 (expand_doubleword_clz): Likewise.
2270 (expand_doubleword_popcount): Likewise.
2271 (expand_doubleword_parity): Likewise.
2272 (expand_absneg_bit): Use scalar_int_mode for local variables.
2273 (prepare_float_lib_cmp): Likewise.
2274 * rtl.h (convert_memory_address_addr_space_1): Take a scalar_int_mode
2275 rather than a machine_mode.
2276 (convert_memory_address_addr_space): Likewise.
2277 (get_mode_bounds): Likewise.
2278 (get_address_mode): Return a scalar_int_mode rather than a
2280 * rtlanal.c (get_address_mode): Likewise.
2281 * stor-layout.c (get_mode_bounds): Take a scalar_int_mode rather
2282 than a machine_mode.
2283 * targhooks.c (default_mode_rep_extended): Likewise.
2284 (default_valid_pointer_mode): Likewise.
2285 (default_addr_space_valid_pointer_mode): Likewise.
2286 (default_eh_return_filter_mode): Return a scalar_int_mode rather
2287 than a machine_mode.
2288 (default_libgcc_cmp_return_mode): Likewise.
2289 (default_libgcc_shift_count_mode): Likewise.
2290 (default_unwind_word_mode): Likewise.
2291 (default_addr_space_pointer_mode): Likewise.
2292 (default_addr_space_address_mode): Likewise.
2293 * targhooks.h (default_eh_return_filter_mode): Likewise.
2294 (default_libgcc_cmp_return_mode): Likewise.
2295 (default_libgcc_shift_count_mode): Likewise.
2296 (default_unwind_word_mode): Likewise.
2297 (default_addr_space_pointer_mode): Likewise.
2298 (default_addr_space_address_mode): Likewise.
2299 (default_mode_rep_extended): Take a scalar_int_mode rather than
2301 (default_valid_pointer_mode): Likewise.
2302 (default_addr_space_valid_pointer_mode): Likewise.
2303 * tree-ssa-address.c (addr_for_mem_ref): Use scalar_int_mode for
2305 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Take a scalar_int_mode
2306 rather than a machine_mode.
2307 * tree-switch-conversion.c (array_value_type): Use scalar_int_mode
2308 for local variables.
2309 * tree-vrp.c (simplify_float_conversion_using_ranges): Likewise.
2310 * var-tracking.c (use_narrower_mode): Take a scalar_int_mode rather
2311 than a machine_mode.
2313 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2314 Alan Hayward <alan.hayward@arm.com>
2315 David Sherwood <david.sherwood@arm.com>
2317 * dojump.c (do_jump_by_parts_greater_rtx): Change the type of
2318 the mode argument to scalar_int_mode.
2319 (do_jump_by_parts_zero_rtx): Likewise.
2320 (do_jump_by_parts_equality_rtx): Likewise.
2321 (do_jump_by_parts_greater): Take a mode argument.
2322 (do_jump_by_parts_equality): Likewise.
2323 (do_jump_1): Update calls accordingly.
2325 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2326 Alan Hayward <alan.hayward@arm.com>
2327 David Sherwood <david.sherwood@arm.com>
2329 * is-a.h (safe_dyn_cast): New function.
2330 * rtl.h (rtx_jump_table_data::get_data_mode): New function.
2331 (jump_table_for_label): Likewise.
2332 * final.c (final_addr_vec_align): Take an rtx_jump_table_data *
2333 instead of an rtx_insn *.
2334 (shorten_branches): Use dyn_cast instead of LABEL_P and
2335 JUMP_TABLE_DATA_P. Use jump_table_for_label and
2336 rtx_jump_table_data::get_data_mode.
2337 (final_scan_insn): Likewise.
2339 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2340 Alan Hayward <alan.hayward@arm.com>
2341 David Sherwood <david.sherwood@arm.com>
2343 * combine.c (try_combine): Use is_a <scalar_int_mode> when
2344 trying to combine a full-register integer set with a subreg
2347 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2348 Alan Hayward <alan.hayward@arm.com>
2349 David Sherwood <david.sherwood@arm.com>
2351 * expr.c (expand_expr_addr_expr): Add a new_tmode local variable
2352 that is always either address_mode or pointer_mode.
2354 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2355 Alan Hayward <alan.hayward@arm.com>
2356 David Sherwood <david.sherwood@arm.com>
2358 * expr.c (expand_expr_real_2): Use word_mode instead of innermode
2359 when the two are known to be equal.
2361 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2362 Alan Hayward <alan.hayward@arm.com>
2363 David Sherwood <david.sherwood@arm.com>
2365 * simplify-rtx.c (simplify_const_unary_operation): Use
2366 is_a <scalar_int_mode> instead of checking for a nonzero
2367 precision. Forcibly convert op_mode to a scalar_int_mode
2368 in that case. More clearly differentiate the operand and
2369 result modes and use the former when deciding what the value
2370 of a count-bits operation should be. Use is_int_mode instead
2371 of checking for a MODE_INT. Remove redundant check for whether
2372 this mode has a zero precision.
2374 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2375 Alan Hayward <alan.hayward@arm.com>
2376 David Sherwood <david.sherwood@arm.com>
2378 * optabs.c (widen_leading): Change the type of the mode argument
2379 to scalar_int_mode. Use opt_scalar_int_mode for the mode iterator.
2380 (widen_bswap): Likewise.
2381 (expand_parity): Likewise.
2382 (expand_ctz): Change the type of the mode argument to scalar_int_mode.
2383 (expand_ffs): Likewise.
2384 (epand_unop): Check for scalar integer modes before calling the
2387 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2388 Alan Hayward <alan.hayward@arm.com>
2389 David Sherwood <david.sherwood@arm.com>
2391 * expr.c (const_scalar_mask_from_tree): Add a mode argument.
2393 (expand_expr_real_1): Update call accordingly.
2395 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2396 Alan Hayward <alan.hayward@arm.com>
2397 David Sherwood <david.sherwood@arm.com>
2399 * expmed.c (store_bit_field_using_insv): Add op0_mode and
2400 value_mode arguments. Use scalar_int_mode internally.
2401 (store_bit_field_1): Rename the new integer mode from imode
2402 to op0_mode and use it instead of GET_MODE (op0). Update calls
2403 to store_split_bit_field, store_bit_field_using_insv and
2404 store_fixed_bit_field.
2405 (store_fixed_bit_field): Add op0_mode and value_mode arguments.
2406 Use scalar_int_mode internally. Use a bit count rather than a mode
2407 when calculating the largest bit size for get_best_mode.
2408 Update calls to store_split_bit_field and store_fixed_bit_field_1.
2409 (store_fixed_bit_field_1): Add mode and value_mode arguments.
2410 Remove assertion that OP0 has a scalar integer mode.
2411 (store_split_bit_field): Add op0_mode and value_mode arguments.
2412 Update calls to extract_fixed_bit_field.
2413 (extract_bit_field_using_extv): Add an op0_mode argument.
2414 Use scalar_int_mode internally.
2415 (extract_bit_field_1): Rename the new integer mode from imode to
2416 op0_mode and use it instead of GET_MODE (op0). Update calls to
2417 extract_split_bit_field, extract_bit_field_using_extv and
2418 extract_fixed_bit_field.
2419 (extract_fixed_bit_field): Add an op0_mode argument. Update calls
2420 to extract_split_bit_field and extract_fixed_bit_field_1.
2421 (extract_fixed_bit_field_1): Add a mode argument. Remove assertion
2422 that OP0 has a scalar integer mode. Use as_a <scalar_int_mode>
2424 (extract_split_bit_field): Add an op0_mode argument. Update call
2425 to extract_fixed_bit_field.
2427 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2428 Alan Hayward <alan.hayward@arm.com>
2429 David Sherwood <david.sherwood@arm.com>
2431 * cse.c (cse_insn): Use opt_scalar_int_mode for the mode iterator.
2432 * explow.c (hard_function_value): Likewise.
2433 * expmed.c (extract_fixed_bit_field_1): Likewise. Move the
2434 convert_to_mode call outside the loop.
2435 * expr.c (alignment_for_piecewise_move): Use opt_scalar_int_mode
2436 for the mode iterator. Require the mode specified by max_pieces
2438 (emit_block_move_via_movmem): Use opt_scalar_int_mode for the
2440 (copy_blkmode_to_reg): Likewise.
2441 (set_storage_via_setmem): Likewise.
2442 * optabs.c (prepare_cmp_insn): Likewise.
2443 * rtlanal.c (init_num_sign_bit_copies_in_rep): Likewise.
2444 * stor-layout.c (finish_bitfield_representative): Likewise.
2446 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2447 Alan Hayward <alan.hayward@arm.com>
2448 David Sherwood <david.sherwood@arm.com>
2450 * rtl.h (subreg_unpromoted_mode, subreg_promoted_mode): New functions.
2451 * expr.c (convert_move): Use them.
2452 (convert_modes): Likewise.
2453 (store_expr_with_bounds): Likewise.
2455 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2456 Alan Hayward <alan.hayward@arm.com>
2457 David Sherwood <david.sherwood@arm.com>
2459 * rtl.h (rtl_hooks::reg_nonzero_bits): Add a scalar_int_mode
2460 parameter for the mode of "x". Remove the "known_x", "known_mode"
2461 and "known_ret" arguments. Change the type of the mode argument
2463 (rtl_hooks:reg_num_sign_bit_copies): Likewise.
2464 * combine.c (reg_nonzero_bits_for_combine): Update accordingly.
2465 (reg_num_sign_bit_copies_for_combine): Likewise.
2466 * rtlanal.c (nonzero_bits1): Likewise.
2467 (num_sign_bit_copies1): Likewise.
2468 * rtlhooks-def.h (reg_nonzero_bits_general): Likewise.
2469 (reg_num_sign_bit_copies_general): Likewise.
2470 * rtlhooks.c (reg_num_sign_bit_copies_general): Likewise.
2471 (reg_nonzero_bits_general): Likewise.
2473 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2474 Alan Hayward <alan.hayward@arm.com>
2475 David Sherwood <david.sherwood@arm.com>
2477 * rtlanal.c (num_sign_bit_copies): Handle VOIDmode here rather
2478 than in subroutines. Return 1 for non-integer modes.
2479 (cached_num_sign_bit_copies): Change the type of the mode parameter
2481 (num_sign_bit_copies1): Likewise. Remove early exit for other mode
2482 classes. Handle CONST_INT_P first and then check whether X also
2483 has a scalar integer mode. Check the same thing for inner registers
2484 of a SUBREG and for values that are being extended or truncated.
2486 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2487 Alan Hayward <alan.hayward@arm.com>
2488 David Sherwood <david.sherwood@arm.com>
2490 * rtlanal.c (nonzero_bits): Handle VOIDmode here rather than
2491 in subroutines. Return the mode mask for non-integer modes.
2492 (cached_nonzero_bits): Change the type of the mode parameter
2494 (nonzero_bits1): Likewise. Remove early exit for other mode
2495 classes. Handle CONST_INT_P first and then check whether X
2496 also has a scalar integer mode.
2498 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2499 Alan Hayward <alan.hayward@arm.com>
2500 David Sherwood <david.sherwood@arm.com>
2502 * expr.c (widest_int_mode_for_size): Make the comment match the code.
2503 Return a scalar_int_mode and assert that the size is greater than
2505 (by_pieces_ninsns): Update accordingly and remove VOIDmode handling.
2506 (op_by_pieces_d::op_by_pieces_d): Likewise.
2507 (op_by_pieces_d::run): Likewise.
2508 (can_store_by_pieces): Likewise.
2510 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2511 Alan Hayward <alan.hayward@arm.com>
2512 David Sherwood <david.sherwood@arm.com>
2514 * combine.c (extract_left_shift): Add a mode argument and update
2516 (make_compound_operation_int): Change the type of the mode parameter
2517 to scalar_int_mode and update the call to extract_left_shift.
2519 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2520 Alan Hayward <alan.hayward@arm.com>
2521 David Sherwood <david.sherwood@arm.com>
2523 * combine.c (simplify_and_const_int): Change the type of the mode
2524 parameter to scalar_int_mode.
2525 (simplify_and_const_int_1): Likewise. Update recursive call.
2527 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2528 Alan Hayward <alan.hayward@arm.com>
2529 David Sherwood <david.sherwood@arm.com>
2531 * combine.c (simplify_compare_const): Check that the mode is a
2532 scalar_int_mode (rather than VOIDmode) before testing its
2534 (simplify_comparison): Move COMPARISON_P handling out of the
2535 loop and restrict the latter part of the loop to scalar_int_modes.
2536 Check is_a <scalar_int_mode> before calling HWI_COMPUTABLE_MODE_P
2537 and when considering SUBREG_REGs. Use is_int_mode instead of
2538 checking GET_MODE_CLASS against MODE_INT.
2540 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2541 Alan Hayward <alan.hayward@arm.com>
2542 David Sherwood <david.sherwood@arm.com>
2544 * combine.c (try_widen_shift_mode): Move check for equal modes to...
2545 (simplify_shift_const_1): ...here. Use scalar_int_mode for
2546 shift_unit_mode and for modes involved in scalar shifts.
2548 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2549 Alan Hayward <alan.hayward@arm.com>
2550 David Sherwood <david.sherwood@arm.com>
2552 * combine.c (force_int_to_mode): New function, split out from...
2553 (force_to_mode): ...here. Keep xmode up-to-date and use it
2554 instead of GET_MODE (x).
2556 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2557 Alan Hayward <alan.hayward@arm.com>
2558 David Sherwood <david.sherwood@arm.com>
2560 * optabs-query.h (extraction_insn::struct_mode): Change type to
2561 opt_scalar_int_mode and update comment.
2562 (extraction_insn::field_mode): Change type to scalar_int_mode.
2563 (extraction_insn::pos_mode): Likewise.
2564 * combine.c (make_extraction): Update accordingly.
2565 * optabs-query.c (get_traditional_extraction_insn): Likewise.
2566 (get_optab_extraction_insn): Likewise.
2567 * recog.c (simplify_while_replacing): Likewise.
2568 * expmed.c (narrow_bit_field_mem): Change the type of the mode
2569 parameter to opt_scalar_int_mode.
2571 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2572 Alan Hayward <alan.hayward@arm.com>
2573 David Sherwood <david.sherwood@arm.com>
2575 * machmode.h (bit_field_mode_iterator::next_mode): Take a pointer
2576 to a scalar_int_mode instead of a machine_mode.
2577 (bit_field_mode_iterator::m_mode): Change type to opt_scalar_int_mode.
2578 (get_best_mode): Return a boolean and use a pointer argument to store
2579 the selected mode. Replace the limit mode parameter with a bit limit.
2580 * expmed.c (adjust_bit_field_mem_for_reg): Use scalar_int_mode
2581 for the values returned by bit_field_mode_iterator::next_mode.
2582 (store_bit_field): Update call to get_best_mode.
2583 (store_fixed_bit_field): Likewise.
2584 (extract_fixed_bit_field): Likewise.
2585 * expr.c (optimize_bitfield_assignment_op): Likewise.
2586 * fold-const.c (optimize_bit_field_compare): Likewise.
2587 (fold_truth_andor_1): Likewise.
2588 * stor-layout.c (bit_field_mode_iterator::next_mode): As above.
2589 Update for new type of m_mode.
2590 (get_best_mode): As above.
2592 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2593 Alan Hayward <alan.hayward@arm.com>
2594 David Sherwood <david.sherwood@arm.com>
2596 * expmed.c (strict_volatile_bitfield_p): Change the type of fieldmode
2597 to scalar_int_mode. Remove check for SCALAR_INT_MODE_P.
2598 (store_bit_field): Check is_a <scalar_int_mode> before calling
2599 strict_volatile_bitfield_p.
2600 (extract_bit_field): Likewise.
2602 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2603 Alan Hayward <alan.hayward@arm.com>
2604 David Sherwood <david.sherwood@arm.com>
2606 * target.def (cstore_mode): Return a scalar_int_mode.
2607 * doc/tm.texi: Regenerate.
2608 * config/sparc/sparc.c (sparc_cstore_mode): Return a scalar_int_mode.
2609 * targhooks.h (default_cstore_mode): Likewise.
2610 * targhooks.c (default_cstore_mode): Likewise, using a forced
2612 * expmed.c (emit_cstore): Expect the target of the cstore to be
2615 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2616 Alan Hayward <alan.hayward@arm.com>
2617 David Sherwood <david.sherwood@arm.com>
2619 * cfgloop.h (rtx_iv): Change type of extend_mode and mode to
2621 (niter_desc): Likewise mode.
2622 (iv_analyze): Add a mode parameter.
2624 (iv_analyze_expr): Pass the mode paraeter before the rtx it describes
2625 and change its type to scalar_int_mode.
2626 * loop-iv.c: Update commentary at head of file.
2627 (iv_constant): Pass the mode paraeter before the rtx it describes
2628 and change its type to scalar_int_mode. Remove VOIDmode handling.
2629 (iv_subreg): Change the type of the mode parameter to scalar_int_mode.
2630 (iv_extend): Likewise.
2631 (shorten_into_mode): Likewise.
2632 (iv_add): Use scalar_int_mode.
2633 (iv_mult): Likewise.
2634 (iv_shift): Likewise.
2635 (canonicalize_iv_subregs): Likewise.
2636 (get_biv_step_1): Pass the outer_mode parameter before the rtx
2637 it describes and change its mode to scalar_int_mode. Also change
2638 the type of the returned inner_mode to scalar_int_mode.
2639 (get_biv_step): Likewise, turning outer_mode from a pointer
2640 into a direct parameter. Update call to get_biv_step_1.
2641 (iv_analyze_biv): Add an outer_mode parameter. Update calls to
2642 iv_constant and get_biv_step.
2643 (iv_analyze_expr): Pass the mode parameter before the rtx it describes
2644 and change its type to scalar_int_mode. Don't initialise iv->mode
2645 to VOIDmode and remove later checks for its still being VOIDmode.
2646 Update calls to iv_analyze_op and iv_analyze_expr. Check
2647 is_a <scalar_int_mode> when changing the mode under consideration.
2648 (iv_analyze_def): Ignore registers that don't have a scalar_int_mode.
2649 Update call to iv_analyze_expr.
2650 (iv_analyze_op): Add a mode parameter. Reject subregs whose
2651 inner register is not also a scalar_int_mode. Update call to
2653 (iv_analyze): Add a mode parameter. Update call to iv_analyze_op.
2654 (biv_p): Add a mode parameter. Update call to iv_analyze_biv.
2655 (iv_number_of_iterations): Use is_a <scalar_int_mode> instead of
2656 separate mode class checks. Update calls to iv_analyze. Remove
2657 fix-up of VOIDmodes after iv_analyze_biv.
2658 * loop-unroll.c (analyze_iv_to_split_insn): Reject registers that
2659 don't have a scalar_int_mode. Update call to biv_p.
2661 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2662 Alan Hayward <alan.hayward@arm.com>
2663 David Sherwood <david.sherwood@arm.com>
2665 * cfgexpand.c (convert_debug_memory_address): Use
2666 as_a <scalar_int_mode>.
2667 * combine.c (expand_compound_operation): Likewise.
2668 (make_extraction): Likewise.
2669 (change_zero_ext): Likewise.
2670 (simplify_comparison): Likewise.
2671 * cse.c (cse_insn): Likewise.
2672 * dwarf2out.c (minmax_loc_descriptor): Likewise.
2673 (mem_loc_descriptor): Likewise.
2674 (loc_descriptor): Likewise.
2675 * expmed.c (init_expmed_one_mode): Likewise.
2676 (synth_mult): Likewise.
2677 (emit_store_flag_1): Likewise.
2678 (expand_divmod): Likewise. Use HWI_COMPUTABLE_MODE_P instead
2679 of a comparison with size.
2680 * expr.c (expand_assignment): Use as_a <scalar_int_mode>.
2681 (reduce_to_bit_field_precision): Likewise.
2682 * function.c (expand_function_end): Likewise.
2683 * internal-fn.c (expand_arith_overflow_result_store): Likewise.
2684 * loop-doloop.c (doloop_modify): Likewise.
2685 * optabs.c (expand_binop): Likewise.
2686 (expand_unop): Likewise.
2687 (expand_copysign_absneg): Likewise.
2688 (prepare_cmp_insn): Likewise.
2689 (maybe_legitimize_operand): Likewise.
2690 * recog.c (const_scalar_int_operand): Likewise.
2691 * rtlanal.c (get_address_mode): Likewise.
2692 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
2693 (simplify_cond_clz_ctz): Likewise.
2694 * tree-nested.c (get_nl_goto_field): Likewise.
2695 * tree.c (build_vector_type_for_mode): Likewise.
2696 * var-tracking.c (use_narrower_mode): Likewise.
2698 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2699 Alan Hayward <alan.hayward@arm.com>
2700 David Sherwood <david.sherwood@arm.com>
2702 * tree.h (SCALAR_INT_TYPE_MODE): New macro.
2703 * builtins.c (expand_builtin_signbit): Use it.
2704 * cfgexpand.c (expand_debug_expr): Likewise.
2705 * dojump.c (do_jump): Likewise.
2706 (do_compare_and_jump): Likewise.
2707 * dwarf2cfi.c (expand_builtin_init_dwarf_reg_sizes): Likewise.
2708 * expmed.c (make_tree): Likewise.
2709 * expr.c (expand_expr_real_2): Likewise.
2710 (expand_expr_real_1): Likewise.
2711 (try_casesi): Likewise.
2712 * fold-const-call.c (fold_const_call_ss): Likewise.
2713 * fold-const.c (unextend): Likewise.
2714 (extract_muldiv_1): Likewise.
2715 (fold_single_bit_test): Likewise.
2716 (native_encode_int): Likewise.
2717 (native_encode_string): Likewise.
2718 (native_interpret_int): Likewise.
2719 * gimple-fold.c (gimple_fold_builtin_memset): Likewise.
2720 * internal-fn.c (expand_addsub_overflow): Likewise.
2721 (expand_neg_overflow): Likewise.
2722 (expand_mul_overflow): Likewise.
2723 (expand_arith_overflow): Likewise.
2724 * match.pd: Likewise.
2725 * stor-layout.c (layout_type): Likewise.
2726 * tree-cfg.c (verify_gimple_assign_ternary): Likewise.
2727 * tree-ssa-math-opts.c (convert_mult_to_widen): Likewise.
2728 * tree-ssanames.c (get_range_info): Likewise.
2729 * tree-switch-conversion.c (array_value_type) Likewise.
2730 * tree-vect-patterns.c (vect_recog_rotate_pattern): Likewise.
2731 (vect_recog_divmod_pattern): Likewise.
2732 (vect_recog_mixed_size_cond_pattern): Likewise.
2733 * tree-vrp.c (extract_range_basic): Likewise.
2734 (simplify_float_conversion_using_ranges): Likewise.
2735 * tree.c (int_fits_type_p): Likewise.
2736 * ubsan.c (instrument_bool_enum_load): Likewise.
2737 * varasm.c (mergeable_string_section): Likewise.
2738 (narrowing_initializer_constant_valid_p): Likewise.
2739 (output_constant): Likewise.
2741 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2742 Alan Hayward <alan.hayward@arm.com>
2743 David Sherwood <david.sherwood@arm.com>
2745 * machmode.h (NARROWEST_INT_MODE): New macro.
2746 * expr.c (alignment_for_piecewise_move): Use it instead of
2747 GET_CLASS_NARROWEST_MODE (MODE_INT).
2748 (push_block): Likewise.
2749 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
2751 * tree-vrp.c (simplify_float_conversion_using_ranges): Likewise.
2753 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2754 Alan Hayward <alan.hayward@arm.com>
2755 David Sherwood <david.sherwood@arm.com>
2757 * postreload.c (move2add_valid_value_p): Change the type of the
2758 mode parameter to scalar_int_mode.
2759 (move2add_use_add2_insn): Add a mode parameter and use it instead
2761 (move2add_use_add3_insn): Likewise.
2762 (reload_cse_move2add): Update accordingly.
2764 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2765 Alan Hayward <alan.hayward@arm.com>
2766 David Sherwood <david.sherwood@arm.com>
2768 * expr.c (expand_expr_real_2): Use scalar_int_mode for the
2770 * lower-subreg.c (resolve_shift_zext): Use is_a <scalar_int_mode>.
2771 * optabs.c (expand_unop): Likewise.
2773 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2774 Alan Hayward <alan.hayward@arm.com>
2775 David Sherwood <david.sherwood@arm.com>
2777 * dwarf2out.c (typed_binop): Change mode parameter to scalar_int_mode.
2778 (clz_loc_descriptor): Likewise. Remove SCALAR_INT_MODE_P check.
2779 (popcount_loc_descriptor): Likewise.
2780 (bswap_loc_descriptor): Likewise.
2781 (rotate_loc_descriptor): Likewise.
2782 (mem_loc_descriptor): Add is_a <scalar_int_mode> checks before
2783 calling the functions above.
2785 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2786 Alan Hayward <alan.hayward@arm.com>
2787 David Sherwood <david.sherwood@arm.com>
2789 * combine.c (sign_extend_short_imm): Add is_a <scalar_int_mode>
2791 (try_combine): Likewise.
2792 (simplify_if_then_else): Likewise.
2793 * cse.c (cse_insn): Likewise.
2794 * dwarf2out.c (mem_loc_descriptor): Likewise.
2795 * emit-rtl.c (gen_lowpart_common): Likewise.
2796 * simplify-rtx.c (simplify_truncation): Likewise.
2797 (simplify_binary_operation_1): Likewise.
2798 (simplify_const_relational_operation): Likewise.
2799 (simplify_ternary_operation): Likewise.
2800 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Likewise.
2802 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2803 Alan Hayward <alan.hayward@arm.com>
2804 David Sherwood <david.sherwood@arm.com>
2806 * cse.c (cse_insn): Add is_a <scalar_int_mode> checks.
2807 * reload.c (push_reload): Likewise.
2808 (find_reloads): Likewise.
2810 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2811 Alan Hayward <alan.hayward@arm.com>
2812 David Sherwood <david.sherwood@arm.com>
2814 * combine.c (find_split_point): Add is_a <scalar_int_mode> checks.
2815 (make_compound_operation_int): Likewise.
2816 (change_zero_ext): Likewise.
2817 * expr.c (convert_move): Likewise.
2818 (convert_modes): Likewise.
2819 * fwprop.c (forward_propagate_subreg): Likewise.
2820 * loop-iv.c (get_biv_step_1): Likewise.
2821 * optabs.c (widen_operand): Likewise.
2822 * postreload.c (move2add_valid_value_p): Likewise.
2823 * recog.c (simplify_while_replacing): Likewise.
2824 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
2825 (simplify_binary_operation_1): Likewise. Remove redundant
2826 mode equality check.
2828 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2829 Alan Hayward <alan.hayward@arm.com>
2830 David Sherwood <david.sherwood@arm.com>
2832 * combine.c (combine_simplify_rtx): Add checks for
2833 is_a <scalar_int_mode>.
2834 (simplify_if_then_else): Likewise.
2835 (make_field_assignment): Likewise.
2836 (simplify_comparison): Likewise.
2837 * ifcvt.c (noce_try_bitop): Likewise.
2838 * loop-invariant.c (canonicalize_address_mult): Likewise.
2839 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
2841 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2842 Alan Hayward <alan.hayward@arm.com>
2843 David Sherwood <david.sherwood@arm.com>
2845 * gimple-fold.c (gimple_fold_builtin_memory_op): Use
2846 is_a <scalar_int_mode> instead of != BLKmode.
2848 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2849 Alan Hayward <alan.hayward@arm.com>
2850 David Sherwood <david.sherwood@arm.com>
2852 * cfgexpand.c (expand_debug_expr): Use is_a <scalar_int_mode>
2853 instead of != VOIDmode.
2854 * combine.c (if_then_else_cond): Likewise.
2855 (change_zero_ext): Likewise.
2856 * dwarf2out.c (mem_loc_descriptor): Likewise.
2857 (loc_descriptor): Likewise.
2858 * rtlanal.c (canonicalize_condition): Likewise.
2859 * simplify-rtx.c (simplify_relational_operation_1): Likewise.
2861 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2862 Alan Hayward <alan.hayward@arm.com>
2863 David Sherwood <david.sherwood@arm.com>
2865 * simplify-rtx.c (simplify_binary_operation_1): Use
2866 is_a <scalar_int_mode> instead of !VECTOR_MODE_P.
2868 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2869 Alan Hayward <alan.hayward@arm.com>
2870 David Sherwood <david.sherwood@arm.com>
2872 * wide-int.h (int_traits<unsigned char>) New class.
2873 (int_traits<unsigned short>) Likewise.
2874 * cfgexpand.c (expand_debug_expr): Use is_a <scalar_int_mode>.
2875 Use GET_MODE_UNIT_PRECISION and remove redundant test for
2877 * combine.c (set_nonzero_bits_and_sign_copies): Use
2878 is_a <scalar_int_mode>.
2879 (find_split_point): Likewise.
2880 (combine_simplify_rtx): Likewise.
2881 (simplify_logical): Likewise.
2882 (expand_compound_operation): Likewise.
2883 (expand_field_assignment): Likewise.
2884 (make_compound_operation): Likewise.
2885 (extended_count): Likewise.
2886 (change_zero_ext): Likewise.
2887 (simplify_comparison): Likewise.
2888 * dwarf2out.c (scompare_loc_descriptor): Likewise.
2889 (ucompare_loc_descriptor): Likewise.
2890 (minmax_loc_descriptor): Likewise.
2891 (mem_loc_descriptor): Likewise.
2892 (loc_descriptor): Likewise.
2893 * expmed.c (init_expmed_one_mode): Likewise.
2894 * lra-constraints.c (lra_constraint_offset): Likewise.
2895 * optabs.c (prepare_libcall_arg): Likewise.
2896 * postreload.c (move2add_note_store): Likewise.
2897 * reload.c (operands_match_p): Likewise.
2898 * rtl.h (load_extend_op): Likewise.
2899 * rtlhooks.c (gen_lowpart_general): Likewise.
2900 * simplify-rtx.c (simplify_truncation): Likewise.
2901 (simplify_unary_operation_1): Likewise.
2902 (simplify_binary_operation_1): Likewise.
2903 (simplify_const_binary_operation): Likewise.
2904 (simplify_const_relational_operation): Likewise.
2905 (simplify_subreg): Likewise.
2906 * stor-layout.c (bitwise_mode_for_mode): Likewise.
2907 * var-tracking.c (adjust_mems): Likewise.
2908 (prepare_call_arguments): Likewise.
2910 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2911 Alan Hayward <alan.hayward@arm.com>
2912 David Sherwood <david.sherwood@arm.com>
2914 * machmode.h (is_int_mode): New fuction.
2915 * combine.c (find_split_point): Use it.
2916 (combine_simplify_rtx): Likewise.
2917 (simplify_if_then_else): Likewise.
2918 (simplify_set): Likewise.
2919 (simplify_shift_const_1): Likewise.
2920 (simplify_comparison): Likewise.
2921 * config/aarch64/aarch64.c (aarch64_rtx_costs): Likewise.
2922 * cse.c (notreg_cost): Likewise.
2923 (cse_insn): Likewise.
2924 * cselib.c (cselib_lookup_1): Likewise.
2925 * dojump.c (do_jump_1): Likewise.
2926 (do_compare_rtx_and_jump): Likewise.
2927 * dse.c (get_call_args): Likewise.
2928 * dwarf2out.c (rtl_for_decl_init): Likewise.
2929 (native_encode_initializer): Likewise.
2930 * expmed.c (emit_store_flag_1): Likewise.
2931 (emit_store_flag): Likewise.
2932 * expr.c (convert_modes): Likewise.
2933 (store_field): Likewise.
2934 (expand_expr_real_1): Likewise.
2935 * fold-const.c (fold_read_from_constant_string): Likewise.
2936 * gimple-ssa-sprintf.c (get_format_string): Likewise.
2937 * optabs-libfuncs.c (gen_int_libfunc): Likewise.
2938 * optabs.c (expand_binop): Likewise.
2939 (expand_unop): Likewise.
2940 (expand_abs_nojump): Likewise.
2941 (expand_one_cmpl_abs_nojump): Likewise.
2942 * simplify-rtx.c (mode_signbit_p): Likewise.
2943 (val_signbit_p): Likewise.
2944 (val_signbit_known_set_p): Likewise.
2945 (val_signbit_known_clear_p): Likewise.
2946 (simplify_relational_operation_1): Likewise.
2947 * tree.c (vector_type_mode): Likewise.
2949 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2950 Alan Hayward <alan.hayward@arm.com>
2951 David Sherwood <david.sherwood@arm.com>
2953 * machmode.h (smallest_mode_for_size): Fix formatting.
2954 (smallest_int_mode_for_size): New function.
2955 * cfgexpand.c (expand_debug_expr): Use smallest_int_mode_for_size
2956 instead of smallest_mode_for_size.
2957 * combine.c (make_extraction): Likewise.
2958 * config/arc/arc.c (arc_expand_movmem): Likewise.
2959 * config/arm/arm.c (arm_expand_divmod_libfunc): Likewise.
2960 * config/i386/i386.c (ix86_get_mask_mode): Likewise.
2961 * config/s390/s390.c (s390_expand_insv): Likewise.
2962 * config/sparc/sparc.c (assign_int_registers): Likewise.
2963 * config/spu/spu.c (spu_function_value): Likewise.
2964 (spu_function_arg): Likewise.
2965 * coverage.c (get_gcov_type): Likewise.
2966 (get_gcov_unsigned_t): Likewise.
2967 * dse.c (find_shift_sequence): Likewise.
2968 * expmed.c (store_bit_field_1): Likewise.
2969 * expr.c (convert_move): Likewise.
2970 (store_field): Likewise.
2971 * internal-fn.c (expand_arith_overflow): Likewise.
2972 * optabs-query.c (get_best_extraction_insn): Likewise.
2973 * optabs.c (expand_twoval_binop_libfunc): Likewise.
2974 * stor-layout.c (layout_type): Likewise.
2975 (initialize_sizetypes): Likewise.
2976 * targhooks.c (default_get_mask_mode): Likewise.
2977 * tree-ssa-loop-manip.c (canonicalize_loop_ivs): Likewise.
2979 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
2980 Alan Hayward <alan.hayward@arm.com>
2981 David Sherwood <david.sherwood@arm.com>
2983 * machmode.h (opt_mode::else_blk): New function.
2984 (int_mode_for_mode): Declare.
2985 * stor-layout.c (int_mode_for_mode): Return an opt_scalar_int_mode.
2986 * builtins.c (expand_builtin_signbit): Adjust for new int_mode_for_mode
2988 * cfgexpand.c (expand_debug_expr): Likewise.
2989 * combine.c (gen_lowpart_or_truncate): Likewise.
2990 (gen_lowpart_for_combine): Likewise.
2991 * config/aarch64/aarch64.c (aarch64_emit_approx_sqrt): Likewise.
2992 * config/avr/avr.c (avr_to_int_mode): Likewise.
2993 (avr_out_plus_1): Likewise.
2994 (avr_out_plus): Likewise.
2995 (avr_out_round): Likewise.
2996 * config/i386/i386.c (ix86_split_to_parts): Likewise.
2997 * config/s390/s390.c (s390_expand_vec_compare_cc): Likewise.
2998 (s390_expand_vcond): Likewise.
2999 * config/spu/spu.c (spu_split_immediate): Likewise.
3000 (spu_expand_mov): Likewise.
3001 * dse.c (get_stored_val): Likewise.
3002 * expmed.c (store_bit_field_1): Likewise.
3003 (convert_extracted_bit_field): Use int_mode_for_mode instead of
3005 (extract_bit_field_1): Adjust for new int_mode_for_mode return type.
3006 (extract_low_bits): Likewise.
3007 * expr.c (emit_group_load_1): Likewise. Separate out the BLKmode
3008 handling rather than repeating the check.
3009 (emit_group_store): Likewise.
3010 (emit_move_via_integer): Adjust for new int_mode_for_mode return type.
3011 * optabs.c (expand_absneg_bit): Likewise.
3012 (expand_copysign_absneg): Likewise.
3013 (expand_copysign_bit): Likewise.
3014 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3015 * tree-vect-slp.c (vect_transform_slp_perm_load): Likewise.
3016 * tree-vect-stmts.c (vect_gen_perm_mask_any): Likewise.
3017 * var-tracking.c (prepare_call_arguments): Likewise.
3018 * config/powerpcspe/powerpcspe.c (rs6000_do_expand_vec_perm): Use
3019 int_mode_for_mode instead of mode_for_size.
3020 * config/rs6000/rs6000.c (rs6000_do_expand_vec_perm): Likewise.
3022 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3023 Alan Hayward <alan.hayward@arm.com>
3024 David Sherwood <david.sherwood@arm.com>
3026 * machmode.h (int_mode_for_size): New function.
3027 * builtins.c (set_builtin_user_assembler_name): Use int_mode_for_size
3028 instead of mode_for_size.
3029 * calls.c (save_fixed_argument_area): Likewise. Make use of BLKmode
3031 * combine.c (expand_field_assignment): Use int_mode_for_size
3032 instead of mode_for_size.
3033 (make_extraction): Likewise.
3034 (simplify_shift_const_1): Likewise.
3035 (simplify_comparison): Likewise.
3036 * dojump.c (do_jump): Likewise.
3037 * dwarf2out.c (mem_loc_descriptor): Likewise.
3038 * emit-rtl.c (init_derived_machine_modes): Likewise.
3039 * expmed.c (flip_storage_order): Likewise.
3040 (convert_extracted_bit_field): Likewise.
3041 * expr.c (copy_blkmode_from_reg): Likewise.
3042 * graphite-isl-ast-to-gimple.c (max_mode_int_precision): Likewise.
3043 * internal-fn.c (expand_mul_overflow): Likewise.
3044 * lower-subreg.c (simple_move): Likewise.
3045 * optabs-libfuncs.c (init_optabs): Likewise.
3046 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
3047 * tree.c (vector_type_mode): Likewise.
3048 * tree-ssa-strlen.c (handle_builtin_memcmp): Likewise.
3049 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Likewise.
3050 * tree-vect-generic.c (expand_vector_parallel): Likewise.
3051 * tree-vect-stmts.c (vectorizable_load): Likewise.
3052 (vectorizable_store): Likewise.
3054 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3055 Alan Hayward <alan.hayward@arm.com>
3056 David Sherwood <david.sherwood@arm.com>
3058 * coretypes.h (pod_mode): New type.
3059 (scalar_int_mode_pod): New typedef.
3060 * machmode.h (pod_mode): New class.
3061 (int_n_data_t::m): Change type to scalar_int_mode_pod.
3062 * genmodes.c (emit_mode_int_n): Update accordingly.
3063 * lower-subreg.h (target_lower_subreg): Change type to
3064 scalar_int_mode_pod.
3065 * gdbhooks.py (build_pretty_printer): Handle pod_mode and
3066 scalar_int_mode_pod.
3068 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3069 Alan Hayward <alan.hayward@arm.com>
3070 David Sherwood <david.sherwood@arm.com>
3072 * config/powerpcspe/powerpcspe.h (rs6000_pmode): Change type from
3073 machine_mode to scalar_int_mode.
3074 * config/powerpcspe/powerpcspe.c (rs6000_pmode): Likewise.
3075 (rs6000_option_override_internal): Remove cast to int.
3076 * config/rs6000/rs6000.h (rs6000_pmode): Change type from
3077 machine_mode to scalar_int_mode.
3078 * config/rs6000/rs6000.c (rs6000_pmode): Likewise.
3079 (rs6000_option_override_internal): Remove cast to int.
3080 * config/s390/s390.h (Pmode): Remove cast to machine_mode.
3081 * config/epiphany/epiphany.h (RTX_OK_FOR_OFFSET_P): Add cast
3083 * config/s390/s390.c (s390_expand_builtin): Likewise.
3084 * coretypes.h (scalar_int_mode): New type.
3085 (opt_scalar_int_mode): New typedef.
3086 * machmode.h (scalar_int_mode): New class.
3087 (scalar_int_mode::includes_p): New function.
3088 (byte_mode): Change type to scalar_int_mode.
3089 (word_mode): Likewise.
3090 (ptr_mode): Likewise.
3091 * emit-rtl.c (byte_mode): Likewise.
3092 (word_mode): Likewise.
3093 (ptr_mode): Likewise.
3094 (init_derived_machine_modes): Update accordingly.
3095 * genmodes.c (get_mode_class): Return scalar_int_mode for MODE_INT
3096 and MODE_PARTIAL_INT.
3097 * gdbhooks.py (build_pretty_printer): Handle scalar_int_mode and
3098 opt_scalar_int_mode.
3100 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3101 Alan Hayward <alan.hayward@arm.com>
3102 David Sherwood <david.sherwood@arm.com>
3104 * target.def (libgcc_floating_mode_supported_p): Take a
3106 * doc/tm.texi: Regenerate.
3107 * targhooks.h (default_libgcc_floating_mode_supported_p): Take a
3109 * targhooks.c (default_libgcc_floating_mode_supported_p): Likewise.
3110 * config/aarch64/aarch64.c (aarch64_libgcc_floating_mode_supported_p):
3113 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3114 Alan Hayward <alan.hayward@arm.com>
3115 David Sherwood <david.sherwood@arm.com>
3117 * target.def (default_floatn_mode): Return an opt_scalar_float_mode.
3118 * doc/tm.texi: Regenerate.
3119 * config/arm/arm.c (arm_floatn_mode): Return an opt_scalar_float_mode.
3120 * config/powerpcspe/powerpcspe.c (rs6000_floatn_mode): Likewise.
3121 * config/rs6000/rs6000.c (rs6000_floatn_mode): Likewise.
3122 * targhooks.h (default_floatn_mode): Likewise.
3123 * targhooks.c (default_floatn_mode): Likewise.
3124 * tree.c (build_common_tree_nodes): Update accordingly.
3126 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3127 Alan Hayward <alan.hayward@arm.com>
3128 David Sherwood <david.sherwood@arm.com>
3130 * machmode.h (mode_iterator::start): Provide overload for opt_modes.
3131 (mode_iterator::iterate_p): Likewise.
3132 (mode_iterator::get_wider): Likewise.
3133 * expr.c (init_expr_target): Use opt_scalar_float_mode.
3135 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3136 Alan Hayward <alan.hayward@arm.com>
3137 David Sherwood <david.sherwood@arm.com>
3139 * coretypes.h (opt_scalar_float_mode): New typedef.
3140 * machmode.h (float_mode_for_size): New function.
3141 * emit-rtl.c (double_mode): Delete.
3142 (init_emit_once): Use float_mode_for_size.
3143 * stor-layout.c (layout_type): Likewise.
3144 * gdbhooks.py (build_pretty_printer): Handle opt_scalar_float_mode.
3146 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3147 Alan Hayward <alan.hayward@arm.com>
3148 David Sherwood <david.sherwood@arm.com>
3150 * output.h (assemble_real): Take a scalar_float_mode.
3151 * config/arm/arm.c (arm_assemble_integer): Update accordingly.
3152 * config/arm/arm.md (consttable_4): Likewise.
3153 (consttable_8): Likewise.
3154 (consttable_16): Likewise.
3155 * config/mips/mips.md (consttable_float): Likewise.
3156 * config/s390/s390.c (s390_output_pool_entry): Likewise.
3157 * varasm.c (assemble_real): Take a scalar_float_mode.
3158 (output_constant_pool_2): Update accordingly.
3159 (output_constant): Likewise.
3161 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3162 Alan Hayward <alan.hayward@arm.com>
3163 David Sherwood <david.sherwood@arm.com>
3165 * tree.h (SCALAR_FLOAT_TYPE_MODE): New macro.
3166 * builtins.c (expand_builtin_signbit): Use it instead of TYPE_MODE.
3167 * fold-const.c (fold_convert_const_real_from_fixed): Likewise.
3168 (native_encode_real): Likewise.
3169 (native_interpret_real): Likewise.
3170 * hsa-brig.c (emit_immediate_scalar_to_buffer): Likewise.
3171 * tree-vrp.c (simplify_float_conversion_using_ranges): Likewise.
3173 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3174 Alan Hayward <alan.hayward@arm.com>
3175 David Sherwood <david.sherwood@arm.com>
3177 * optabs-libfuncs.c (gen_trunc_conv_libfunc): Use is_a
3178 <scalar_float_mode>. Simplify.
3179 (gen_extend_conv_libfunc): Likewise.
3181 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3182 Alan Hayward <alan.hayward@arm.com>
3183 David Sherwood <david.sherwood@arm.com>
3185 * coretypes.h (scalar_float_mode): New type.
3186 * machmode.h (mode_traits::from_int): Use machine_mode if
3187 USE_ENUM_MODES is defined.
3188 (is_a): New function.
3190 (dyn_cast): Likewise.
3191 (scalar_float_mode): New class.
3192 (scalar_float_mode::includes_p): New function.
3193 (is_float_mode): Likewise.
3194 * gdbhooks.py (MachineModePrinter): New class.
3195 (build_pretty_printer): Use it for scalar_float_mode.
3196 * real.h (FLOAT_MODE_FORMAT): Use as_a <scalar_float_mode>.
3197 (format_helper::format_helper): Turn into a template.
3198 * genmodes.c (get_mode_class): New function.
3199 (emit_insn_modes_h): Give modes the class returned by get_mode_class,
3200 or machine_mode if none.
3201 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Use
3202 as_a <scalar_float_mode>.
3203 * dwarf2out.c (mem_loc_descriptor): Likewise.
3204 (insert_float): Likewise.
3205 (add_const_value_attribute): Likewise.
3206 * simplify-rtx.c (simplify_immed_subreg): Likewise.
3207 * optabs.c (expand_absneg_bit): Take a scalar_float_mode.
3208 (expand_unop): Update accordingly.
3209 (expand_abs_nojump): Likewise.
3210 (expand_copysign_absneg): Take a scalar_float_mode.
3211 (expand_copysign_bit): Likewise.
3212 (expand_copysign): Update accordingly.
3214 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3215 Alan Hayward <alan.hayward@arm.com>
3216 David Sherwood <david.sherwood@arm.com>
3218 * coretypes.h (opt_mode): New class.
3219 * machmode.h (opt_mode): Likewise.
3220 (opt_mode::else_void): New function.
3221 (opt_mode::require): Likewise.
3222 (opt_mode::exists): Likewise.
3223 (GET_MODE_WIDER_MODE): Turn into a function and return an opt_mode.
3224 (GET_MODE_2XWIDER_MODE): Likewise.
3225 (mode_iterator::get_wider): Update accordingly.
3226 (mode_iterator::get_2xwider): Likewise.
3227 (mode_iterator::get_known_wider): Likewise, turning into a template.
3228 * combine.c (make_extraction): Update use of GET_MODE_WIDER_MODE,
3229 forcing a wider mode to exist.
3230 * config/cr16/cr16.h (LONG_REG_P): Likewise.
3231 * rtlanal.c (init_num_sign_bit_copies_in_rep): Likewise.
3232 * config/c6x/c6x.c (c6x_rtx_costs): Update use of
3233 GET_MODE_2XWIDER_MODE, forcing a wider mode to exist.
3234 * lower-subreg.c (init_lower_subreg): Likewise.
3235 * optabs-libfuncs.c (init_sync_libfuncs_1): Likewise, but not
3236 on the final iteration.
3237 * config/i386/i386.c (ix86_expand_set_or_movmem): Check whether
3238 a wider mode exists before asking for a move pattern.
3239 (get_mode_wider_vector): Update use of GET_MODE_WIDER_MODE,
3240 forcing a wider mode to exist.
3241 (expand_vselect_vconcat): Update use of GET_MODE_2XWIDER_MODE,
3242 returning false if no such mode exists.
3243 * config/ia64/ia64.c (expand_vselect_vconcat): Likewise.
3244 * config/mips/mips.c (mips_expand_vselect_vconcat): Likewise.
3245 * expmed.c (init_expmed_one_mode): Update use of GET_MODE_WIDER_MODE.
3246 Avoid checking for a MODE_INT if we already know the mode is not a
3248 (extract_high_half): Update use of GET_MODE_WIDER_MODE,
3249 forcing a wider mode to exist.
3250 (expmed_mult_highpart_optab): Likewise.
3251 (expmed_mult_highpart): Likewise.
3252 * expr.c (expand_expr_real_2): Update use of GET_MODE_WIDER_MODE,
3254 * lto-streamer-in.c (lto_input_mode_table): Likewise.
3255 * optabs-query.c (find_widening_optab_handler_and_mode): Likewise.
3256 * stor-layout.c (bit_field_mode_iterator::next_mode): Likewise.
3257 * internal-fn.c (expand_mul_overflow): Update use of
3258 GET_MODE_2XWIDER_MODE.
3259 * omp-low.c (omp_clause_aligned_alignment): Likewise.
3260 * tree-ssa-math-opts.c (convert_mult_to_widen): Update use of
3261 GET_MODE_WIDER_MODE.
3262 (convert_plusminus_to_widen): Likewise.
3263 * tree-switch-conversion.c (array_value_type): Likewise.
3264 * var-tracking.c (emit_note_insn_var_location): Likewise.
3265 * tree-vrp.c (simplify_float_conversion_using_ranges): Likewise.
3266 Return false inside rather than outside the loop if no wider mode
3268 * optabs.c (expand_binop): Update use of GET_MODE_WIDER_MODE
3269 and GET_MODE_2XWIDER_MODE
3270 (can_compare_p): Use else_void.
3271 * gdbhooks.py (OptMachineModePrinter): New class.
3272 (build_pretty_printer): Use it for opt_mode.
3274 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3275 Alan Hayward <alan.hayward@arm.com>
3276 David Sherwood <david.sherwood@arm.com>
3278 * tree-switch-conversion.c (array_value_type): Only read TYPE_MODE
3279 once. Use get_narrowest_mode instead of GET_CLASS_NARROWEST_MODE.
3281 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3282 Alan Hayward <alan.hayward@arm.com>
3283 David Sherwood <david.sherwood@arm.com>
3285 * machmode.h (mode_traits): New structure.
3286 (get_narrowest_mode): New function.
3287 (mode_iterator::start): Likewise.
3288 (mode_iterator::iterate_p): Likewise.
3289 (mode_iterator::get_wider): Likewise.
3290 (mode_iterator::get_known_wider): Likewise.
3291 (mode_iterator::get_2xwider): Likewise.
3292 (FOR_EACH_MODE_IN_CLASS): New mode iterator.
3293 (FOR_EACH_MODE): Likewise.
3294 (FOR_EACH_MODE_FROM): Likewise.
3295 (FOR_EACH_MODE_UNTIL): Likewise.
3296 (FOR_EACH_WIDER_MODE): Likewise.
3297 (FOR_EACH_2XWIDER_MODE): Likewise.
3298 * builtins.c (expand_builtin_strlen): Use new mode iterators.
3299 * combine.c (simplify_comparison): Likewise
3300 * config/i386/i386.c (type_natural_mode): Likewise.
3301 * cse.c (cse_insn): Likewise.
3302 * dse.c (find_shift_sequence): Likewise.
3303 * emit-rtl.c (init_derived_machine_modes): Likewise.
3304 (init_emit_once): Likewise.
3305 * explow.c (hard_function_value): Likewise.
3306 * expmed.c (extract_fixed_bit_field_1): Likewise.
3307 (extract_bit_field_1): Likewise.
3308 (expand_divmod): Likewise.
3309 (emit_store_flag_1): Likewise.
3310 * expr.c (init_expr_target): Likewise.
3311 (convert_move): Likewise.
3312 (alignment_for_piecewise_move): Likewise.
3313 (widest_int_mode_for_size): Likewise.
3314 (emit_block_move_via_movmem): Likewise.
3315 (copy_blkmode_to_reg): Likewise.
3316 (set_storage_via_setmem): Likewise.
3317 (compress_float_constant): Likewise.
3318 * omp-low.c (omp_clause_aligned_alignment): Likewise.
3319 * optabs-query.c (get_best_extraction_insn): Likewise.
3320 * optabs.c (expand_binop): Likewise.
3321 (expand_twoval_unop): Likewise.
3322 (expand_twoval_binop): Likewise.
3323 (widen_leading): Likewise.
3324 (widen_bswap): Likewise.
3325 (expand_parity): Likewise.
3326 (expand_unop): Likewise.
3327 (prepare_cmp_insn): Likewise.
3328 (prepare_float_lib_cmp): Likewise.
3329 (expand_float): Likewise.
3330 (expand_fix): Likewise.
3331 (expand_sfix_optab): Likewise.
3332 * postreload.c (move2add_use_add2_insn): Likewise.
3333 * reg-stack.c (reg_to_stack): Likewise.
3334 * reginfo.c (choose_hard_reg_mode): Likewise.
3335 * rtlanal.c (init_num_sign_bit_copies_in_rep): Likewise.
3336 * stor-layout.c (mode_for_size): Likewise.
3337 (smallest_mode_for_size): Likewise.
3338 (mode_for_vector): Likewise.
3339 (finish_bitfield_representative): Likewise.
3340 * tree-ssa-math-opts.c (target_supports_divmod_p): Likewise.
3341 * tree-vect-generic.c (type_for_widest_vector_mode): Likewise.
3342 * tree-vect-stmts.c (vectorizable_conversion): Likewise.
3343 * var-tracking.c (prepare_call_arguments): Likewise.
3345 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3346 Alan Hayward <alan.hayward@arm.com>
3347 David Sherwood <david.sherwood@arm.com>
3349 * genconditions.c (write_header): Add a "#define USE_ENUM_MODES".
3350 * genmodes.c (emit_insn_modes_h): Define FOOmode to E_FOOmode if
3351 USE_ENUM_MODES is defined and to ((void) 0, E_FOOmode) otherwise.
3352 * machmode.h (mode_size): Move earlier in file.
3353 (mode_precision): Likewise.
3354 (mode_inner): Likewise.
3355 (mode_nunits): Likewise.
3356 (mode_unit_size): Likewise.
3357 (unit_unit_precision): Likewise.
3358 (mode_wider): Likewise.
3359 (mode_2xwider): Likewise.
3360 (machine_mode): New class.
3361 (mode_to_bytes): New function.
3362 (mode_to_bits): Likewise.
3363 (mode_to_precision): Likewise.
3364 (mode_to_inner): Likewise.
3365 (mode_to_unit_size): Likewise.
3366 (mode_to_unit_precision): Likewise.
3367 (mode_to_nunits): Likewise.
3368 (GET_MODE_SIZE): Use mode_to_bytes.
3369 (GET_MODE_BITSIZE): Use mode_to_bits.
3370 (GET_MODE_PRECISION): Use mode_to_precision.
3371 (GET_MODE_INNER): Use mode_to_inner.
3372 (GET_MODE_UNIT_SIZE): Use mode_to_unit_size.
3373 (GET_MODE_UNIT_PRECISION): Use mode_to_unit_precision.
3374 (GET_MODE_NUNITS): Use mode_to_nunits.
3375 * system.h (ALWAYS_INLINE): New macro.
3376 * config/powerpcspe/powerpcspe-c.c
3377 (altivec_resolve_overloaded_builtin): Use machine_mode instead of
3378 int for arg1_mode and arg2_mode.
3380 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3381 Alan Hayward <alan.hayward@arm.com>
3382 David Sherwood <david.sherwood@arm.com>
3384 * config/aarch64/aarch64-builtins.c (aarch64_simd_builtin_std_type):
3385 Prefix mode names with E_ in case statements.
3386 * config/aarch64/aarch64-elf.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3387 * config/aarch64/aarch64.c (aarch64_split_simd_combine): Likewise.
3388 (aarch64_split_simd_move): Likewise.
3389 (aarch64_gen_storewb_pair): Likewise.
3390 (aarch64_gen_loadwb_pair): Likewise.
3391 (aarch64_gen_store_pair): Likewise.
3392 (aarch64_gen_load_pair): Likewise.
3393 (aarch64_get_condition_code_1): Likewise.
3394 (aarch64_constant_pool_reload_icode): Likewise.
3395 (get_rsqrte_type): Likewise.
3396 (get_rsqrts_type): Likewise.
3397 (get_recpe_type): Likewise.
3398 (get_recps_type): Likewise.
3399 (aarch64_gimplify_va_arg_expr): Likewise.
3400 (aarch64_simd_container_mode): Likewise.
3401 (aarch64_emit_load_exclusive): Likewise.
3402 (aarch64_emit_store_exclusive): Likewise.
3403 (aarch64_expand_compare_and_swap): Likewise.
3404 (aarch64_gen_atomic_cas): Likewise.
3405 (aarch64_emit_bic): Likewise.
3406 (aarch64_emit_atomic_swap): Likewise.
3407 (aarch64_emit_atomic_load_op): Likewise.
3408 (aarch64_evpc_trn): Likewise.
3409 (aarch64_evpc_uzp): Likewise.
3410 (aarch64_evpc_zip): Likewise.
3411 (aarch64_evpc_ext): Likewise.
3412 (aarch64_evpc_rev): Likewise.
3413 (aarch64_evpc_dup): Likewise.
3414 (aarch64_gen_ccmp_first): Likewise.
3415 (aarch64_gen_ccmp_next): Likewise.
3416 * config/alpha/alpha.c (alpha_scalar_mode_supported_p): Likewise.
3417 (alpha_emit_xfloating_libcall): Likewise.
3418 (emit_insxl): Likewise.
3419 (alpha_arg_type): Likewise.
3420 * config/arc/arc.c (arc_vector_mode_supported_p): Likewise.
3421 (arc_preferred_simd_mode): Likewise.
3422 (arc_secondary_reload): Likewise.
3423 (get_arc_condition_code): Likewise.
3424 (arc_print_operand): Likewise.
3425 (arc_legitimate_constant_p): Likewise.
3426 * config/arc/arc.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3427 * config/arc/arc.md (casesi_load): Likewise.
3428 (casesi_compact_jump): Likewise.
3429 * config/arc/predicates.md (proper_comparison_operator): Likewise.
3430 (cc_use_register): Likewise.
3431 * config/arm/aout.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3432 * config/arm/arm-builtins.c (arm_simd_builtin_std_type): Likewise.
3433 (arm_init_iwmmxt_builtins): Likewise.
3434 * config/arm/arm.c (thumb1_size_rtx_costs): Likewise.
3435 (neon_expand_vector_init): Likewise.
3436 (arm_attr_length_move_neon): Likewise.
3437 (maybe_get_arm_condition_code): Likewise.
3438 (arm_emit_vector_const): Likewise.
3439 (arm_preferred_simd_mode): Likewise.
3440 (arm_output_iwmmxt_tinsr): Likewise.
3441 (thumb1_output_casesi): Likewise.
3442 (thumb2_output_casesi): Likewise.
3443 (arm_emit_load_exclusive): Likewise.
3444 (arm_emit_store_exclusive): Likewise.
3445 (arm_expand_compare_and_swap): Likewise.
3446 (arm_evpc_neon_vuzp): Likewise.
3447 (arm_evpc_neon_vzip): Likewise.
3448 (arm_evpc_neon_vrev): Likewise.
3449 (arm_evpc_neon_vtrn): Likewise.
3450 (arm_evpc_neon_vext): Likewise.
3451 (arm_validize_comparison): Likewise.
3452 * config/arm/neon.md (neon_vc<cmp_op><mode>): Likewise.
3453 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Likewise.
3454 * config/avr/avr.c (avr_rtx_costs_1): Likewise.
3455 * config/c6x/c6x.c (c6x_vector_mode_supported_p): Likewise.
3456 (c6x_preferred_simd_mode): Likewise.
3457 * config/epiphany/epiphany.c (get_epiphany_condition_code): Likewise.
3458 (epiphany_rtx_costs): Likewise.
3459 * config/epiphany/predicates.md (proper_comparison_operator):
3461 * config/frv/frv.c (condexec_memory_operand): Likewise.
3462 (frv_emit_move): Likewise.
3463 (output_move_single): Likewise.
3464 (output_condmove_single): Likewise.
3465 (frv_hard_regno_mode_ok): Likewise.
3466 (frv_matching_accg_mode): Likewise.
3467 * config/h8300/h8300.c (split_adds_subs): Likewise.
3468 (h8300_rtx_costs): Likewise.
3469 (h8300_print_operand): Likewise.
3470 (compute_mov_length): Likewise.
3471 (output_logical_op): Likewise.
3472 (compute_logical_op_length): Likewise.
3473 (compute_logical_op_cc): Likewise.
3474 (h8300_shift_needs_scratch_p): Likewise.
3475 (output_a_shift): Likewise.
3476 (compute_a_shift_length): Likewise.
3477 (compute_a_shift_cc): Likewise.
3478 (expand_a_rotate): Likewise.
3479 (output_a_rotate): Likewise.
3480 * config/i386/i386.c (classify_argument): Likewise.
3481 (function_arg_advance_32): Likewise.
3482 (function_arg_32): Likewise.
3483 (function_arg_64): Likewise.
3484 (function_value_64): Likewise.
3485 (ix86_gimplify_va_arg): Likewise.
3486 (ix86_legitimate_constant_p): Likewise.
3487 (put_condition_code): Likewise.
3488 (split_double_mode): Likewise.
3489 (ix86_avx256_split_vector_move_misalign): Likewise.
3490 (ix86_expand_vector_logical_operator): Likewise.
3491 (ix86_split_idivmod): Likewise.
3492 (ix86_expand_adjust_ufix_to_sfix_si): Likewise.
3493 (ix86_build_const_vector): Likewise.
3494 (ix86_build_signbit_mask): Likewise.
3495 (ix86_match_ccmode): Likewise.
3496 (ix86_cc_modes_compatible): Likewise.
3497 (ix86_expand_branch): Likewise.
3498 (ix86_expand_sse_cmp): Likewise.
3499 (ix86_expand_sse_movcc): Likewise.
3500 (ix86_expand_int_sse_cmp): Likewise.
3501 (ix86_expand_vec_perm_vpermi2): Likewise.
3502 (ix86_expand_vec_perm): Likewise.
3503 (ix86_expand_sse_unpack): Likewise.
3504 (ix86_expand_int_addcc): Likewise.
3505 (ix86_split_to_parts): Likewise.
3506 (ix86_vectorize_builtin_gather): Likewise.
3507 (ix86_vectorize_builtin_scatter): Likewise.
3508 (avx_vpermilp_parallel): Likewise.
3509 (inline_memory_move_cost): Likewise.
3510 (ix86_tieable_integer_mode_p): Likewise.
3511 (x86_maybe_negate_const_int): Likewise.
3512 (ix86_expand_vector_init_duplicate): Likewise.
3513 (ix86_expand_vector_init_one_nonzero): Likewise.
3514 (ix86_expand_vector_init_one_var): Likewise.
3515 (ix86_expand_vector_init_concat): Likewise.
3516 (ix86_expand_vector_init_interleave): Likewise.
3517 (ix86_expand_vector_init_general): Likewise.
3518 (ix86_expand_vector_set): Likewise.
3519 (ix86_expand_vector_extract): Likewise.
3520 (emit_reduc_half): Likewise.
3521 (ix86_emit_i387_round): Likewise.
3522 (ix86_mangle_type): Likewise.
3523 (ix86_expand_round_sse4): Likewise.
3524 (expand_vec_perm_blend): Likewise.
3525 (canonicalize_vector_int_perm): Likewise.
3526 (ix86_expand_vec_one_operand_perm_avx512): Likewise.
3527 (expand_vec_perm_1): Likewise.
3528 (expand_vec_perm_interleave3): Likewise.
3529 (expand_vec_perm_even_odd_pack): Likewise.
3530 (expand_vec_perm_even_odd_1): Likewise.
3531 (expand_vec_perm_broadcast_1): Likewise.
3532 (ix86_vectorize_vec_perm_const_ok): Likewise.
3533 (ix86_expand_vecop_qihi): Likewise.
3534 (ix86_expand_mul_widen_hilo): Likewise.
3535 (ix86_expand_sse2_abs): Likewise.
3536 (ix86_expand_pextr): Likewise.
3537 (ix86_expand_pinsr): Likewise.
3538 (ix86_preferred_simd_mode): Likewise.
3539 (ix86_simd_clone_compute_vecsize_and_simdlen): Likewise.
3540 * config/i386/sse.md (*andnot<mode>3): Likewise.
3541 (<mask_codefor><code><mode>3<mask_name>): Likewise.
3542 (*<code><mode>3): Likewise.
3543 * config/ia64/ia64.c (ia64_expand_vecint_compare): Likewise.
3544 (ia64_expand_atomic_op): Likewise.
3545 (ia64_arg_type): Likewise.
3546 (ia64_mode_to_int): Likewise.
3547 (ia64_scalar_mode_supported_p): Likewise.
3548 (ia64_vector_mode_supported_p): Likewise.
3549 (expand_vec_perm_broadcast): Likewise.
3550 * config/iq2000/iq2000.c (iq2000_move_1word): Likewise.
3551 (iq2000_function_arg_advance): Likewise.
3552 (iq2000_function_arg): Likewise.
3553 * config/m32c/m32c.c (m32c_preferred_reload_class): Likewise.
3554 * config/m68k/m68k.c (output_dbcc_and_branch): Likewise.
3555 (m68k_libcall_value): Likewise.
3556 (m68k_function_value): Likewise.
3557 (sched_attr_op_type): Likewise.
3558 * config/mcore/mcore.c (mcore_output_move): Likewise.
3559 * config/microblaze/microblaze.c (microblaze_function_arg_advance):
3561 (microblaze_function_arg): Likewise.
3562 * config/mips/mips.c (mips16_build_call_stub): Likewise.
3563 (mips_print_operand): Likewise.
3564 (mips_mode_ok_for_mov_fmt_p): Likewise.
3565 (mips_vector_mode_supported_p): Likewise.
3566 (mips_preferred_simd_mode): Likewise.
3567 (mips_expand_vpc_loongson_even_odd): Likewise.
3568 (mips_expand_vec_unpack): Likewise.
3569 (mips_expand_vi_broadcast): Likewise.
3570 (mips_expand_vector_init): Likewise.
3571 (mips_expand_vec_reduc): Likewise.
3572 (mips_expand_msa_cmp): Likewise.
3573 * config/mips/mips.md (casesi_internal_mips16_<mode>): Likewise.
3574 * config/mn10300/mn10300.c (mn10300_print_operand): Likewise.
3575 (cc_flags_for_mode): Likewise.
3576 * config/msp430/msp430.c (msp430_print_operand): Likewise.
3577 * config/nds32/nds32-md-auxiliary.c (nds32_mem_format): Likewise.
3578 (nds32_output_casesi_pc_relative): Likewise.
3579 * config/nds32/nds32.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3580 * config/nvptx/nvptx.c (nvptx_ptx_type_from_mode): Likewise.
3581 (nvptx_gen_unpack): Likewise.
3582 (nvptx_gen_pack): Likewise.
3583 (nvptx_gen_shuffle): Likewise.
3584 (nvptx_gen_wcast): Likewise.
3585 (nvptx_preferred_simd_mode): Likewise.
3586 * config/pa/pa.c (pa_secondary_reload): Likewise.
3587 * config/pa/predicates.md (base14_operand): Likewise.
3588 * config/powerpcspe/powerpcspe-c.c
3589 (altivec_resolve_overloaded_builtin): Likewise.
3590 * config/powerpcspe/powerpcspe.c (rs6000_setup_reg_addr_masks):
3592 (rs6000_preferred_simd_mode): Likewise.
3593 (output_vec_const_move): Likewise.
3594 (rs6000_expand_vector_extract): Likewise.
3595 (rs6000_split_vec_extract_var): Likewise.
3596 (reg_offset_addressing_ok_p): Likewise.
3597 (rs6000_legitimate_offset_address_p): Likewise.
3598 (rs6000_legitimize_address): Likewise.
3599 (rs6000_emit_set_const): Likewise.
3600 (rs6000_const_vec): Likewise.
3601 (rs6000_emit_move): Likewise.
3602 (spe_build_register_parallel): Likewise.
3603 (rs6000_darwin64_record_arg_recurse): Likewise.
3604 (swap_selector_for_mode): Likewise.
3605 (spe_init_builtins): Likewise.
3606 (paired_init_builtins): Likewise.
3607 (altivec_init_builtins): Likewise.
3608 (do_load_for_compare): Likewise.
3609 (rs6000_generate_compare): Likewise.
3610 (rs6000_expand_float128_convert): Likewise.
3611 (emit_load_locked): Likewise.
3612 (emit_store_conditional): Likewise.
3613 (rs6000_output_function_epilogue): Likewise.
3614 (rs6000_handle_altivec_attribute): Likewise.
3615 (rs6000_function_value): Likewise.
3616 (emit_fusion_gpr_load): Likewise.
3617 (emit_fusion_p9_load): Likewise.
3618 (emit_fusion_p9_store): Likewise.
3619 * config/powerpcspe/predicates.md (easy_fp_constant): Likewise.
3620 (fusion_gpr_mem_load): Likewise.
3621 (fusion_addis_mem_combo_load): Likewise.
3622 (fusion_addis_mem_combo_store): Likewise.
3623 * config/rs6000/predicates.md (easy_fp_constant): Likewise.
3624 (fusion_gpr_mem_load): Likewise.
3625 (fusion_addis_mem_combo_load): Likewise.
3626 (fusion_addis_mem_combo_store): Likewise.
3627 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
3629 * config/rs6000/rs6000-string.c (do_load_for_compare): Likewise.
3630 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Likewise.
3631 (rs6000_preferred_simd_mode): Likewise.
3632 (output_vec_const_move): Likewise.
3633 (rs6000_expand_vector_extract): Likewise.
3634 (rs6000_split_vec_extract_var): Likewise.
3635 (reg_offset_addressing_ok_p): Likewise.
3636 (rs6000_legitimate_offset_address_p): Likewise.
3637 (rs6000_legitimize_address): Likewise.
3638 (rs6000_emit_set_const): Likewise.
3639 (rs6000_const_vec): Likewise.
3640 (rs6000_emit_move): Likewise.
3641 (rs6000_darwin64_record_arg_recurse): Likewise.
3642 (swap_selector_for_mode): Likewise.
3643 (paired_init_builtins): Likewise.
3644 (altivec_init_builtins): Likewise.
3645 (rs6000_expand_float128_convert): Likewise.
3646 (emit_load_locked): Likewise.
3647 (emit_store_conditional): Likewise.
3648 (rs6000_output_function_epilogue): Likewise.
3649 (rs6000_handle_altivec_attribute): Likewise.
3650 (rs6000_function_value): Likewise.
3651 (emit_fusion_gpr_load): Likewise.
3652 (emit_fusion_p9_load): Likewise.
3653 (emit_fusion_p9_store): Likewise.
3654 * config/rx/rx.c (rx_gen_move_template): Likewise.
3655 (flags_from_mode): Likewise.
3656 * config/s390/predicates.md (s390_alc_comparison): Likewise.
3657 (s390_slb_comparison): Likewise.
3658 * config/s390/s390.c (s390_handle_vectorbool_attribute): Likewise.
3659 (s390_vector_mode_supported_p): Likewise.
3660 (s390_cc_modes_compatible): Likewise.
3661 (s390_match_ccmode_set): Likewise.
3662 (s390_canonicalize_comparison): Likewise.
3663 (s390_emit_compare_and_swap): Likewise.
3664 (s390_branch_condition_mask): Likewise.
3665 (s390_rtx_costs): Likewise.
3666 (s390_secondary_reload): Likewise.
3667 (__SECONDARY_RELOAD_CASE): Likewise.
3668 (s390_expand_cs): Likewise.
3669 (s390_preferred_simd_mode): Likewise.
3670 * config/s390/vx-builtins.md (vec_packsu_u<mode>): Likewise.
3671 * config/sh/sh.c (sh_print_operand): Likewise.
3672 (dump_table): Likewise.
3673 (sh_secondary_reload): Likewise.
3674 * config/sh/sh.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3675 * config/sh/sh.md (casesi_worker_1): Likewise.
3676 (casesi_worker_2): Likewise.
3677 * config/sparc/predicates.md (icc_comparison_operator): Likewise.
3678 (fcc_comparison_operator): Likewise.
3679 * config/sparc/sparc.c (sparc_expand_move): Likewise.
3680 (emit_soft_tfmode_cvt): Likewise.
3681 (sparc_preferred_simd_mode): Likewise.
3682 (output_cbranch): Likewise.
3683 (sparc_print_operand): Likewise.
3684 (sparc_expand_vec_perm_bmask): Likewise.
3685 (vector_init_bshuffle): Likewise.
3686 * config/spu/spu.c (spu_scalar_mode_supported_p): Likewise.
3687 (spu_vector_mode_supported_p): Likewise.
3688 (spu_expand_insv): Likewise.
3689 (spu_emit_branch_or_set): Likewise.
3690 (spu_handle_vector_attribute): Likewise.
3691 (spu_builtin_splats): Likewise.
3692 (spu_builtin_extract): Likewise.
3693 (spu_builtin_promote): Likewise.
3694 (spu_expand_sign_extend): Likewise.
3695 * config/tilegx/tilegx.c (tilegx_scalar_mode_supported_p): Likewise.
3696 (tilegx_simd_int): Likewise.
3697 * config/tilepro/tilepro.c (tilepro_scalar_mode_supported_p): Likewise.
3698 (tilepro_simd_int): Likewise.
3699 * config/v850/v850.c (const_double_split): Likewise.
3700 (v850_print_operand): Likewise.
3701 (ep_memory_offset): Likewise.
3702 * config/vax/vax.c (vax_rtx_costs): Likewise.
3703 (vax_output_int_move): Likewise.
3704 (vax_output_int_add): Likewise.
3705 (vax_output_int_subtract): Likewise.
3706 * config/visium/predicates.md (visium_branch_operator): Likewise.
3707 * config/visium/visium.c (rtx_ok_for_offset_p): Likewise.
3708 (visium_print_operand_address): Likewise.
3709 * config/visium/visium.h (ASM_OUTPUT_ADDR_DIFF_ELT): Likewise.
3710 * config/xtensa/xtensa.c (xtensa_mem_offset): Likewise.
3711 (xtensa_expand_conditional_branch): Likewise.
3712 (xtensa_copy_incoming_a7): Likewise.
3713 (xtensa_output_literal): Likewise.
3714 * dfp.c (decimal_real_maxval): Likewise.
3715 * targhooks.c (default_libgcc_floating_mode_supported_p): Likewise.
3717 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3718 Alan Hayward <alan.hayward@arm.com>
3719 David Sherwood <david.sherwood@arm.com>
3721 * genmodes.c (mode_size_inline): Add an E_ prefix to mode names.
3722 (mode_nunits_inline): Likewise.
3723 (mode_inner_inline): Likewise.
3724 (mode_unit_size_inline): Likewise.
3725 (mode_unit_precision_inline): Likewise.
3726 (emit_insn_modes_h): Likewise. Also emit a #define of the
3728 (emit_mode_wider): Add an E_ prefix to mode names.
3729 (emit_mode_complex): Likewise.
3730 (emit_mode_inner): Likewise.
3731 (emit_mode_adjustments): Likewise.
3732 (emit_mode_int_n): Likewise.
3733 * config/aarch64/aarch64-builtins.c (v8qi_UP, v4hi_UP, v4hf_UP)
3734 (v2si_UP, v2sf_UP, v1df_UP, di_UP, df_UP, v16qi_UP, v8hi_UP, v8hf_UP)
3735 (v4si_UP, v4sf_UP, v2di_UP, v2df_UP, ti_UP, oi_UP, ci_UP, xi_UP)
3736 (si_UP, sf_UP, hi_UP, hf_UP, qi_UP): Likewise.
3737 (CRC32_BUILTIN, ENTRY): Likewise.
3738 * config/aarch64/aarch64.c (aarch64_push_regs): Likewise.
3739 (aarch64_pop_regs): Likewise.
3740 (aarch64_process_components): Likewise.
3741 * config/alpha/alpha.c (alpha_emit_conditional_move): Likewise.
3742 * config/arm/arm-builtins.c (v8qi_UP, v4hi_UP, v4hf_UP, v2si_UP)
3743 (v2sf_UP, di_UP, v16qi_UP, v8hi_UP, v8hf_UP, v4si_UP, v4sf_UP)
3744 (v2di_UP, ti_UP, ei_UP, oi_UP, hf_UP, si_UP, void_UP): Likewise.
3745 * config/arm/arm.c (arm_init_libfuncs): Likewise.
3746 * config/i386/i386-builtin-types.awk (ix86_builtin_type_vect_mode):
3748 * config/i386/i386-builtin.def (pcmpestr): Likewise.
3749 (pcmpistr): Likewise.
3750 * config/microblaze/microblaze.c (double_memory_operand): Likewise.
3751 * config/mmix/mmix.c (mmix_output_condition): Likewise.
3752 * config/powerpcspe/powerpcspe.c (rs6000_init_hard_regno_mode_ok):
3754 * config/rl78/rl78.c (mduc_regs): Likewise.
3755 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Likewise.
3756 (htm_expand_builtin): Likewise.
3757 * config/sh/sh.h (REGISTER_NATURAL_MODE): Likewise.
3758 * config/sparc/sparc.c (emit_save_or_restore_regs): Likewise.
3759 * config/xtensa/xtensa.c (print_operand): Likewise.
3760 * expmed.h (NUM_MODE_PARTIAL_INT): Likewise.
3761 (NUM_MODE_VECTOR_INT): Likewise.
3762 * genoutput.c (null_operand): Likewise.
3763 (output_operand_data): Likewise.
3764 * genrecog.c (print_parameter_value): Likewise.
3765 * lra.c (debug_operand_data): Likewise.
3767 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3769 * dwarf2out.c (scompare_loc_descriptor_wide)
3770 (scompare_loc_descriptor_narrow): New functions, split out from...
3771 (scompare_loc_descriptor): ...here.
3772 * expmed.c (emit_store_flag_int): New function, split out from...
3773 (emit_store_flag): ...here.
3775 2017-08-30 Richard Biener <rguenther@suse.de>
3777 * dwarf2out.c (dwarf2out_finish): Remove setting AT_pubnames.
3778 (dwarf2out_early_finish): Move setting of AT_pubnames from
3779 early debug output to early finish.
3781 2017-08-30 Jozef Lawrynowicz <jozef.l@somniumtech.com>
3783 * gcc/config/msp430/msp430.h: Pass -mcode/data-region to the linker
3784 and -mdata-region to the assembler.
3786 2017-08-30 Richard Biener <rguenther@suse.de>
3788 * dwarf2out.c (add_dwarf_attr): Check we don't add duplicate
3790 (gen_subprogram_die): Add DW_AT_object_pointer only early.
3791 (dwarf2out_early_global_decl): Only generate a DIE for the
3792 abstract origin if it doesn't already exist or is a declaration DIE.
3793 (resolve_addr): Do not add the linkage name twice when
3794 generating a stub DIE for the DW_TAG_GNU_call_site target.
3796 2017-08-30 Richard Sandiford <richard.sandiford@linaro.org>
3798 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
3799 Use machine_mode rather than int for arg1_mode.
3801 2017-08-29 Michael Meissner <meissner@linux.vnet.ibm.com>
3804 * config/rs6000/rs6000.c (rs6000_expand_binop_builtin): Insure
3805 that the second argument of the built-in functions to unpack
3806 128-bit scalar types to 64-bit values is 0 or 1. Change to use a
3807 switch statement instead a lot of if statements.
3808 * config/rs6000/rs6000.md (unpack<mode>, FMOVE128_VSX iterator):
3809 Allow 64-bit values to be in Altivec registers as well as
3810 traditional floating point registers.
3811 (pack<mode>, FMOVE128_VSX iterator): Likewise.
3813 2017-08-29 Alexander Monakov <amonakov@ispras.ru>
3815 * ira-costs.c (record_address_regs): Handle both operands of PLUS for
3816 MAX_REGS_PER_ADDRESS == 1.
3818 2017-08-29 Uros Bizjak <ubizjak@gmail.com>
3820 * config/i386/i386.opt (flag_fentry): Do not init to -1.
3821 * config/i386/i386.c (ix86_option_override_internal): Simplify
3822 setting of opts->x_flag_entry.
3824 2017-08-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3825 Jakub Jelinek <jakub@redhat.com>
3826 Richard Biener <rguenther@suse.de>
3828 PR tree-optimization/81503
3829 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Ensure
3830 folded constant fits in the target type; reorder tests for clarity.
3832 2017-08-29 Martin Liska <mliska@suse.cz>
3834 * passes.def: Include pass_lower_switch.
3835 * stmt.c (dump_case_nodes): Remove and move to
3836 tree-switch-conversion.
3837 (case_values_threshold): Likewise.
3838 (expand_switch_as_decision_tree_p): Likewise.
3839 (emit_case_decision_tree): Likewise.
3840 (expand_case): Likewise.
3841 (balance_case_nodes): Likewise.
3842 (node_has_low_bound): Likewise.
3843 (node_has_high_bound): Likewise.
3844 (node_is_bounded): Likewise.
3845 (emit_case_nodes): Likewise.
3846 (struct simple_case_node): New struct.
3847 (add_case_node): Remove.
3848 (emit_case_dispatch_table): Use vector instead of case_list.
3849 (reset_out_edges_aux): Remove.
3850 (compute_cases_per_edge): Likewise.
3851 (expand_case): Build list of simple_case_node.
3852 (expand_sjlj_dispatch_table): Use it.
3853 * tree-switch-conversion.c (struct case_node): Moved from
3854 stmt.c and adjusted.
3855 (emit_case_nodes): Likewise.
3856 (node_has_low_bound): Likewise.
3857 (node_has_high_bound): Likewise.
3858 (node_is_bounded): Likewise.
3859 (case_values_threshold): Likewise.
3860 (reset_out_edges_aux): Likewise.
3861 (compute_cases_per_edge): Likewise.
3862 (add_case_node): Likewise.
3863 (dump_case_nodes): Likewise.
3864 (balance_case_nodes): Likewise.
3865 (expand_switch_as_decision_tree_p): Likewise.
3866 (emit_jump): Likewise.
3867 (emit_case_decision_tree): Likewise.
3868 (try_switch_expansion): Likewise.
3869 (do_jump_if_equal): Likewise.
3870 (emit_cmp_and_jump_insns): Likewise.
3871 (fix_phi_operands_for_edge): New function.
3872 (record_phi_operand_mapping): Likewise.
3873 (class pass_lower_switch): New pass.
3874 (pass_lower_switch::execute): New function.
3875 (make_pass_lower_switch): Likewise.
3876 (conditional_probability):
3877 * timevar.def: Add TV_TREE_SWITCH_LOWERING.
3878 * tree-pass.h: Add make_pass_lower_switch.
3880 2017-08-29 Jozef Lawrynowicz <jozef.l@somniumtech.com>
3883 * gcc/config/msp430/msp430.c (msp430_attr): Mark interrupt
3886 2017-08-29 Richard Biener <rguenther@suse.de>
3888 * dwarf2out.c (add_dwarf_attr): When checking is enabled verify
3889 we do not add a DW_AT_inline attribute twice.
3890 (gen_subprogram_die): Remove code setting DW_AT_inline on
3891 DECL_ABSTRACT_P nodes.
3893 2017-08-29 Richard Sandiford <richard.sandiford@linaro.org>
3895 * gimplify.c (gimplify_call_expr): Copy the nothrow flag to
3896 calls to internal functions.
3897 (gimplify_modify_expr): Likewise.
3898 * tree-call-cdce.c (use_internal_fn): Likewise.
3899 * tree-ssa-math-opts.c (pass_cse_reciprocals::execute): Likewise.
3900 (convert_to_divmod): Set the nothrow flag.
3901 * tree-if-conv.c (predicate_mem_writes): Likewise.
3902 * tree-vect-stmts.c (vectorizable_mask_load_store): Likewise.
3903 (vectorizable_call): Likewise.
3904 (vectorizable_store): Likewise.
3905 (vectorizable_load): Likewise.
3906 * tree-vect-patterns.c (vect_recog_pow_pattern): Likewise.
3907 (vect_recog_mask_conversion_pattern): Likewise.
3909 2017-08-29 Martin Liska <mliska@suse.cz>
3912 * gcc.c (driver_handle_option): Add new argument.
3913 * opts-common.c (handle_option): Pass
3914 target_option_override_hook.
3915 * opts-global.c (lang_handle_option): Add new option.
3916 (set_default_handlers): Add new argument.
3917 (decode_options): Likewise.
3918 * opts.c (target_handle_option): Likewise.
3919 (common_handle_option): Call target_option_override_hook.
3920 * opts.h (struct cl_option_handler_func): Add hook for
3921 target option override.
3922 (struct cl_option_handlers): Likewise.
3923 (set_default_handlers): Add new argument.
3924 (decode_options): Likewise.
3925 (common_handle_option): Likewise.
3926 (target_handle_option): Likewise.
3927 * toplev.c (toplev::main): Pass targetm.target_option.override
3930 2017-08-29 Richard Biener <rguenther@suse.de>
3931 Dominik Infuehr <dominik.infuehr@theobroma-systems.com>
3933 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Properly confine
3934 life to the active subtree.
3936 2017-08-28 Jeff Law <law@redhat.com>
3938 * tree-ssa-dom.c (edge_info::record_simple_equiv): Call
3939 derive_equivalences.
3940 (derive_equivalences_from_bit_ior, record_temporary_equivalences):
3942 (edge_info::derive_equivalences): New private member function
3944 * tree-ssa-dom.c (class edge_info): Changed from a struct
3945 to a class. Add ctor/dtor, methods and data members.
3946 (edge_info::edge_info): Renamed from allocate_edge_info.
3947 Initialize additional members.
3948 (edge_info::~edge_info): New.
3949 (free_dom_edge_info): Delete the edge info.
3950 (record_edge_info): Use new class & associated member functions.
3951 Tighten forms for testing for edge equivalences.
3952 (record_temporary_equivalences): Iterate over the simple
3953 equivalences rather than assuming there's only one per edge.
3954 (cprop_into_successor_phis): Iterate over the simple
3955 equivalences rather than assuming there's only one per edge.
3956 (optimize_stmt): Use operand_equal_p rather than pointer
3957 equality for mini-DSE code.
3959 2017-08-28 Nathan Sidwell <nathan@acm.org>
3961 * gcc.c (execute): Fold SIGPIPE handling into switch
3962 statement. Adjust internal error message.
3964 2017-08-28 Richard Biener <rguenther@suse.de>
3967 * dwarf2out.c (gen_remaining_tmpl_value_param_die_attributes):
3968 Do nothing for removed DIEs.
3970 2017-08-28 Richard Biener <rguenther@suse.de>
3972 PR tree-optimization/81977
3973 * tree-ssa-sccvn.c (vn_reference_lookup_3): Fix look through
3976 2017-08-28 Alexander Monakov <amonakov@ispras.ru>
3979 * doc/md.texi (mem_thread_fence): Remove mention of mode. Rewrite.
3980 * optabs.c (expand_mem_thread_fence): Emit a compiler barrier when
3981 using targetm.gen_mem_thread_fence.
3983 2017-08-27 Uros Bizjak <ubizjak@gmail.com>
3986 * config/i386/i386.md (*<btsc><mode>): Change operand 2
3987 predicate to register_operand. Reorder operands.
3988 (*btr<mode>): Ditto.
3989 (*<btsc><mode>_mask): Change operand 3 predicate to register_operand.
3990 (*btr<mode>_mask): Ditto.
3992 2017-08-25 Steven Munroe <munroesj@gcc.gnu.org>
3994 * config.gcc (powerpc*-*-*): Add xmmintrin.h and mm_malloc.h.
3995 * config/rs6000/xmmintrin.h: New file.
3996 * config/rs6000/x86intrin.h [__ALTIVEC__]: Include xmmintrin.h.
3998 2017-08-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4001 * config/rs6000/rs6000-p8swap.c (find_alignment_op): Add reference
4002 parameter and_insn and return it.
4003 (recombine_lvx_pattern): Insert a copy to ensure availability of
4004 the base register of the copied masking operation at the point of
4005 the instruction replacement.
4006 (recombine_stvx_pattern): Likewise.
4008 2017-08-25 Michael Meissner <meissner@linux.vnet.ibm.com>
4010 * config/rs6000/rs6000.opt (-mpower9-dform-scalar): Delete
4011 undocumented switches.
4012 (-mpower9-dform-vector): Likewise.
4013 (-mpower9-dform): Likewise.
4014 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Update
4015 comments to delete references to -mpower9-dform* switches.
4016 * config/rs6000/predicates.md (vsx_quad_dform_memory_operand):
4017 Delete reference to -mpower9-dform* switches, test for
4018 -mpower9-vector instead.
4019 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Likewise.
4020 (OTHER_P9_VECTOR_MASKS): Likewise.
4021 (POWERPC_MASKS): Likewise.
4022 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Change
4023 tests against -mpower9-dform* to -mpower9-vector. Delete code
4024 that checked for -mpower9-dform* consistancy with other options.
4025 Add test for -mpower9-misc to enable other power9 switches.
4026 (rs6000_init_hard_regno_mode_ok): Likewise.
4027 (rs6000_option_override_internal): Likewise.
4028 (rs6000_emit_prologue): Likewise.
4029 (rs6000_emit_epilogue): Likewise.
4030 (rs6000_opt_masks): Delete -mpower9-dform-{scalar,vector}.
4031 (rs6000_disable_incompatiable_switches): Delete -mpower9-dform.
4032 (emit_fusion_p9_load): Change tests for -mpower9-dform-scalar
4034 (emit_fusion_p9_store): Likewise.
4035 * config/rs6000/rs6000.h (TARGET_P9_DFORM_SCALAR): Delete
4036 resetting these macros if the assembler does not support ISA 3.0
4038 (TARGET_P9_DFORM_VECTOR): Likewise.
4039 * config/rs6000/rs6000.md (peepholes to optimize altivec memory):
4040 Change to use -mpower9-vector instead of -mpower9-dform-scalar.
4042 2017-08-25 Alan Modra <amodra@gmail.com>
4044 PR rtl-optimization/81747
4045 * cse.c (cse_extended_basic_block): Don't attempt to record
4046 equivalences for degenerate conditional branches that branch
4047 to their fall-through.
4049 2017-08-24 Martin Sebor <msebor@redhat.com>
4052 * gimple-fold.c (size_must_be_zero_p): New function.
4053 (gimple_fold_builtin_memory_op): Call it.
4055 2017-08-24 Steven Munroe <munroesj@gcc.gnu.org>
4057 * config/rs6000/mm_malloc.h: New file.
4059 2017-08-24 Bin Cheng <bin.cheng@arm.com>
4061 PR tree-optimization/81913
4062 * tree-ssa-loop-niter.c (number_of_iterations_cond): Skip niter
4063 analysis when either IVs in condition can wrap.
4065 2017-08-24 Uros Bizjak <ubizjak@gmail.com>
4067 * dwarf2out.c (MAX_ARTIFICIAL_LABEL_BYTES): Increase to 40.
4068 * dwarf2cfi.c (MAX_ARTIFICIAL_LABEL_BYTES): Remove.
4070 2017-08-24 Richard Biener <rguenther@suse.de>
4073 * targhooks.c (default_target_can_inline_p): Properly
4074 use target_option_default_node when no DECL_FUNCTION_SPECIFIC_TARGET
4075 is present and always compare.
4076 * config/i386/i386.c (ix86_valid_target_attribute_tree): Do not
4077 infer -mfpmath=sse from TARGET_SSE_P.
4078 (ix86_can_inline_p): Properly use target_option_default_node when
4079 no DECL_FUNCTION_SPECIFIC_TARGET is present and always compare.
4081 2017-08-24 Richard Biener <rguenther@suse.de>
4084 * dwarf2out.c (output_die): Handle flag_generate_offload like
4086 (output_comp_unit): Likewise.
4087 (gen_array_type_die): Likewise.
4088 (dwarf2out_early_finish): Likewise.
4089 (note_variable_value_in_expr): Likewise.
4090 (dwarf2out_finish): Likewise. Adjust assert.
4091 * cgraphunit.c (symbol_table::compile): Move setting of
4092 flag_generate_offload earlier ...
4093 (symbol_table::finalize_compilation_unit): ... here, before
4094 early debug finalization.
4096 2017-08-24 Richard Biener <rguenther@suse.de>
4098 * config/i386/i386.c: Include symbol-summary.h, ipa-prop.h
4099 and ipa-fnsummary.h.
4100 (ix86_can_inline_p): When ix86_fpmath flags do not match
4101 check whether the callee uses FP math at all.
4103 2017-08-24 Aldy Hernandez <aldyh@redhat.com>
4106 * tree-ssanames.c (get_nonzero_bits): Use element_precision
4107 instead of TYPE_PRECISION.
4109 2017-08-24 Richard Sandiford <richard.sandiford@linaro.org>
4110 Alan Hayward <alan.hayward@arm.com>
4111 David Sherwood <david.sherwood@arm.com>
4113 * combine.c (make_extraction): Use subreg_offset_from_lsb.
4115 2017-08-23 Daniel Santos <daniel.santos@pobox.com>
4117 * config/i386/i386.h (ix86_frame::stack_realign_allocate_offset):
4119 (ix86_frame::stack_realign_allocate): New field.
4120 (struct machine_frame_state): Modify comments.
4121 (machine_frame_state::sp_realigned_fp_end): New field.
4122 * config/i386/i386.c (ix86_compute_frame_layout): Rework stack frame
4124 (sp_valid_at): Add assertion to assure no attempt to access invalid
4125 offset of a realigned stack.
4126 (fp_valid_at): Likewise.
4127 (choose_baseaddr): Modify comments.
4128 (ix86_emit_outlined_ms2sysv_save): Adjust to changes in
4129 ix86_expand_prologue.
4130 (ix86_expand_prologue): Modify stack realignment and allocation.
4131 (ix86_expand_epilogue): Modify comments.
4132 * doc/sourcebuild.texi: Add documentation for target selectors avx2,
4133 avx2_runtime, avx512f, and avx512f_runtime.
4135 2017-08-23 Uros Bizjak <ubizjak@gmail.com>
4137 * config/i386/i386.opt: Remove unneeded Init(0) initializations.
4138 (mstackrealign): Do not init to -1.
4139 * config/i386/i386.c (ix86_option_override_internal):
4140 Check opts_set, not opts when setting default value of
4141 opts->x_ix86_force_align_arg_pointer.
4143 2017-08-23 Richard Biener <rguenther@suse.de>
4145 * function.c (fndecl_name): Use verbosity 1 (no arguments) for
4146 lang_hooks.decl_printable_name.
4147 * print-rtl-function.c (print_rtx_function): Likewise.
4148 * tree-pretty-print.c (dump_function_header): Likewise.
4150 2017-08-23 Richard Biener <rguenther@suse.de>
4153 * dwarf2out.c (dwarf2out_abstract_function): Handle LTO with
4154 -g0 at compile-time.
4156 2017-08-23 Tamar Christina <tamar.christina@arm.com>
4159 * doc/sourcebuild.texi (Other hardware attributes):
4162 2017-08-23 Tamar Christina <tamar.christina@arm.com>
4165 * tree-ssa-math-opts.c (convert_expand_mult_copysign):
4166 Fix single-use check.
4168 2017-08-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
4170 * gcc.c (execute): Only test for SIGKILL and SIGQUIT if available.
4172 2017-08-22 Daniel Santos <daniel.santos@pobox.com>
4174 * doc/install.texi: Modify to add more details on running selected
4177 2017-08-22 Daniel Santos <daniel.santos@pobox.com>
4179 * config/i386/i386.c (ix86_option_override_internal): Error when -mx32
4180 is combined with -mabi=ms.
4181 (ix86_function_type_abi): Limit errors for mixing -mx32 with attribute
4184 2017-08-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4186 PR tree-optimization/81488
4187 * gimple-ssa-strength-reduction (struct slsr_cand_d): Add visited
4188 and cached_basis fields.
4189 (MAX_SPREAD): New constant.
4190 (alloc_cand_and_find_basis): Initialize new fields.
4191 (clear_visited): New function.
4192 (create_phi_basis_1): Rename from create_phi_basis, set visited
4193 and cached_basis fields.
4194 (create_phi_basis): New wrapper function.
4195 (phi_add_costs_1): Rename from phi_add_costs, add spread
4196 parameter, set visited field, short-circuit when limits reached.
4197 (phi_add_costs): New wrapper function.
4198 (record_phi_increments_1): Rename from record_phi_increments, set
4200 (record_phi_increments): New wrapper function.
4201 (phi_incr_cost_1): Rename from phi_incr_cost, set visited field.
4202 (phi_incr_cost): New wrapper function.
4203 (all_phi_incrs_profitable_1): Rename from
4204 all_phi_incrs_profitable, set visited field.
4205 (all_phi_incrs_profitable): New wrapper function.
4207 2017-08-22 Richard Sandiford <richard.sandiford@linaro.org>
4208 Alan Hayward <alan.hayward@arm.com>
4209 David Sherwood <david.sherwood@arm.com>
4211 * rtl.h (paradoxical_subreg_p): Define inline, and add a version
4212 that takes the outer and inner modes.
4213 * doc/rtl.texi: Use paradoxical_subreg_p instead of a GET_MODE_SIZE
4214 comparison as the canonical test for a paradoxical subreg.
4215 * combine.c (simplify_set): Use paradoxical_subreg_p.
4216 (make_extraction): Likewise.
4217 (force_to_mode): Likewise.
4218 (rtx_equal_for_field_assignment_p): Likewise.
4219 (gen_lowpart_for_combine): Likewise.
4220 (simplify_comparison): Likewise.
4221 * cse.c (equiv_constant): Likewise.
4222 * expmed.c (store_bit_field_1): Likewise.
4223 * final.c (alter_subreg): Likewise.
4224 * fwprop.c (propagate_rtx): Likewise.
4225 (forward_propagate_subreg): Likewise.
4226 * ira-conflicts.c (ira_build_conflicts): Likewise.
4227 * lower-subreg.c (simplify_gen_subreg_concatn): Likewise.
4228 * lra-constraints.c (curr_insn_transform): Likewise.
4229 (split_reg): Likewise.
4230 * lra-eliminations.c (move_plus_up): Likewise.
4231 (lra_eliminate_regs_1): Likewise.
4232 * recog.c (general_operand): Likewise.
4233 * ree.c (combine_reaching_defs): Likewise.
4234 * reload.c (push_reload): Likewise.
4235 (find_reloads): Likewise.
4236 * reload1.c (elimination_effects): Likewise.
4237 (compute_reload_subreg_offset): Likewise.
4238 (choose_reload_regs): Likewise.
4239 * rtlanal.c (subreg_lsb_1): Likewise.
4240 * simplify-rtx.c (simplify_unary_operation_1): Likewise.
4241 (simplify_subreg): Likewise.
4242 * var-tracking.c (track_loc_p): Likewise.
4243 * emit-rtl.c (byte_lowpart_offset): Likewise.
4244 (paradoxical_subreg_p): Delete out-of-line definition.
4246 2017-08-22 Jeff Law <law@redhat.com>
4248 PR tree-optimization/81741
4249 PR tree-optimization/71947
4250 * tree-ssa-dom.c: Include tree-inline.h.
4251 (record_temporary_equivalences): Only record SSA_NAME = SSA_NAME
4252 equivalences if one is more expensive to compute than the other.
4253 * tree-ssa-scopedtables.h (class const_or_copies): Make
4254 record_const_or_copy_raw method private.
4255 (class avail_exprs_stack): New method simplify_binary_operation.
4256 * tree-ssa-scopedtables.c (avail_exprs_stack::lookup_avail_expr): Call
4257 avail_exprs_stack::simplify_binary_operation as needed.
4258 (avail_exprs_stack::simplify_binary_operation): New function.
4260 2017-08-22 Sebastian Huber <sebastian.huber@embedded-brains.de>
4262 * config.gcc (powerpc-*-rtems*): Add rs6000/linux64.opt.
4263 * config/rs6000/rtems.h (ASM_PREFERRED_EH_DATA_FORMAT): New define.
4264 (DOT_SYMBOLS): Likewise.
4265 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
4266 (RELOCATABLE_NEEDS_FIXUP): Likewise.
4267 (RS6000_ABI_NAME): Likewise.
4268 (TARGET_CMODEL): Likewise.
4269 (TOC_SECTION_ASM_OP): Likewise.
4270 (SET_CMODEL): New macro.
4271 (SUBSUBTARGET_OVERRIDE_OPTIONS): Evaluate cmodel options.
4273 2017-08-22 Richard Biener <rguenther@suse.de>
4275 * tree-inline.c (remap_type_1): Change asserts on TYPE_SIZE[_UNIT]
4276 to allow for free-lang-data replacements similar to verify_type_variant.
4278 2017-08-22 Richard Sandiford <richard.sandiford@linaro.org>
4279 Alan Hayward <alan.hayward@arm.com>
4280 David Sherwood <david.sherwood@arm.com>
4282 * config/aarch64/aarch64.md (casesi): Use DImode rather than
4283 VOIDmode for the LABEL_REF.
4285 2017-08-22 Richard Biener <rguenther@suse.de>
4287 * tree-cfg.c (gimple_split_edge): Avoid reallocating target PHI nodes.
4289 2017-08-22 Richard Biener <rguenther@suse.de>
4291 * common.opt (feliminate-dwarf2-dups): Ignore.
4292 * doc/invoke.texi (feliminate-dwarf2-dups): Remove documentation.
4293 * dwarf2out.c (push_new_compile_unit, pop_compile_unit,
4294 same_die_p_wrap, compute_section_prefix,
4295 is_symbol_die, assign_symbol_names, break_out_includes): Remove.
4296 (comdat_symbol_id, comdat_symbol_number): Likewise.
4297 (cu_hash_table_entry, cu_hash_table_entry_hasher, cu_hash_type):
4299 (check_duplicate_cu, record_comdat_symbol_number): Likewise.
4300 (output_die): Mark unreachable path unreachable.
4301 (dwarf2out_start_source_file): Do not create DW_TAG_GNU_BINCL.
4302 (dwarf2out_end_source_file): Do not create DW_TAG_GNU_EINCL.
4303 (dwarf2out_init): Remove code handling flag_eliminate_dwarf2_dups.
4304 (dwarf2out_early_finish): Likewise.
4306 2017-08-22 Aldy Hernandez <aldyh@redhat.com>
4308 * wide-int.h (hwi_with_prec::hwi_with_prec): Sign extend.
4310 2017-08-22 Georg-Johann Lay <avr@gjlay.de>
4313 * config/avr/avr.c (avr_handle_addr_attribute): Early return if
4314 not VAR_P. Filter attribute warnings with OPT_Wattributes.
4315 (avr_attribute_table) <io, io_low, address>: Initialize
4316 .decl_required with true.
4318 2017-08-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4320 * config/rs6000/rs6000-cpus.def (-mvsx-scalar-float): Delete
4321 undocumented debugging options.
4322 (-mvsx-scalar-double): Likewise.
4323 (-mallow-df-permute): Likewise.
4324 (-mvectorize-builtins): Likewise.
4325 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Likewise.
4326 (rs6000_builtin_vectorized_function): Likewise.
4327 (rs6000_builtin_md_vectorized_function): Likewise.
4328 (rs6000_opt_vars): Likewise.
4330 2017-08-21 Uros Bizjak <ubizjak@gmail.com>
4333 * config/i386/i386.md (*btsq_imm): Rename from *btsq.
4334 (*btrq_imm): Rename from *btrq.
4335 (*btcq_imm): Rename from *btcq.
4336 (btsc): New code attribute.
4337 (*<btsc><mode>): New insn pattern.
4338 (*btr<mode>): Ditto.
4339 (*<btsc><mode>_mask): New insn_and_split pattern.
4340 (*btr<mode>_mask): Ditto.
4342 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4343 Alan Hayward <alan.hayward@arm.com>
4344 David Sherwood <david.sherwood@arm.com>
4346 * function.c (pad_below): Simplify padding calculation.
4348 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4349 Alan Hayward <alan.hayward@arm.com>
4350 David Sherwood <david.sherwood@arm.com>
4352 * target.def (function_prologue): Remove frame size argument.
4353 (function_epilogue): Likewise.
4354 * doc/tm.texi: Regenerate.
4355 * final.c (final_start_function): Update call to function_prologue.
4356 (final_end_function): Update call to function_epilogue.
4357 (default_function_pro_epilogue): Remove frame size argument.
4358 * output.h (default_function_pro_epilogue): Likewise.
4359 * config/arm/arm.c (arm_output_function_epilogue): Likewise.
4360 (arm_output_function_prologue): Likewise.
4361 * config/frv/frv.c (frv_function_prologue): Likewise.
4362 (frv_function_epilogue): Likewise.
4363 * config/i386/i386.c (ix86_output_function_epilogue): Likewise.
4364 * config/ia64/ia64.c (ia64_output_function_prologue): Likewise.
4365 (ia64_output_function_epilogue): Likewise.
4366 * config/m32r/m32r.c (m32r_output_function_prologue): Likewise.
4367 (m32r_output_function_epilogue): Likewise.
4368 * config/microblaze/microblaze.c (microblaze_function_prologue)
4369 (microblaze_function_epilogue): Likewise.
4370 * config/mips/mips.c (mips_output_function_prologue): Likewise.
4371 (mips_output_function_epilogue): Likewise.
4372 * config/mmix/mmix.c (mmix_target_asm_function_prologue): Likewise.
4373 (mmix_target_asm_function_epilogue): Likewise.
4374 * config/msp430/msp430.c (msp430_start_function): Likewise.
4375 * config/nds32/nds32.c (nds32_asm_function_prologue): Likewise.
4376 (nds32_asm_function_epilogue): Likewise.
4377 * config/nios2/nios2.c (nios2_asm_function_prologue): Likewise.
4378 * config/pa/pa.c (pa_output_function_prologue): Likewise.
4379 (pa_output_function_epilogue): Likewise.
4380 * config/powerpcspe/powerpcspe.c (rs6000_output_function_prologue)
4381 (rs6000_output_function_epilogue): Likewise.
4382 * config/rl78/rl78.c (rl78_start_function): Likewise.
4383 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Likewise.
4384 (rs6000_output_function_epilogue): Likewise.
4385 * config/rx/rx.c (rx_output_function_prologue): Likewise.
4386 * config/sh/sh.c (sh_output_function_epilogue): Likewise.
4387 * config/sparc/sparc.c (sparc_asm_function_prologue): Likewise.
4388 (sparc_asm_function_epilogue): Likewise.
4390 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4392 * tree.h (type_has_mode_precision_p): New function.
4393 * convert.c (convert_to_integer_1): Use it.
4394 * expr.c (expand_expr_real_2): Likewise.
4395 (expand_expr_real_1): Likewise.
4396 * fold-const.c (fold_single_bit_test_into_sign_test): Likewise.
4397 * match.pd: Likewise.
4398 * tree-ssa-forwprop.c (simplify_rotate): Likewise.
4399 * tree-ssa-math-opts.c (convert_mult_to_fma): Likewise.
4400 * tree-tailcall.c (process_assignment): Likewise.
4401 * tree-vect-loop.c (vectorizable_reduction): Likewise.
4402 * tree-vect-patterns.c (vect_recog_vector_vector_shift_pattern)
4403 (vect_recog_mult_pattern, vect_recog_divmod_pattern): Likewise.
4404 * tree-vect-stmts.c (vectorizable_conversion): Likewise.
4405 (vectorizable_assignment): Likewise.
4406 (vectorizable_shift): Likewise.
4407 (vectorizable_operation): Likewise.
4408 * tree-vrp.c (register_edge_assert_for_2): Likewise.
4410 2017-08-21 Wilco Dijkstra <wdijkstr@arm.com>
4412 * match.pd: Add pow (C, x) simplification.
4414 2017-08-21 Richard Biener <rguenther@suse.de>
4416 PR tree-optimization/81900
4417 * tree-ssa-pre.c (compute_antic_aux): Properly compute changed
4418 for blocks with abnormal predecessors.
4419 (compute_antic): Do not set visited flag prematurely.
4421 2017-08-21 Georg-Johann Lay <avr@gjlay.de>
4424 * config/avr/avr.c (avr_set_current_function): Typo in diagnostic.
4426 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4428 * stor-layout.h (vector_type_mode): Move to...
4429 * tree.h (vector_type_mode): ...here.
4430 * stor-layout.c (vector_type_mode): Move to...
4431 * tree.c (vector_type_mode): ...here. Include rtl.h and regs.h.
4433 2017-08-21 Richard Biener <rguenther@suse.de>
4435 * debug.h (struct gcc_debug_hooks): Add die_ref_for_decl and
4436 register_external_die hooks.
4437 (debug_false_tree_charstarstar_uhwistar): Declare.
4438 (debug_nothing_tree_charstar_uhwi): Likewise.
4439 * debug.c (do_nothing_debug_hooks): Adjust.
4440 (debug_false_tree_charstarstar_uhwistar): New do nothing.
4441 (debug_nothing_tree_charstar_uhwi): Likewise.
4442 * dbxout.c (dbx_debug_hooks): Adjust.
4443 (xcoff_debug_hooks): Likewise.
4444 * sdbout.c (sdb_debug_hooks): Likewise.
4445 * vmsdbgout.c (vmsdbg_debug_hooks): Likewise.
4446 * dwarf2out.c (macinfo_label_base): New global.
4447 (dwarf2out_register_external_die): New function for the
4448 register_external_die hook.
4449 (dwarf2out_die_ref_for_decl): Likewise for die_ref_for_decl.
4450 (dwarf2_debug_hooks): Use them.
4451 (dwarf2_lineno_debug_hooks): Adjust.
4452 (struct die_struct): Add with_offset flag.
4453 (DEBUG_LTO_DWO_INFO_SECTION, DEBUG_LTO_INFO_SECTION,
4454 DEBUG_LTO_DWO_ABBREV_SECTION, DEBUG_LTO_ABBREV_SECTION,
4455 DEBUG_LTO_DWO_MACINFO_SECTION, DEBUG_LTO_MACINFO_SECTION,
4456 DEBUG_LTO_DWO_MACRO_SECTION, DEBUG_LTO_MACRO_SECTION,
4457 DEBUG_LTO_LINE_SECTION, DEBUG_LTO_DWO_STR_OFFSETS_SECTION,
4458 DEBUG_LTO_STR_DWO_SECTION, DEBUG_STR_LTO_SECTION): New macros
4459 defining section names for the early LTO debug variants.
4460 (reset_indirect_string): New helper.
4461 (add_AT_external_die_ref): Helper for dwarf2out_register_external_die.
4462 (print_dw_val): Add support for offsetted symbol references.
4463 (get_ultimate_context): Split out from is_cxx.
4464 (is_cxx): Use get_ultimate_context.
4465 (is_fortran): Add decl overload.
4466 (compute_comp_unit_symbol): Split out worker from
4467 compute_section_prefix.
4468 (compute_section_prefix): Call compute_comp_unit_symbol and
4469 set comdat_type_p here.
4470 (output_die): Skip DIE symbol output for the LTO added one.
4471 Handle DIE symbol references with offset.
4472 (output_comp_unit): Guard section name mangling properly.
4473 For LTO debug sections emit a symbol at the section beginning
4474 which we use to refer to its DIEs.
4475 (add_abstract_origin_attribute): For DIEs registered via
4476 dwarf2out_register_external_die directly refer to the early
4477 DIE rather than indirectly through the shadow one we created.
4478 Remove obsolete call to dwarf2out_abstract_function for
4479 non-function/block origins.
4480 (gen_array_type_die): When generating early LTO debug do
4481 not emit DW_AT_string_length.
4482 (gen_formal_parameter_die): Do not re-create DIEs for PARM_DECLs
4483 late when in LTO. As suggested place a gcc_unreachable for
4484 the DECL_ABSTRACT_P case.
4485 (gen_subprogram_die): Avoid another specification DIE
4486 for early built declarations/definitions for the late LTO case.
4487 (gen_variable_die): Add type references for late duplicated VLA dies
4489 (gen_inlined_subroutine_die): Do not call dwarf2out_abstract_function,
4490 we have the abstract instance already.
4491 (process_scope_var): Adjust decl DIE contexts in LTO which
4492 first puts them in limbo.
4493 (gen_decl_die): Do not generate type DIEs late apart from
4494 types for VLAs or for decls we do not yet have a DIE. Do not
4495 call dwarf2out_abstract_function late.
4496 (dwarf2out_early_global_decl): Make sure to create DIEs
4497 for abstract instances of a decl first.
4498 (dwarf2out_late_global_decl): Adjust comment.
4499 (output_macinfo_op): With multiple macro sections use
4500 macinfo_label_base to distinguish labels.
4501 (output_macinfo): Likewise. Update macinfo_label_base.
4502 Pass in the line info label.
4503 (note_variable_value_in_expr): When generating LTO resolve
4504 all variable values here by generating DIEs as needed.
4505 (init_sections_and_labels): Add early LTO debug flag parameter
4506 and generate different sections and names if set. Add generation
4507 counter for the labels so we can have multiple of them.
4508 (reset_dies): Helper to allow DIEs to be output multiple times.
4509 (dwarf2out_finish): When outputting DIEs to the fat part of an
4510 LTO object first reset DIEs.
4511 (dwarf2out_early_finish): Output early DIEs when generating LTO.
4512 (modified_type_die): Check for decl_ultimate_origin being self
4514 (gen_type_die_with_usage): Likewise.
4515 (gen_typedef_die): Allow decl_ultimate_origin being self.
4516 (set_decl_abstract_flags): Remove.
4517 (set_block_abstract_flags): Likewise.
4518 (dwarf2out_abstract_function): Treat the early generated DIEs
4519 as the abstract copy and only add DW_AT_inline and
4520 DW_AT_artificial here and call set_decl_origin_self.
4521 If the DIE has an abstract origin don't do anything.
4522 * tree.c (free_lang_data): Build a dummy TRANSLATION_UNIT_DECL
4523 if we have none yet (Go fails to build one, PR78628).
4524 (variably_modified_type_p): Prevent endless recursion for Ada
4525 cyclic pointer types.
4526 * lto-streamer-in.c: Include debug.h.
4527 (dref_queue): New global.
4528 (lto_read_tree_1): Stream in DIE references.
4529 (lto_input_tree): Register DIE references.
4530 (input_function): Stream DECL_DEBUG_ARGS.
4531 * lto-streamer-out.c: Include debug.h.
4532 (lto_write_tree_1): Output DIE references.
4533 (DFS::DFS_write_tree_body): Follow DECL_ABSTRACT_ORIGIN.
4534 Force a TRANSLATION_UNIT_DECL DECL_CONTEXT for file-scope decls.
4535 (output_function): Stream DECL_DEBUG_ARGS.
4536 * tree-streamer-in.c (lto_input_ts_decl_common_tree_pointers):
4537 Stream DECL_ABSTRACT_ORIGIN.
4538 * tree-streamer-out.c (write_ts_decl_common_tree_pointers): Likewise.
4539 (write_ts_decl_minimal_tree_pointers): Force a TRANSLATION_UNIT_DECL
4540 DECL_CONTEXT for file-scope decls.
4541 * lto-streamer.h (struct dref_entry): Declare.
4542 (dref_queue): Likewise.
4543 * cfgexpand.c (pass_expand::execute): Do not call the
4544 outlining_inline_function hook here.
4545 * lto-wrapper.c (debug_obj): New global.
4546 (tool_cleanup): Unlink it if required.
4547 (debug_objcopy): New function.
4548 (run_gcc): Handle early debug sections in the IL files by
4549 extracting them to separate files, partially linkin them and
4550 feeding the result back as result to the linker.
4551 * config/darwin.h (DEBUG_LTO_INFO_SECTION, DEBUG_LTO_ABBREV_SECTION,
4552 DEBUG_LTO_MACINFO_SECTION, DEBUG_LTO_LINE_SECTION,
4553 DEBUG_STR_LTO_SECTION, DEBUG_LTO_MACRO_SECTION): Put early debug
4554 sections into a separate segment.
4555 * config/darwin.c (darwin_asm_named_section): Handle __GNU_DWARF_LTO
4557 (darwin_asm_dwarf_section): Likewise.
4558 (darwin_asm_output_dwarf_offset): Likewise.
4559 * config/i386/i386.c (make_resolver_func): Set DECL_IGNORED_P.
4561 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4562 Alan Hayward <alan.hayward@arm.com>
4563 David Sherwood <david.sherwood@arm.com>
4565 * read-md.h (md_reader::record_potential_iterator_use): Replace
4566 pointer argument with an rtx and an index.
4567 * read-rtl.c (iterator_group::apply_iterator): Likewise.
4568 (apply_mode_iterator): Likewise.
4569 (apply_code_iterator): Likewise.
4570 (apply_int_iterator): Likewise.
4571 (apply_subst_iterator): Likewise.
4572 (record_iterator_use): Likewise.
4573 (record_attribute_use): Likewise.
4574 (md_reader::record_potential_iterator_use): Likewise. Update calls
4575 to record_iterator_use and apply_iterator.
4576 (iterator_use): Replace ptr with x and index.
4577 (attribute_use): Likewise.
4578 (apply_attribute_uses): Update calls to apply_iterator.
4579 (apply_iterators): Likewise. Update initialization of iterator_use.
4580 (rtx_reader::read_rtx_code): Update calls to record_iterator_use
4581 and record_potential_iterator_use.
4582 (rtx_reader::read_rtx_operand): Likewise.
4584 2017-08-21 Richard Sandiford <richard.sandiford@linaro.org>
4585 Alan Hayward <alan.hayward@arm.com>
4586 David Sherwood <david.sherwood@arm.com>
4588 * varasm.c (const_rtx_hash_1): Don't hash in the mode of a
4591 2017-08-21 Richard Biener <rguenther@suse.de>
4594 * tree-ssa-alias.c (stmt_kills_ref_p): Handle array accesses
4595 at struct end conservatively when comparing common bases.
4597 2017-08-21 Richard Biener <rguenther@suse.de>
4599 * tree-ssa-loop-im.c (struct lim_aux_data): Add ref index member.
4600 (mem_ref_in_stmt): Remove.
4601 (determine_max_movement): Use ref index to get at the reference.
4602 (invariantness_dom_walker::before_dom_children): Deal with
4603 lim data already initialized.
4604 (gather_mem_refs_stmt): Initialize lim data and record ref index.
4606 2017-08-19 Uros Bizjak <ubizjak@gmail.com>
4608 * config/i386/i386.h (OPTION_MASK_ISA_ROUND): Remove.
4609 (TARGET_ISA_ROUND): Ditto.
4610 (TARGET_ROUND): Ditto.
4611 * config/i386/i386.c: Substitute TARGET_ROUND with TARGET_SSE4_1.
4612 * config/i386/i386.md: Ditto.
4613 * config/i386/sse.md: Ditto.
4614 * config/i386/i386-builtin.def: Substitute OPTION_MASK_ISA_ROUND
4615 with OPTION_MASK_ISA_SSE4_1.
4617 2017-08-19 Uros Bizjak <ubizjak@gmail.com>
4620 * doc/extend.texi (x86 Built-in Functions): Correct the name of
4621 __builtin_ia32_lzcnt_u16.
4623 2017-08-18 Peter Bergner <bergner@vnet.ibm.com>
4626 * config/rs6000/rs6000.c (rs6000_activate_target_options): New function.
4627 (rs6000_set_current_function): Rewrite function to use it.
4629 2017-08-18 H.J. Lu <hongjiu.lu@intel.com>
4632 * print-tree.c (print_node): Support DECL_WARN_IF_NOT_ALIGN
4633 and TYPE_WARN_IF_NOT_ALIGN.
4634 * stor-layout.c (do_type_align): Merge DECL_WARN_IF_NOT_ALIGN.
4635 (handle_warn_if_not_align): New.
4636 (place_union_field): Call handle_warn_if_not_align.
4637 (place_field): Call handle_warn_if_not_align.
4638 Copy TYPE_WARN_IF_NOT_ALIGN.
4639 (finish_builtin_struct): Copy TYPE_WARN_IF_NOT_ALIGN.
4640 (layout_type): Likewise.
4641 * tree-core.h (tree_type_common): Add warn_if_not_align. Set
4643 (tree_decl_common): Add warn_if_not_align.
4644 * tree.c (build_range_type_1): Copy TYPE_WARN_IF_NOT_ALIGN.
4645 * tree.h (TYPE_WARN_IF_NOT_ALIGN): New.
4646 (SET_TYPE_WARN_IF_NOT_ALIGN): Likewise.
4647 (DECL_WARN_IF_NOT_ALIGN): Likewise.
4648 (SET_DECL_WARN_IF_NOT_ALIGN): Likewise.
4649 * doc/extend.texi: Document warn_if_not_aligned attribute.
4650 * doc/invoke.texi: Document -Wif-not-aligned and -Wpacked-not-aligned.
4652 2017-08-17 Martin Liska <mliska@suse.cz>
4655 * tree-loop-distribution.c (ddrs_table): Change type to pointer type.
4656 (get_data_dependence): Use it as pointer type.
4657 (distribute_loop): Likewise.
4659 2017-08-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4661 * config/rs6000/altivec.md (UNSPEC_VMRGOW_DIRECT): New constant.
4662 (p8_vmrgew_v4sf_direct): Generalize to p8_vmrgew_<mode>_direct.
4663 (p8_vmrgow_<mode>_direct): New define_insn.
4664 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Properly
4665 handle endianness for vmrgew and vmrgow permute patterns.
4667 2017-08-17 Peter Bergner <bergner@vnet.ibm.com>
4669 * config/rs6000/altivec.md (VParity): Remove TARGET_VSX_TIMODE.
4670 * config/rs6000/rs6000-cpus.def: Remove comment.
4671 (ISA_2_7_MASKS_SERVER): Delete OPTION_MASK_VSX_TIMODE;
4672 (POWERPC_MASKS): Likewise.
4673 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Remove unneeded
4674 use of TARGET_VSX_TIMODE.
4675 (rs6000_setup_reg_addr_masks): Change TARGET_VSX_TIMODE to TARGET_VSX.
4676 (rs6000_init_hard_regno_mode_ok): Remove unneeded uses of
4677 TARGET_VSX_TIMODE. Change use of TARGET_VSX_TIMODE to TARGET_VSX.
4678 (rs6000_option_override_internal): Remove dead code.
4679 (rs6000_legitimize_address): Change TARGET_VSX_TIMODE to TARGET_VSX.
4680 (rs6000_legitimize_reload_address): Likewise.
4681 (rs6000_legitimate_address_p): Likewise.
4682 (rs6000_opt_masks): Delete "vsx-timode".
4683 (rs6000_disable_incompatible_switches): Remove mention of -mvsx-timode
4684 from function comment.
4685 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Delete.
4686 * config/rs6000/rs6000.md (FMOVE128_GPR): Remove TARGET_VSX_TIMODE.
4687 (V16QI, V8HI, V4SI, V4SF, V2DI, V2DF, V1TI): Remove useless empty
4689 * config/rs6000/rs6000.opt (mvsx-timode): Replace with stub.
4690 * config/rs6000/vector.md (VEC_IP): Remove TARGET_VSX_TIMODE.
4691 * config/rs6000/vsx.md (VSX_LE_128): Likewise.
4694 (define_peephole2): Likewise.
4696 2017-08-17 Martin Sebor <msebor@redhat.com>
4699 * pretty-print.c (pp_format): Use strnlen in %.*s to avoid reading
4700 past the end of an array.
4701 (test_pp_format): Add test cases.
4703 2017-08-17 Richard Sandiford <richard.sandiford@linaro.org>
4705 * internal-fn.def (CLRSB, CLZ, CTZ, FFS, PARITY, POPCOUNT): Add
4706 missing ECF_NOTHROW flags.
4708 2017-08-17 Peter Bergner <bergner@vnet.ibm.com>
4711 * config/rs6000/vsx.md (*vsx_le_permute_<mode>): Add support for
4712 operands residing in integer registers.
4713 (*vsx_le_perm_load_<mode>): Likewise.
4714 (*vsx_le_perm_store_<mode>): Likewise.
4715 (define_peephole2): Add peepholes to optimize the above.
4717 2017-08-17 Marek Polacek <polacek@redhat.com>
4720 * fold-const.c (operand_equal_for_comparison_p): Remove code that used
4721 to mimic what shorten_compare did. Change the return type to bool.
4722 (fold_cond_expr_with_comparison): Update call to
4723 operand_equal_for_comparison_p.
4724 (fold_ternary_loc): Likewise.
4726 2017-08-17 Jackson Woodruff <jackson.woodruff@arm.com>
4728 * aarch64-simd.md (mov<mode>): No longer force zero immediate into
4730 (*aarch64_simd_mov<mode>): Add new case for stp using zero immediate.
4732 2017-08-17 Richard Biener <rguenther@suse.de>
4734 * tree-ssa-structalias.c (solve_graph): When propagating
4735 to successors update the graphs succ edges and avoid duplicate work.
4737 2017-08-17 Maxim Ostapenko <m.ostapenko@samsung.com>
4740 * config/i386/i386.c (ix86_option_override_internal): Save target
4741 specific options after ix86_stack_protector_guard_reg was changed.
4743 2017-08-17 Richard Biener <rguenther@suse.de>
4745 PR tree-optimization/81827
4746 * tree-ssa-structalias.c (struct variable_info): Add is_reg_var flag.
4747 (new_var_info): Initialize it conservatively.
4748 (get_call_vi): Mark register vars.
4749 (new_scalar_tmp_constraint_exp): Likewise.
4750 (handle_rhs_call): Likewise.
4751 (handle_const_call): Likewise.
4752 (create_function_info_for): Likewise.
4753 (solve_constraints): Sort varinfos to separate register from
4754 non-register vars to pack points-to solution bitmaps during
4757 2017-08-17 Marek Polacek <polacek@redhat.com>
4759 * gimplify.c (gimplify_adjust_omp_clauses): Compare with 0 instead of 1.
4761 2017-08-17 Richard Biener <rguenther@suse.de>
4763 * tree-vrp.c (vrp_int_const_binop): Do not set *overflow_p
4764 to true when overflow is undefined and we saturated the result.
4766 2017-08-17 Alan Modra <amodra@gmail.com>
4769 * config/rs6000/rs6000.c (rs6000_savres_strategy): Revert 2017-08-09.
4770 Don't use store multiple if only one reg needs saving.
4771 (interesting_frame_related_regno): New function.
4772 (rs6000_frame_related): Don't emit frame info for regs that
4774 (rs6000_emit_epilogue): Likewise.
4776 2017-08-16 Nathan Sidwell <nathan@acm.org>
4778 * tree-core.h (tree_type_non_common): Rename binfo to lang_1.
4779 * tree.h (TYPE_BINFO): Use type_non_common.maxval.
4780 (TYPE_LANG_SLOT_1): Use type_non_common.lang_1, for any type.
4781 * tree.c (free_lang_data_in_type): Use else-if chain. Always
4782 clear TYPE_LANG_1. Remove obsolete member-function stripping.
4783 (find_decls_types_r): Comment about TYPE_MAX_VALUES_RAW.
4784 (verify_type): Adjust for TYPE_BINFO move.
4785 * lto-streamer-out.c (DFS::DFS_write_tree_body): No need to
4786 process TYPE_BINFO directly.
4787 (hash_tree): Likewise.
4788 * tree-streamer-in.c (lto_input_ts_type_non_common_tree_pointers):
4790 * tree-streamer-out.c (write_ts_type_non_common_tree_pointers):
4793 2017-08-16 David Malcolm <dmalcolm@redhat.com>
4795 * diagnostic-show-locus.c (colorizer::m_caret): Remove unused field.
4797 2017-08-16 Uros Bizjak <ubizjak@gmail.com>
4800 * config/i386/i386.md (*anddi_1_btr): Change predicates of
4801 operand 0 and operand 1 to nomimmediate_operand. Add "m" constraint.
4802 Add ix86_binary_operator_ok to insn constraint.
4803 (*iordi_1_bts): Ditto.
4804 (*xordi_1_btc): Ditto.
4805 (*btsq): Change predicate of operand 0 to nonimmediate_operand.
4806 Update corresponding peephole2 pattern.
4810 2017-08-16 Bin Cheng <bin.cheng@arm.com>
4812 PR tree-optimization/81832
4813 * tree-ssa-loop-ch.c (should_duplicate_loop_header_p): Don't
4814 copy loop header which has IFN_LOOP_DIST_ALIAS call.
4816 2017-08-16 Marek Polacek <polacek@redhat.com>
4819 * fold-const.c (fold_indirect_ref_1): Restore original behavior
4820 regarding size_zero_node.
4822 2017-08-16 Martin Liska <mliska@suse.cz>
4825 * config.gcc: Respect previously set extra_objs in case
4828 2017-08-16 Richard Sandiford <richard.sandiford@linaro.org>
4830 PR tree-optimization/81835
4831 * tree-vect-loop.c (vect_is_simple_reduction): Simply checks for
4832 the phi SSA_NAME. Check that the condition in a COND_EXPR does
4833 not depend on the phi.
4835 2017-08-16 Alan Modra <amodra@gmail.com>
4837 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Delete
4840 2017-08-16 Alan Modra <amodra@gmail.com>
4842 * config/rs6000/rs6000.c (rs6000_reg_live_or_pic_offset_p): Merge..
4843 (save_reg_p): ..into this. Update all callers.
4844 (first_reg_to_save): Simplify.
4846 2017-08-16 Alan Modra <amodra@gmail.com>
4848 * config/rs6000/rs6000.c (rs6000_savres_strategy): Don't restore
4851 2017-08-15 Joseph Myers <joseph@codesourcery.com>
4855 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Only unroll for
4856 constant count if that count is less than 32.
4858 2017-08-15 Nathan Sidwell <nathan@acm.org>
4860 * gcc.c (execute): Emit friendlier message if inferior is killed
4861 by an external cause.
4863 2017-08-15 Richard Biener <rguenther@suse.de>
4865 PR tree-optimization/81790
4866 * tree-ssa-sccvn.c (vn_lookup_simplify_result): Handle both
4867 CONSTRUCTORs from simplifying and VN.
4869 2017-08-14 Martin Sebor <msebor@redhat.com>
4871 * builtin-attrs.def: Add comments.
4873 2017-08-14 Martin Sebor <msebor@redhat.com>
4876 * doc/extend.texi (attribute nonstring): Document new attribute.
4878 2017-08-14 Martin Sebor <msebor@redhat.com>
4881 * tree-diagnostic.c (default_tree_printer): Handle %G.
4882 * gimple-pretty-print.h (percent_G_format): Declare new function.
4883 * gimple-pretty-print.c (percent_G_format): Define.
4884 * tree-pretty-print.c (percent_K_format): Add argument.
4886 2017-08-14 Martin Sebor <msebor@redhat.com>
4888 PR translation/79998
4889 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call):
4890 Remove a stray space.
4892 2017-08-14 Uros Bizjak <ubizjak@gmail.com>
4895 * config/i386/i386.md (*anddi_1_btr): New insn_and_split pattern.
4896 (*iordi_1_bts): Ditto.
4897 (*xordi_1_btc): Ditto.
4899 2017-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4902 * config/rs6000/linux64.h (INVALID_64BIT): Use quoted strings.
4903 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
4905 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Use
4906 quoted strings, and make more translator-friendly.
4907 (darwin_rs6000_override_options): Likewise.
4908 (rs6000_option_override_internal): Likewise.
4909 (rs6000_return_in_memory): Fix overlong line.
4910 (init_cmulative_args): Use quoted strings, and make more
4911 translator-friendly.
4912 (rs6000_pass_by_reference): Fix overlong line.
4913 (def_builtin): Use quoted strings.
4914 (altivec_expand_predicate_builtin): Use quoted strings, and make
4915 more translator-friendly.
4916 (htm_expand_builtin): Use quoted strings.
4917 (cpu_expand_builtin): Use quoted strings, and make more
4918 translator-friendly.
4919 (altivec_expand_builtin): Likewise.
4920 (paired_expand_predicate_builtin): Likewise.
4921 (rs6000_invalid_builtin): Likewise.
4922 (builtin_function_type): Use quoted strings.
4923 (rs6000_expand_split_stack_prologue): Use quoted strings, and make
4924 more translator-friendly.
4925 (rs6000_trampoline_init): Likewise.
4926 (rs6000_handle_altivec_attribute): Likewise.
4927 (rs6000_inner_target_options): Use quoted strings.
4928 (rs6000_disable_incompatible_switches): Likewise.
4929 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Use quoted
4930 strings, and make more translator-friendly.
4931 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use quoted strings.
4933 2017-08-14 Bin Cheng <bin.cheng@arm.com>
4935 PR tree-optimization/81799
4936 * tree-loop-distribution.c (version_loop_by_alias_check): Force
4937 cond_expr to simple gimple operand.
4939 2017-08-14 Wilco Dijkstra <wdijkstr@arm.com>
4942 * doc/sourcebuild.texi (autoincdec): Add autoincdec description.
4944 2017-08-14 Georg-Johann Lay <avr@gjlay.de>
4948 * config/avr/avr.opt (mgas-isr-prologues): New Var avr_gasisr_prologues.
4949 * config/avr/avr.md (gasisr, *gasisr): Use it instead of
4950 TARGET_GASISR_PROLOGUES.
4951 * config/avr/avr.c (avr_option_override): Same.
4952 (avr_pass_pre_proep::execute): Same.
4954 2017-08-13 H.J. Lu <hongjiu.lu@intel.com>
4957 * config/i386/i386.c (ix86_finalize_stack_frame_flags): Replace
4958 frame pointer with stack pointer - UNITS_PER_WORD in debug insns.
4960 2017-08-13 Uros Bizjak <ubizjak@gmail.com>
4962 * config/i386/i386.md (*load_tp_<mode>): Redefine as
4963 define_insn_and_split. Split to a memory load from 0 in
4964 DEFAULT_TLS_SEG_REG address space. Merge with *load_tp_x32
4965 using PTR mode iterator.
4966 (*load_tp_x32_zext"): Redefine as define_insn_and_split.
4967 Split to a memory load from 0 in DEFAULT_TLS_SEG_REG address space.
4968 (*add_tp_<mode>): Redefine as define_insn_and_split.
4969 Split to an add with a memory load from 0 in DEFAULT_TLS_SEG_REG
4970 address space. Merge with *add_tp_x32 using PTR mode iterator.
4971 (*add_tp_x32_zext"): Redefine as define_insn_and_split.
4972 Split to an add with a memory load from 0 in
4973 DEFAULT_TLS_SEG_REG address space.
4975 2017-08-12 Andrew Pinski <apinski@cavium.com>
4977 * config/aarch64/aarch64-option-extensions.def (rdma):
4978 Fix feature string to what Linux prints out in /proc/cpuinfo.
4980 2017-08-12 Pierre-Marie de Rodat <derodat@adacore.com>
4983 * dwarf2out.c (modified_type_die): For C typedef types that have
4984 an ultimate origin, process the ultimate origin instead of the
4986 (gen_typedef_die): Assert that input DECLs have no ultimate
4988 (gen_type_die_with_usage): For typedef variants that have an
4989 ultimate origin, just call gen_decl_die on the original DECL.
4990 (process_scope_var): Avoid creating DIEs for local typedefs and
4991 concrete static variables.
4993 2017-08-12 Alan Modra <amodra@gmail.com>
4997 * config/rs6000/sysv4.h (STARTFILE_LINUX_SPEC): Upgrade to
4998 match gnu-user.h startfile.
4999 (ENDFILE_LINUX_SPEC): Similarly.
5001 2017-08-11 Thomas Schwinge <thomas@codesourcery.com>
5004 * config/nvptx/nvptx.c (nvptx_override_options_after_change):
5006 (TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE): Remove definition.
5008 2017-08-11 Tamar Christina <tamar.christina@arm.com>
5009 * config/aarch64/aarch64.md (mov<mode>): Change.
5010 (*movhf_aarch64, *movsf_aarch64, *movdf_aarch64):
5011 aarch64_reg_or_fp_float into aarch64_reg_or_fp_zero.
5012 * config/aarch64/predicates.md (aarch64_reg_or_fp_float): Removed.
5014 2017-08-11 Eric Botcazou <ebotcazou@adacore.com>
5016 * tree-sra.c (build_access_from_expr_1): Use more precise diagnostics
5017 for storage order barriers.
5019 2017-08-11 Martin Liska <mliska@suse.cz>
5022 * tree-chkp.c (chkp_get_bounds_for_decl_addr): Do not instrument
5023 variables of void type.
5025 2017-08-11 Martin Liska <mliska@suse.cz>
5027 * asan.c (asan_protect_global): Replace ASM_OUTPUT_DEF with
5028 TARGET_SUPPORTS_ALIASES.
5029 * cgraph.c (cgraph_node::create_same_body_alias): Likewise.
5030 * ipa-visibility.c (can_replace_by_local_alias): Likewise.
5031 (optimize_weakref): Likewise.
5032 * symtab.c (symtab_node::noninterposable_alias): Likewise.
5033 * varpool.c (varpool_node::create_extra_name_alias): Likewise.
5034 * defaults.h: Introduce TARGET_SUPPORTS_ALIASES.
5036 2017-08-11 Martin Liska <mliska@suse.cz>
5039 * config/i386/i386.c (make_resolver_func): Do complete
5040 refactoring of the function.
5042 2017-08-10 Uros Bizjak <ubizjak@gmail.com>
5045 * config/i386/i386.opt (mstack-protector-guard-symbol=): New option
5046 * config/i386/i386.c (ix86_stack_protect_guard): Use
5047 ix86_stack_protect_guard_symbol_str to generate varible declaration.
5048 * doc/invoke.texi (x86 Options): Document
5049 -mstack-protector-guard-symbol= option.
5051 2017-08-10 Uros Bizjak <ubizjak@gmail.com>
5053 * config/i386/i386-protos.h (ix86_split_stack_guard): New prototype.
5054 * config/i386/i386.c (ix86_split_stack_guard): New function.
5055 (ix86_xpand_split_stack_prologue): Call ix86_split_stack_guard.
5056 (ix86_legitimate_address_p) <case UNSPEC_STACK_CHECK>: Remove.
5057 (i386_asm_output_addr_const_extra) <case UNSPEC_STACK_CHECK>: Ditto.
5058 (optput_pic_addr_const): Remove UNSPEC_STACK_CHECK handling.
5059 * config/i386/i386.md (unspec): Remove UNSPEC_STACK_CHECK.
5060 (split_stack_space_check): Call ix86_split_stack_guard.
5062 2017-08-10 Martin Sebor <msebor@redhat.com>
5064 * print-tree.c (print_node): Print location using the established
5066 Replace spaces with colons.
5067 (debug_raw, debug): Ditto.
5069 2017-08-10 Martin Sebor <msebor@redhat.com>
5072 * pretty-print.c (pp_format): Correct the handling of %s precision.
5074 2017-08-10 H.J. Lu <hongjiu.lu@intel.com>
5077 * config/i386/i386.c (ix86_finalize_stack_realign_flags): Renamed
5079 (ix86_finalize_stack_frame_flags): This. Also clear
5080 frame_pointer_needed if -fno-omit-frame-pointer is used without
5082 (ix86_expand_prologue): Replace ix86_finalize_stack_realign_flags
5083 with ix86_finalize_stack_frame_flags.
5084 (ix86_expand_epilogue): Likewise.
5085 (ix86_expand_split_stack_prologue): Likewise.
5086 * doc/invoke.texi: Add a note for -fno-omit-frame-pointer.
5088 2017-08-10 Martin Liska <mliska@suse.cz>
5091 * c-attribs.c (handle_target_attribute):
5092 Report warning for an empty string argument of target attribute.
5094 2017-08-09 Jakub Jelinek <jakub@redhat.com>
5097 * omp-low.c (omp_copy_decl): Don't remap FORCED_LABEL or DECL_NONLOCAL
5099 * tree-cfg.c (move_stmt_op): Don't adjust DECL_CONTEXT of FORCED_LABEL
5100 or DECL_NONLOCAL labels.
5101 (move_stmt_r) <case GIMPLE_LABEL>: Adjust DECL_CONTEXT of FORCED_LABEL
5102 or DECL_NONLOCAL labels here.
5104 2017-08-09 Will Schmidt <will_schmidt@vnet.ibm.com>
5106 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add blurb
5107 to indicate when early gimple folding has been disabled.
5108 (rs6000_gimple_fold_builtin): Add debug content.
5109 (rs6000_invalid_builtin): Fix whitespace.
5110 (rs6000_expand_builtin): Fix whitespace.
5111 * config/rs6000/rs6000.opt: Add option for -mfold-gimple.
5113 2017-08-09 Segher Boessenkool <segher@kernel.crashing.org>
5116 * config/rs6000/rs6000.c (rs6000_savres_strategy): Don't use
5117 SAVE_MULTIPLE if not all the registers that saves, should be saved.
5119 2017-08-09 Jim Wilson <jim.wilson@linaro.org>
5121 * config/aarch64/aarch64-cores.def (falkor): Use falkor pipeline.
5122 (qdf24xx): Likewise.
5123 * config/aarch64/aarch64.md: Include falkor.md.
5124 * config/aarch64/falkor.md: New.
5126 2017-08-09 Marek Polacek <polacek@redhat.com>
5129 * diagnostic-core.h (emit_diagnostic_valist): Add declaration.
5130 * diagnostic.c (emit_diagnostic): Add a comment.
5131 (emit_diagnostic_valist): New function.
5133 2017-08-09 Marek Polacek <polacek@redhat.com>
5136 * input.c (make_location): New overload.
5137 * input.h (make_location): Declare.
5139 2017-08-08 Alan Modra <amodra@gmail.com>
5140 H.J. Lu <hongjiu.lu@intel.com>
5143 * gcc.c (NO_PIE_SPEC): Delete.
5144 (PIE_SPEC): Define as !no-pie/pie. Move static|shared|r
5146 (LINK_PIE_SPEC): ..to here.
5147 (LINK_COMMAND_SPEC): Support -no-pie.
5148 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Correct
5149 chain of crtbegin*.o selection, update for PIE_SPEC changes and
5151 (GNU_USER_TARGET_ENDFILE_SPEC): Similarly.
5152 * config/sol2.h (STARTFILE_CRTBEGIN_SPEC): Similarly.
5153 (ENDFILE_CRTEND_SPEC): Similarly.
5155 2017-08-08 Uros Bizjak <ubizjak@gmail.com>
5158 * config/i386/i386.opt (mstack-protector-guard-reg=): New option
5159 (mstack-protector-guard-offset=): Ditto.
5160 * config/i386/i386.c (ix86_option_override): Handle
5161 -mstack-protector-guard-reg= and -mstack-protector-guard-offset=
5163 (ix86_stack_protect_guard): Use ix86_stack_protect_guard_reg and
5164 ix86_stack_protect_guard_offset variables.
5165 (TARGET_STACK_PROTECT_GUARD): Always define.
5166 * doc/invoke.texi (x86 Options): Document -mstack-protector-guard-reg=
5167 and -mstack-protector-guard-offset= options.
5169 2017-08-08 Bin Cheng <bin.cheng@arm.com>
5171 * tree-ssa-loop-ivopts.c (relate_compare_use_with_all_cands): Handle
5172 boundary case for the last candidate.
5174 2017-08-08 Bin Cheng <bin.cheng@arm.com>
5176 * doc/invoke.texi: Document -ftree-loop-distribution for O3.
5177 * opts.c (default_options_table): Add OPT_ftree_loop_distribution.
5179 2017-08-08 Tamar Christina <tamar.christina@arm.com>
5182 * config/aarch64/aarch64.md (xorsign<mode>3): New optabs.
5183 * config/aarch64/aarch64-builtins.c
5184 (aarch64_builtin_vectorized_function): Added CASE_CFN_XORSIGN.
5185 * config/aarch64/aarch64-simd-builtins.def: Added xorsign BINOP.
5186 * config/aarch64/aarch64-simd.md: Added xorsign<mode>3
5188 2017-08-08 Tamar Christina <tamar.christina@arm.com>
5189 Andrew Pinski <pinskia@gmail.com>
5192 * internal-fn.def (XORSIGN): New.
5193 * optabs.def (xorsign_optab): New.
5194 * tree-ssa-math-opts.c (is_copysign_call_with_1): New.
5195 (convert_expand_mult_copysign): New.
5196 (pass_optimize_widening_mul::execute): Call
5197 convert_expand_mult_copysign.
5199 2017-08-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5201 PR tree-optimization/81354
5202 * gimple-ssa-strength-reduction.c (create_add_on_incoming_edge):
5203 Insert on edges rather than explicitly creating landing pads.
5204 (analyze_candidates_and_replace): Commit edge inserts.
5206 2017-08-08 Richard Biener <rguenther@suse.de>
5209 * tree-ssa-loop-niter.c: Include tree-dfa.h.
5210 (expand_simple_operations): Also look through ADDR_EXPRs with
5211 MEM_REF bases treating them as POINTER_PLUS_EXPR.
5213 2017-08-08 Richard Biener <rguenther@suse.de>
5215 PR tree-optimization/81723
5216 * tree-vect-slp.c (struct bst_traits): New hash traits.
5217 (bst_fail): New global.
5218 (vect_build_slp_tree_2): New worker, split out from ...
5219 (vect_build_slp_tree): ... this now wrapping it with using
5220 bst_fail set to cache SLP tree build fails. Properly handle
5222 (vect_analyze_slp_instance): Allocate and free bst_fail.
5224 2017-08-08 Martin Liska <mliska@suse.cz>
5227 * ipa-icf-gimple.c (func_checker::compare_cst_or_decl): Consider
5228 LABEL_DECLs that can be from a different function.
5230 2017-08-08 Bin Cheng <bin.cheng@arm.com>
5232 PR tree-optimization/81744
5233 * tree-predcom.c (prepare_finalizers_chain): Deep copy expr of
5234 loop's number of iterations.
5236 2017-08-08 Martin Liska <mliska@suse.cz>
5238 * asan.c: Include header files.
5239 * attribs.c (build_decl_attribute_variant): New function moved
5241 (build_type_attribute_qual_variant): Likewise.
5242 (cmp_attrib_identifiers): Likewise.
5243 (simple_cst_list_equal): Likewise.
5244 (omp_declare_simd_clauses_equal): Likewise.
5245 (attribute_value_equal): Likewise.
5246 (comp_type_attributes): Likewise.
5247 (build_type_attribute_variant): Likewise.
5248 (lookup_ident_attribute): Likewise.
5249 (remove_attribute): Likewise.
5250 (merge_attributes): Likewise.
5251 (merge_type_attributes): Likewise.
5252 (merge_decl_attributes): Likewise.
5253 (merge_dllimport_decl_attributes): Likewise.
5254 (handle_dll_attribute): Likewise.
5255 (attribute_list_equal): Likewise.
5256 (attribute_list_contained): Likewise.
5257 * attribs.h (lookup_attribute): New function moved from tree.[ch].
5258 (lookup_attribute_by_prefix): Likewise.
5259 * bb-reorder.c: Include header files.
5260 * builtins.c: Likewise.
5261 * calls.c: Likewise.
5262 * cfgexpand.c: Likewise.
5263 * cgraph.c: Likewise.
5264 * cgraphunit.c: Likewise.
5265 * convert.c: Likewise.
5266 * dwarf2out.c: Likewise.
5267 * final.c: Likewise.
5268 * fold-const.c: Likewise.
5269 * function.c: Likewise.
5270 * gimple-expr.c: Likewise.
5271 * gimple-fold.c: Likewise.
5272 * gimple-pretty-print.c: Likewise.
5273 * gimple.c: Likewise.
5274 * gimplify.c: Likewise.
5275 * hsa-common.c: Likewise.
5276 * hsa-gen.c: Likewise.
5277 * internal-fn.c: Likewise.
5278 * ipa-chkp.c: Likewise.
5279 * ipa-cp.c: Likewise.
5280 * ipa-devirt.c: Likewise.
5281 * ipa-fnsummary.c: Likewise.
5282 * ipa-inline.c: Likewise.
5283 * ipa-visibility.c: Likewise.
5285 * lto-cgraph.c: Likewise.
5286 * omp-expand.c: Likewise.
5287 * omp-general.c: Likewise.
5288 * omp-low.c: Likewise.
5289 * omp-offload.c: Likewise.
5290 * omp-simd-clone.c: Likewise.
5291 * opts-global.c: Likewise.
5292 * passes.c: Likewise.
5293 * predict.c: Likewise.
5294 * sancov.c: Likewise.
5295 * sanopt.c: Likewise.
5296 * symtab.c: Likewise.
5297 * toplev.c: Likewise.
5298 * trans-mem.c: Likewise.
5299 * tree-chkp.c: Likewise.
5300 * tree-eh.c: Likewise.
5301 * tree-into-ssa.c: Likewise.
5302 * tree-object-size.c: Likewise.
5303 * tree-parloops.c: Likewise.
5304 * tree-profile.c: Likewise.
5305 * tree-ssa-ccp.c: Likewise.
5306 * tree-ssa-live.c: Likewise.
5307 * tree-ssa-loop.c: Likewise.
5308 * tree-ssa-sccvn.c: Likewise.
5309 * tree-ssa-structalias.c: Likewise.
5310 * tree-ssa.c: Likewise.
5311 * tree-streamer-in.c: Likewise.
5312 * tree-vectorizer.c: Likewise.
5313 * tree-vrp.c: Likewise.
5315 * ubsan.c: Likewise.
5316 * varasm.c: Likewise.
5317 * varpool.c: Likewise.
5318 * tree.c: Remove functions moved to attribs.[ch].
5320 * config/aarch64/aarch64.c: Add attrs.h header file.
5321 * config/alpha/alpha.c: Likewise.
5322 * config/arc/arc.c: Likewise.
5323 * config/arm/arm.c: Likewise.
5324 * config/avr/avr.c: Likewise.
5325 * config/bfin/bfin.c: Likewise.
5326 * config/c6x/c6x.c: Likewise.
5327 * config/cr16/cr16.c: Likewise.
5328 * config/cris/cris.c: Likewise.
5329 * config/darwin.c: Likewise.
5330 * config/epiphany/epiphany.c: Likewise.
5331 * config/fr30/fr30.c: Likewise.
5332 * config/frv/frv.c: Likewise.
5333 * config/ft32/ft32.c: Likewise.
5334 * config/h8300/h8300.c: Likewise.
5335 * config/i386/winnt.c: Likewise.
5336 * config/ia64/ia64.c: Likewise.
5337 * config/iq2000/iq2000.c: Likewise.
5338 * config/lm32/lm32.c: Likewise.
5339 * config/m32c/m32c.c: Likewise.
5340 * config/m32r/m32r.c: Likewise.
5341 * config/m68k/m68k.c: Likewise.
5342 * config/mcore/mcore.c: Likewise.
5343 * config/microblaze/microblaze.c: Likewise.
5344 * config/mips/mips.c: Likewise.
5345 * config/mmix/mmix.c: Likewise.
5346 * config/mn10300/mn10300.c: Likewise.
5347 * config/moxie/moxie.c: Likewise.
5348 * config/msp430/msp430.c: Likewise.
5349 * config/nds32/nds32-isr.c: Likewise.
5350 * config/nds32/nds32.c: Likewise.
5351 * config/nios2/nios2.c: Likewise.
5352 * config/nvptx/nvptx.c: Likewise.
5353 * config/pa/pa.c: Likewise.
5354 * config/pdp11/pdp11.c: Likewise.
5355 * config/powerpcspe/powerpcspe.c: Likewise.
5356 * config/riscv/riscv.c: Likewise.
5357 * config/rl78/rl78.c: Likewise.
5358 * config/rx/rx.c: Likewise.
5359 * config/s390/s390.c: Likewise.
5360 * config/sh/sh.c: Likewise.
5361 * config/sol2.c: Likewise.
5362 * config/sparc/sparc.c: Likewise.
5363 * config/spu/spu.c: Likewise.
5364 * config/stormy16/stormy16.c: Likewise.
5365 * config/tilegx/tilegx.c: Likewise.
5366 * config/tilepro/tilepro.c: Likewise.
5367 * config/v850/v850.c: Likewise.
5368 * config/vax/vax.c: Likewise.
5369 * config/visium/visium.c: Likewise.
5370 * config/xtensa/xtensa.c: Likewise.
5372 2017-08-07 Michael Meissner <meissner@linux.vnet.ibm.com>
5375 * config/rs6000/vsx.md (vsx_concat_<mode>, VSX_D): Cleanup
5376 constraints since the -mupper-regs-* switches have been
5378 (vsx_concat_<mode>_1): New combiner insns to recognize inserting
5379 into a vector from a double word element that was extracted from
5380 another vector, and eliminate extra XXPERMDI instructions.
5381 (vsx_concat_<mode>_2): Likewise.
5382 (vsx_concat_<mode>_3): Likewise.
5383 (vsx_set_<mode>, VSX_D): Rewrite vector set in terms of vector
5384 concat to allow optimizing inserts from previous extracts.
5386 2017-08-07 Uros Bizjak <ubizjak@gmail.com>
5388 * config/i386/i386.c (ix86_stack_protect_guard): Generate
5389 memory reference to a SSP offset in TLS address space.
5390 (ix86_print_operand) <case '@'>: Remove.
5391 (ix86_print_operand_punct_valid_p): Remove '@' code.
5392 * config/i386/i386.md (unspec): Remove UNSPEC_SP_TLS_SET and
5394 (stack_tls_protect_set_<mode>): Remove.
5395 (stack_protect_set): Do not call gen_stack_tls_protect_set_<mode>.
5396 (stack_tls_protect_test_<mode>): Remove.
5397 (stack_protect_test): Do not call gen_stack_tls_protect_test_<mode>.
5399 2017-08-07 Olivier Hainque <hainque@adacore.com>
5402 * config/vxworksae.h (VXWORKS_HAVE_TLS): Define.
5404 2017-08-07 Douglas Rupp <rupp@adacore.com>
5406 * Makefile.in (install-mkheaders): Fix typo, where the multi_dir
5407 variable was referenced as multidir in command.
5409 2017-08-07 Jakub Jelinek <jakub@redhat.com>
5412 * gimplify.c (goa_stabilize_expr): Handle BIT_INSERT_EXPR and
5415 2017-08-07 Martin Liska <mliska@suse.cz>
5417 * config/m32c/m32c.c: Add include of stringpool.h and attribs.h.
5418 * config/rl78/rl78.c: Add include of attribs.h.
5419 * config/sh/sh.c: Likewise.
5420 * config/v850/v850.c: Likewise.
5422 2017-08-07 Tom de Vries <tom@codesourcery.com>
5425 * omp-expand.c (expand_oacc_for): Ensure diff_type is large enough.
5427 2017-08-07 Martin Liska <mliska@suse.cz>
5429 * config/mips/mips.c: Include attribs.h.
5431 2017-08-07 Thomas Koenig <tkoenig@gcc.gnu.org>
5434 * doc/invoke.texi: Document change in behvaior for -Ofast for
5437 2017-08-07 Wilco Dijkstra <wdijkstr@arm.com>
5439 * config/aarch64/aarch64.c (aarch64_pushwb_single_reg):
5441 (aarch64_pop_regs): Likewise.
5442 (aarch64_gen_load_pair): Likewise.
5443 (aarch64_save_callee_saves): Likewise.
5444 (aarch64_restore_callee_saves): Likewise.
5446 2017-08-07 H.J. Lu <hongjiu.lu@intel.com>
5448 * config/i386/i386.c: Revert the last change.
5450 2017-08-07 H.J. Lu <hongjiu.lu@intel.com>
5453 * config/i386/i386.c (ix86_finalize_stack_realign_flags): Renamed
5455 (ix86_finalize_stack_frame_flags): This. Also clear
5456 frame_pointer_needed if -fno-omit-frame-pointer is used without
5458 (ix86_expand_prologue): Replace ix86_finalize_stack_realign_flags
5459 with ix86_finalize_stack_frame_flags.
5460 (ix86_expand_epilogue): Likewise.
5461 (ix86_expand_split_stack_prologue): Likewise.
5463 2017-08-07 H.J. Lu <hongjiu.lu@intel.com>
5466 * config/i386/i386.c (get_builtin_code_for_version): Set priority
5467 to P_AES for Westmere.
5469 2017-08-07 Jonathan Yong <10walls@gmail.com>
5471 * config/i386/mingw.opt (fset-stack-executable): Removed.
5472 * config/i386/cygming.opt (fset-stack-executable): Moved
5474 * config/i386/cygwin.h: Define CHECK_EXECUTE_STACK_ENABLED.
5476 2017-08-07 Segher Boessenkool <segher@kernel.crashing.org>
5478 * print-rtl.c (print_exp): Print NOT as "~" instead of as "!".
5480 2017-08-07 Marek Polacek <polacek@redhat.com>
5483 * fold-const.c (fold_indirect_ref_1): Check type_domain.
5485 2017-08-07 Martin Liska <mliska@suse.cz>
5487 * attribs.h (canonicalize_attr_name): New function.
5488 (cmp_attribs): Move from c-format.c and adjusted.
5489 (is_attribute_p): Moved from tree.h.
5490 * tree-inline.c: Add new includes.
5491 * tree.c (cmp_attrib_identifiers): Use cmp_attribs.
5492 (private_is_attribute_p): Remove.
5493 (private_lookup_attribute): Likewise.
5494 (private_lookup_attribute_by_prefix): Simplify.
5495 (remove_attribute): Use is_attribute_p.
5496 * tree.h: Remove removed declarations.
5498 2017-08-07 Jakub Jelinek <jakub@redhat.com>
5501 * stmt.c (emit_case_dispatch_table): Add DEFAULT_EDGE argument,
5502 instead of computing it in the function. Formatting fix.
5503 (expand_case): Don't rely on default_edge being the first edge,
5504 clear it if removing it, pass default_edge to
5505 emit_case_dispatch_table.
5506 (expand_sjlj_dispatch_table): Pass NULL as DEFAULT_EDGE, formatting
5509 2017-08-06 Uros Bizjak <ubizjak@gmail.com>
5511 * config/alpha/alpha.c (alpha_reorg): If trap is the last active
5512 insn in the function, emit NOP after the insn.
5514 2017-08-06 Tom de Vries <tom@codesourcery.com>
5516 * omp-expand.c (expand_oacc_for): Add missing edge probability for tile
5519 2017-08-06 Tom de Vries <tom@codesourcery.com>
5521 * omp-expand.c (expand_oacc_for): Add missing edge probability for chunk
5524 2017-08-04 Yury Gribov <tetra2005@gmail.com>
5526 PR tree-optimization/57371
5527 * match.pd: New pattern.
5529 2017-08-04 Marek Polacek <polacek@redhat.com>
5532 * fold-const.c (fold_indirect_ref_1): For ((int *)&a + 4 -> a[1],
5533 perform the computation in offset_int.
5535 2017-08-04 Richard Sandiford <richard.sandiford@linaro.org>
5537 PR tree-optimization/81136
5538 * tree-vectorizer.h: Include tree-hash-traits.h.
5539 (vec_base_alignments): New typedef.
5540 (vec_info): Add a base_alignments field.
5541 (vect_record_base_alignments): Declare.
5542 * tree-data-ref.h (data_reference): Add an is_conditional_in_stmt
5544 (DR_IS_CONDITIONAL_IN_STMT): New macro.
5545 (create_data_ref): Add an is_conditional_in_stmt argument.
5546 * tree-data-ref.c (create_data_ref): Likewise. Use it to initialize
5547 the is_conditional_in_stmt field.
5548 (data_ref_loc): Add an is_conditional_in_stmt field.
5549 (get_references_in_stmt): Set the is_conditional_in_stmt field.
5550 (find_data_references_in_stmt): Update call to create_data_ref.
5551 (graphite_find_data_references_in_stmt): Likewise.
5552 * tree-ssa-loop-prefetch.c (determine_loop_nest_reuse): Likewise.
5553 * tree-vect-data-refs.c (vect_analyze_data_refs): Likewise.
5554 (vect_record_base_alignment): New function.
5555 (vect_record_base_alignments): Likewise.
5556 (vect_compute_data_ref_alignment): Adjust base_addr and aligned_to
5557 for nested statements even if we fail to compute a misalignment.
5558 Use pooled base alignments for unconditional references.
5559 (vect_find_same_alignment_drs): Compare base addresses instead
5561 (vect_analyze_data_refs_alignment): Call vect_record_base_alignments.
5562 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5564 2017-08-04 Richard Sandiford <richard.sandiford@linaro.org>
5566 * tree-vectorizer.h (vec_info): Add a constructor and destructor.
5567 Add an explicit name for the enum. Use auto_vec for slp_instances
5569 (_loop_vec_info): Add a constructor and destructor. Use auto_vec
5571 (_bb_vec_info): Add a constructor and destructor.
5572 (vinfo_for_stmt): Return NULL for uids of -1 as well.
5573 (destroy_loop_vec_info): Delete.
5574 (vect_destroy_datarefs): Likewise.
5575 * tree-vectorizer.c (vect_destroy_datarefs): Delete.
5576 (vec_info::vec_info): New function.
5577 (vec_info::~vec_info): Likewise.
5578 (vectorize_loops): Use delete instead of destroy_loop_vec_info.
5579 * tree-parloops.c (gather_scalar_reductions): Use delete instead of
5580 destroy_loop_vec_info.
5581 * tree-vect-loop.c (new_loop_vec_info): Replace with...
5582 (_loop_vec_info::_loop_vec_info): ...this.
5583 (destroy_loop_vec_info): Replace with...
5584 (_loop_vec_info::~_loop_vec_info): ...this. Unconditionally delete
5585 the stmt_vec_infos. Leave handling of vec_info information to its
5586 destructor. Remove explicit vector releases.
5587 (vect_analyze_loop_form): Use new instead of new_loop_vec_info.
5588 (vect_analyze_loop): Use delete instead of destroy_loop_vec_info.
5589 * tree-vect-slp.c (new_bb_vec_info): Replace with...
5590 (_bb_vec_info::_bb_vec_info): ...this. Don't reserve space in
5591 BB_VINFO_GROUPED_STORES or BB_VINFO_SLP_INSTANCES.
5592 (destroy_bb_vec_info): Replace with...
5593 (_bb_vec_info::~_bb_vec_info): ...this. Leave handling of vec_info
5594 information to its destructor.
5595 (vect_slp_analyze_bb_1): Use new and delete instead of
5596 new_bb_vec_info and destroy_bb_vec_info.
5597 (vect_slp_bb): Replace 2 calls to destroy_bb_vec_info with a
5600 2017-08-04 Richard Sandiford <richard.sandiford@linaro.org>
5602 * tree-data-ref.h (subscript): Add access_fn field.
5603 (data_dependence_relation): Add could_be_independent_p.
5604 (SUB_ACCESS_FN, DDR_COULD_BE_INDEPENDENT_P): New macros.
5605 (same_access_functions): Move to tree-data-ref.c.
5606 * tree-data-ref.c (ref_contains_union_access_p): New function.
5607 (access_fn_component_p): Likewise.
5608 (access_fn_components_comparable_p): Likewise.
5609 (dr_analyze_indices): Add a reference to access_fn_component_p.
5610 (dump_data_dependence_relation): Use SUB_ACCESS_FN instead of
5612 (constant_access_functions): Likewise.
5613 (add_other_self_distances): Likewise.
5614 (same_access_functions): Likewise. (Moved from tree-data-ref.h.)
5615 (initialize_data_dependence_relation): Use XCNEW and remove
5616 explicit zeroing of DDR_REVERSED_P. Look for a subsequence
5617 of access functions that have the same type. Allow the
5618 subsequence to end with different bases in some circumstances.
5619 Record the chosen access functions in SUB_ACCESS_FN.
5620 (build_classic_dist_vector_1): Replace ddr_a and ddr_b with
5621 a_index and b_index. Use SUB_ACCESS_FN instead of DR_ACCESS_FN.
5622 (subscript_dependence_tester_1): Likewise dra and drb.
5623 (build_classic_dist_vector): Update calls accordingly.
5624 (subscript_dependence_tester): Likewise.
5625 * tree-ssa-loop-prefetch.c (determine_loop_nest_reuse): Check
5626 DDR_COULD_BE_INDEPENDENT_P.
5627 * tree-vectorizer.h (LOOP_REQUIRES_VERSIONING_FOR_ALIAS): Test
5628 comp_alias_ddrs instead of may_alias_ddrs.
5629 * tree-vect-data-refs.c (vect_analyze_possibly_independent_ddr):
5631 (vect_analyze_data_ref_dependence): Use it if
5632 DDR_COULD_BE_INDEPENDENT_P, but fall back to using the recorded
5633 distance vectors if that fails.
5634 (dependence_distance_ge_vf): New function.
5635 (vect_prune_runtime_alias_test_list): Use it. Don't clear
5636 LOOP_VINFO_MAY_ALIAS_DDRS.
5638 2017-08-04 Richard Biener <rguenther@suse.de>
5641 * fold-const.c (fold_binary_loc): Properly restrict
5642 minus_var0 && minus_var1 case when associating undefined overflow
5645 2017-08-04 Tom de Vries <tom@codesourcery.com>
5647 * omp-simd-clone.c (simd_clone_adjust): Add missing edge probability.
5649 2017-08-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5651 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
5652 Don't start diagnostic messages with a capital letter.
5653 * config/rs6000/rs6000.c (rs6000_option_override_internal):
5655 (rs6000_invalid_builtin): Likewise.
5656 (rs6000_trampoline_init): Likewise.
5658 2017-08-03 Jakub Jelinek <jakub@redhat.com>
5661 * bb-reorder.c (pass_partition_blocks::execute): Return TODO_df_finish
5662 after setting changeable df flags.
5664 2017-08-03 Richard Biener <rguenther@suse.de>
5666 * tree-ssa-reassoc.c (should_break_up_subtract): Also break
5667 up if the use is in USE - X.
5669 2017-08-03 Alexander Monakov <amonakov@ispras.ru>
5671 * toplev.c (dumpfile.h): New include.
5672 (internal_error_reentered): New static function. Use it...
5673 (internal_error_function): ...here to handle reentered internal_error.
5675 2017-08-03 Richard Biener <rguenther@suse.de>
5678 * fold-const.c (split_tree): Add minus_var and minus_con
5679 arguments, remove unused loc arg. Never generate NEGATE_EXPRs
5680 here but always use minus_*.
5681 (associate_trees): Assert we never associate with MINUS_EXPR
5682 and NULL first operand. Do not recurse for PLUS_EXPR operands
5683 when associating as MINUS_EXPR either.
5684 (fold_binary_loc): Track minus_var and minus_con.
5686 2017-08-03 Tom de Vries <tom@codesourcery.com>
5689 * tree-streamer-in.c (lto_input_ts_function_decl_tree_pointers): If
5690 ACCEL_COMPILER, apply finish_options on
5691 DECL_FUNCTION_SPECIFIC_OPTIMIZATION.
5693 2017-08-03 Tom de Vries <tom@codesourcery.com>
5696 * config/nvptx/nvptx.c (nvptx_option_override): Emit sorry if
5697 function_entry_patch_area_size > 0.
5699 2017-08-03 Jakub Jelinek <jakub@redhat.com>
5702 * calls.c (alloc_max_size): Use HOST_WIDE_INT_UC (10??)
5703 instead of 10??LU, perform unit multiplication in wide_int,
5704 don't change alloc_object_size_limit if the limit is larger
5707 PR tree-optimization/81655
5708 PR tree-optimization/81588
5709 * tree-ssa-reassoc.c (optimize_range_tests_var_bound): Handle also
5710 the case when ranges[i].low and high are 1 for unsigned type with
5714 * omp-low.c (diagnose_sb_0): Handle flag_openmp_simd like flag_openmp.
5715 (pass_diagnose_omp_blocks::gate): Enable also for flag_openmp_simd.
5717 2017-08-03 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
5719 * tree-vrp.h: Add include guard.
5721 2017-08-02 Uros Bizjak <ubizjak@gmail.com>
5724 * config/i386/i386.md (unspecv): Add UNSPECV_UD2.
5725 (ud2): New insn pattern.
5726 * config/i386/i386.c (ix86_expand_epilogue):
5727 For naked functions, generate ud2 instead of trap insn.
5729 2017-08-02 Marek Polacek <polacek@redhat.com>
5732 * alloc-pool.h (base_pool_allocator): Initialize m_elt_size.
5734 2017-08-02 Tom de Vries <tom@codesourcery.com>
5735 Cesar Philippidis <cesar@codesourcery.com>
5737 * config/nvptx/nvptx.c (nvptx_lockless_update, nvptx_lockfull_update):
5738 Add missing edge probabilities.
5740 2017-08-02 Tamar Christina <tamar.christina@arm.com>
5742 * config/aarch64/aarch64.c (aarch64_reinterpret_float_as_int):
5745 2017-08-02 Jakub Jelinek <jakub@redhat.com>
5748 * function.c (thread_prologue_and_epilogue_insns): Determine blocks
5749 for find_many_sub_basic_blocks bitmap by looking up BLOCK_FOR_INSN
5750 of first NONDEBUG_INSN_P in each of the split_prologue_seq and
5751 prologue_seq sequences - if any.
5753 2017-08-02 Richard Biener <rguenther@suse.de>
5755 * tree-vect-stmts.c (vectorizable_store): Perform vector extracts
5756 via vectors if supported, integer extracts via punning if supported
5757 or otherwise vector extracts.
5759 2017-08-02 Richard Biener <rguenther@suse.de>
5761 * tree-ssa-pre.c (bitmap_insert_into_set_1): Remove and inline
5763 (bitmap_insert_into_set): ... this.
5765 2017-08-02 Richard Biener <rguenther@suse.de>
5767 PR tree-optimization/81633
5769 2015-08-17 Alan Hayward <alan.hayward@arm.com>
5771 PR tree-optimization/71752
5772 * tree-vect-slp.c (vect_get_slp_defs): Handle null operands.
5774 2017-08-01 Daniel Santos <daniel.santos@pobox.com>
5776 * config/i386/i386.h (ix86_frame::outlined_save_offset): Remove field.
5777 (machine_function::call_ms2sysv_pad_out): Remove field.
5778 * config/i386/i386.c (xlogue_layout::get_stack_space_used): Modify.
5779 (ix86_compute_frame_layout): Likewise.
5781 2017-08-01 H.J. Lu <hongjiu.lu@intel.com>
5784 * config/i386/i386.c (ix86_set_func_type): Disallow naked
5785 attribute with interrupt attribute.
5787 2017-08-01 Andrew Pinski <apinski@cavium.com>
5789 * tree-ssa-scopedtables.c (hashable_expr_equal_p): Check
5790 BIT_INSERT_EXPR's operand 1
5791 to see if the types precision matches.
5793 2017-08-01 Martin Liska <mliska@suse.cz>
5796 * builtins.c (expand_builtin_memcpy_args): Remove.
5797 (expand_builtin_memcpy): Call newly added function
5798 expand_builtin_memory_copy_args.
5799 (expand_builtin_memcpy_with_bounds): Likewise.
5800 (expand_builtin_mempcpy): Remove last argument.
5801 (expand_builtin_mempcpy_with_bounds): Likewise.
5802 (expand_builtin_memory_copy_args): New function created from
5803 expand_builtin_mempcpy_args with small modifications.
5804 (expand_builtin_mempcpy_args): Remove.
5805 (expand_builtin_stpcpy): Remove unused argument.
5806 (expand_builtin): Likewise.
5807 (expand_builtin_with_bounds): Likewise.
5809 2017-08-01 Martin Liska <mliska@suse.cz>
5812 Make mempcpy more optimal (PR middle-end/70140).
5814 2017-08-01 Jakub Jelinek <jakub@redhat.com>
5817 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): For
5818 __builtin_vec_cmpne verify both arguments are compatible vectors
5819 before looking at TYPE_MODE on the element type. For __builtin_vec_ld
5820 verify arg1_type is a pointer or array type. For __builtin_vec_st,
5821 move computation of aligned to after checking the argument types.
5825 * config/rs6000/vsx.md (vextract_fp_from_shorth,
5826 vextract_fp_from_shortl): Add element mode after mode in gen_vec_init*
5829 2017-08-01 Jerome Lambourg <lambourg@adacore.com>
5830 Doug Rupp <rupp@adacore.com>
5831 Olivier Hainque <hainque@adacore.com>
5833 * config.gcc (arm-wrs-vxworks*): Rework to handle arm-wrs-vxworks7 as
5834 well as arm-wrs-vxworks. Update target_cpu_name from arm6 (arch v3) to
5836 * config/arm/vxworks.h (MAYBE_TARGET_BPABI_CPP_BUILTINS): New, helper
5837 for TARGET_OS_CPP_BUILTIN.
5838 (TARGET_OS_CPP_BUILTIN): Invoke MAYBE_TARGET_BPABI_CPP_BUILTINS(),
5839 refine CPU definitions for arm_arch5 and add those for arm_arch6 and
5841 (MAYBE_ASM_ABI_SPEC): New, helper for SUBTARGET_EXTRA_ASM_SPEC,
5842 passing required abi options to the assembler for EABI configurations.
5843 (EXTRA_CC1_SPEC): New macro, to help prevent the implicit production
5844 of .text.hot and .text.unlikely sections for kernel modules when
5845 using ARM style exceptions.
5846 (CC1_SPEC): Remove obsolete attempt at mimicking Diab toolchain
5847 options. Add EXTRA_CC1_SPEC.
5848 (VXWORKS_ENDIAN_SPEC): Adjust comment and remove handling of Diab
5850 (DWARF2_UNWIND_INFO): Redefine to handle the pre/post VxWorks 7
5852 (ARM_TARGET2_DWARF_FORMAT): Define.
5853 * config/arm/t-vxworks: Adjust multilib control to removal of the
5854 Diab command line options.
5856 2017-08-01 Martin Liska <mliska@suse.cz>
5858 PR gcov-profile/81561
5859 * gcov.c (unblock): Make unblocking safe as we need to preserve
5860 index correspondence of blocks and block_lists.
5862 2017-08-01 Richard Biener <rguenther@suse.de>
5864 PR tree-optimization/81181
5865 * tree-ssa-pre.c (compute_antic_aux): Defer clean() to ...
5866 (compute_antic): ... end of iteration here.
5868 2017-08-01 James Greenhalgh <james.greenhalgh@arm.com>
5870 * common.opt (ftree-vectorize): No longer set flag_tree_vectorize.
5871 (ftree-loop-vectorize): Set as EnabledBy ftree-vectorize.
5872 (ftree-slp-vectorize): Likewise.
5873 * omp-expand (expand_omp_simd): Remove flag_tree_vectorize, as it
5874 can no longer be set independent of flag_tree_loop_vectorize.
5875 * omp-general.c (emp_max_vf): Likewise.
5876 * opts.c (enable_fdo_optimizations): Remove references to
5877 flag_tree_vectorize, these are now implicit.
5878 (common_handle_option): Remove handling for OPT_ftree_vectorize,
5879 and leave it for the options machinery.
5881 2017-08-01 Martin Liska <mliska@suse.cz>
5884 * builtins.c (expand_builtin_memcpy_args): Remove.
5885 (expand_builtin_memcpy): Call newly added function
5886 expand_builtin_memory_copy_args.
5887 (expand_builtin_memcpy_with_bounds): Likewise.
5888 (expand_builtin_mempcpy): Remove last argument.
5889 (expand_builtin_mempcpy_with_bounds): Likewise.
5890 (expand_builtin_memory_copy_args): New function created from
5891 expand_builtin_mempcpy_args with small modifications.
5892 (expand_builtin_mempcpy_args): Remove.
5893 (expand_builtin_stpcpy): Remove unused argument.
5894 (expand_builtin): Likewise.
5895 (expand_builtin_with_bounds): Likewise.
5897 2017-08-01 Uros Bizjak <ubizjak@gmail.com>
5900 * config/i386/i386.c (ix86_print_operand_address_as): For -masm=intel
5901 print "ds:" only for immediates in generic address space.
5903 2017-08-01 Uros Bizjak <ubizjak@gmail.com>
5906 * config/i386/i386.c (ix86_funciton_naked): New prototype.
5907 (ix86_function_ok_for_sibcall): Return false for naked functions.
5909 2017-08-01 Richard Biener <rguenther@suse.de>
5911 * tree-ssa-pre.c (print_pre_expr): Handle NULL expr.
5912 (compute_antic): Seed worklist with exit block predecessors.
5913 * cfganal.c (dfs_find_deadend): For a cycle return the source
5914 of the edge closing it.
5916 2017-08-01 Tamar Christina <tamar.christina@arm.com>
5918 * config/aarch64/aarch64.c
5919 (aarch64_can_const_movi_rtx_p): Move 0 check.
5921 2017-08-01 Bin Cheng <bin.cheng@arm.com>
5923 * tree.h (POINTER_TYPE_OVERFLOW_UNDEFINED): Delete.
5924 * fold-const.c (fold_comparison, fold_binary_loc): Delete use of
5926 * match.pd: Ditto in address comparison pattern.
5928 2017-08-01 Bin Cheng <bin.cheng@arm.com>
5930 PR tree-optimization/81627
5931 * tree-predcom.c (prepare_finalizers): Always rewrite into loop
5932 closed ssa form for store-store chain.
5934 2017-08-01 Bin Cheng <bin.cheng@arm.com>
5936 PR tree-optimization/81620
5937 * tree-predcom.c (add_ref_to_chain): Don't set has_max_use_after
5938 for store-store chain.
5940 2017-08-01 Jakub Jelinek <jakub@redhat.com>
5942 PR tree-optimization/81588
5943 * tree-ssa-reassoc.c (optimize_range_tests_var_bound): If
5944 ranges[i].in_p, invert comparison code ccode. For >/>=,
5945 swap rhs1 and rhs2 and comparison code unconditionally,
5946 for </<= don't do that. Don't swap rhs1/rhs2 again if
5947 ranges[i].in_p, instead invert comparison code ccode if
5948 opcode or oe->rank is BIT_IOR_EXPR.
5951 * optabs.def (vec_extract_optab, vec_init_optab): Change from
5952 a direct optab to conversion optab.
5953 * optabs.c (expand_vector_broadcast): Use convert_optab_handler
5954 with GET_MODE_INNER as last argument instead of optab_handler.
5955 * expmed.c (extract_bit_field_1): Likewise. Use vector from
5956 vector extraction if possible and optab is available.
5957 * expr.c (store_constructor): Use convert_optab_handler instead
5958 of optab_handler. Use vector initialization from smaller
5959 vectors if possible and optab is available.
5960 * tree-vect-stmts.c (vectorizable_load): Likewise.
5961 * doc/md.texi (vec_extract, vec_init): Document that the optabs
5963 * config/i386/i386.c (ix86_expand_vector_init): Handle expansion
5964 of vec_init from half-sized vectors with the same element mode.
5965 * config/i386/sse.md (ssehalfvecmode): Add V4TI case.
5966 (ssehalfvecmodelower, ssescalarmodelower): New mode attributes.
5967 (reduc_plus_scal_v8df, reduc_plus_scal_v4df, reduc_plus_scal_v2df,
5968 reduc_plus_scal_v16sf, reduc_plus_scal_v8sf, reduc_plus_scal_v4sf,
5969 reduc_<code>_scal_<mode>, reduc_umin_scal_v8hi): Add element mode
5970 after mode in gen_vec_extract* calls.
5971 (vec_extract<mode>): Renamed to ...
5972 (vec_extract<mode><ssescalarmodelower>): ... this.
5973 (vec_extract<mode><ssehalfvecmodelower>): New expander.
5974 (rotl<mode>3, rotr<mode>3, <shift_insn><mode>3, ashrv2di3): Add
5975 element mode after mode in gen_vec_init* calls.
5976 (VEC_INIT_HALF_MODE): New mode iterator.
5977 (vec_init<mode>): Renamed to ...
5978 (vec_init<mode><ssescalarmodelower>): ... this.
5979 (vec_init<mode><ssehalfvecmodelower>): New expander.
5980 * config/i386/mmx.md (vec_extractv2sf): Renamed to ...
5981 (vec_extractv2sfsf): ... this.
5982 (vec_initv2sf): Renamed to ...
5983 (vec_initv2sfsf): ... this.
5984 (vec_extractv2si): Renamed to ...
5985 (vec_extractv2sisi): ... this.
5986 (vec_initv2si): Renamed to ...
5987 (vec_initv2sisi): ... this.
5988 (vec_extractv4hi): Renamed to ...
5989 (vec_extractv4hihi): ... this.
5990 (vec_initv4hi): Renamed to ...
5991 (vec_initv4hihi): ... this.
5992 (vec_extractv8qi): Renamed to ...
5993 (vec_extractv8qiqi): ... this.
5994 (vec_initv8qi): Renamed to ...
5995 (vec_initv8qiqi): ... this.
5996 * config/rs6000/vector.md (VEC_base_l): New mode attribute.
5997 (vec_init<mode>): Renamed to ...
5998 (vec_init<mode><VEC_base_l>): ... this.
5999 (vec_extract<mode>): Renamed to ...
6000 (vec_extract<mode><VEC_base_l>): ... this.
6001 * config/rs6000/paired.md (vec_initv2sf): Renamed to ...
6002 (vec_initv2sfsf): ... this.
6003 * config/rs6000/altivec.md (splitter, altivec_copysign_v4sf3,
6004 vec_unpacku_hi_v16qi, vec_unpacku_hi_v8hi, vec_unpacku_lo_v16qi,
6005 vec_unpacku_lo_v8hi, mulv16qi3, altivec_vreve<mode>2): Add
6006 element mode after mode in gen_vec_init* calls.
6007 * config/aarch64/aarch64-simd.md (vec_init<mode>): Renamed to ...
6008 (vec_init<mode><Vel>): ... this.
6009 (vec_extract<mode>): Renamed to ...
6010 (vec_extract<mode><Vel>): ... this.
6011 * config/aarch64/iterators.md (Vel): New mode attribute.
6012 * config/s390/s390.c (s390_expand_vec_strlen, s390_expand_vec_movstr):
6013 Add element mode after mode in gen_vec_extract* calls.
6014 * config/s390/vector.md (non_vec_l): New mode attribute.
6015 (vec_extract<mode>): Renamed to ...
6016 (vec_extract<mode><non_vec_l>): ... this.
6017 (vec_init<mode>): Renamed to ...
6018 (vec_init<mode><non_vec_l>): ... this.
6019 * config/s390/s390-builtins.def (s390_vlgvb, s390_vlgvh, s390_vlgvf,
6020 s390_vlgvf_flt, s390_vlgvg, s390_vlgvg_dbl): Add element mode after
6022 * config/arm/iterators.md (V_elem_l): New mode attribute.
6023 * config/arm/neon.md (vec_extract<mode>): Renamed to ...
6024 (vec_extract<mode><V_elem_l>): ... this.
6025 (vec_extractv2di): Renamed to ...
6026 (vec_extractv2didi): ... this.
6027 (vec_init<mode>): Renamed to ...
6028 (vec_init<mode><V_elem_l>): ... this.
6029 (reduc_plus_scal_<mode>, reduc_plus_scal_v2di, reduc_smin_scal_<mode>,
6030 reduc_smax_scal_<mode>, reduc_umin_scal_<mode>,
6031 reduc_umax_scal_<mode>, neon_vget_lane<mode>, neon_vget_laneu<mode>):
6032 Add element mode after gen_vec_extract* calls.
6033 * config/mips/mips-msa.md (vec_init<mode>): Renamed to ...
6034 (vec_init<mode><unitmode>): ... this.
6035 (vec_extract<mode>): Renamed to ...
6036 (vec_extract<mode><unitmode>): ... this.
6037 * config/mips/loongson.md (vec_init<mode>): Renamed to ...
6038 (vec_init<mode><unitmode>): ... this.
6039 * config/mips/mips-ps-3d.md (vec_initv2sf): Renamed to ...
6040 (vec_initv2sfsf): ... this.
6041 (vec_extractv2sf): Renamed to ...
6042 (vec_extractv2sfsf): ... this.
6043 (reduc_plus_scal_v2sf, reduc_smin_scal_v2sf, reduc_smax_scal_v2sf):
6044 Add element mode after gen_vec_extract* calls.
6045 * config/mips/mips.md (unitmode): New mode iterator.
6046 * config/spu/spu.c (spu_expand_prologue, spu_allocate_stack,
6047 spu_builtin_extract): Add element mode after gen_vec_extract* calls.
6048 * config/spu/spu.md (inner_l): New mode attribute.
6049 (vec_init<mode>): Renamed to ...
6050 (vec_init<mode><inner_l>): ... this.
6051 (vec_extract<mode>): Renamed to ...
6052 (vec_extract<mode><inner_l>): ... this.
6053 * config/sparc/sparc.md (veltmode): New mode iterator.
6054 (vec_init<VMALL:mode>): Renamed to ...
6055 (vec_init<VMALL:mode><VMALL:veltmode>): ... this.
6056 * config/ia64/vect.md (vec_initv2si): Renamed to ...
6057 (vec_initv2sisi): ... this.
6058 (vec_initv2sf): Renamed to ...
6059 (vec_initv2sfsf): ... this.
6060 (vec_extractv2sf): Renamed to ...
6061 (vec_extractv2sfsf): ... this.
6062 * config/powerpcspe/vector.md (VEC_base_l): New mode attribute.
6063 (vec_init<mode>): Renamed to ...
6064 (vec_init<mode><VEC_base_l>): ... this.
6065 (vec_extract<mode>): Renamed to ...
6066 (vec_extract<mode><VEC_base_l>): ... this.
6067 * config/powerpcspe/paired.md (vec_initv2sf): Renamed to ...
6068 (vec_initv2sfsf): ... this.
6069 * config/powerpcspe/altivec.md (splitter, altivec_copysign_v4sf3,
6070 vec_unpacku_hi_v16qi, vec_unpacku_hi_v8hi, vec_unpacku_lo_v16qi,
6071 vec_unpacku_lo_v8hi, mulv16qi3): Add element mode after mode in
6072 gen_vec_init* calls.
6074 2017-08-01 Richard Biener <rguenther@suse.de>
6076 PR tree-optimization/81297
6077 * tree-vrp.c (get_single_symbol): Remove assert, instead drop
6078 TREE_OVERFLOW from INTEGER_CSTs.
6080 2017-07-31 Segher Boessenkool <segher@kernel.crashing.org>
6082 * config/rs6000/rs6000.c (enum rs6000_reg_type): Delete trailing comma.
6084 2017-07-31 Carl Love <cel@us.ibm.com>
6086 * config/rs6000/rs6000-c: Add support for built-in functions
6087 vector signed char vec_xl_be (signed long long, signed char *);
6088 vector unsigned char vec_xl_be (signed long long, unsigned char *);
6089 vector signed int vec_xl_be (signed long long, signed int *);
6090 vector unsigned int vec_xl_be (signed long long, unsigned int *);
6091 vector signed long long vec_xl_be (signed long long, signed long long *);
6092 vector unsigned long long vec_xl_be (signed long long, unsigned long long *);
6093 vector signed short vec_xl_be (signed long long, signed short *);
6094 vector unsigned short vec_xl_be (signed long long, unsigned short *);
6095 vector double vec_xl_be (signed long long, double *);
6096 vector float vec_xl_be (signed long long, float *);
6097 * config/rs6000/altivec.h (vec_xl_be): Add #define.
6098 * config/rs6000/rs6000-builtin.def (XL_BE_V16QI, XL_BE_V8HI, XL_BE_V4SI)
6099 XL_BE_V2DI, XL_BE_V4SF, XL_BE_V2DF, XL_BE): Add definitions
6101 * config/rs6000/rs6000.c (altivec_expand_xl_be_builtin): Add function.
6102 (altivec_expand_builtin): Add switch statement to call
6103 altivec_expand_xl_be for each builtin.
6104 (altivec_init_builtins): Add def_builtin for _builtin_vsx_le_be_v8hi,
6105 __builtin_vsx_le_be_v4si, __builtin_vsx_le_be_v2di,
6106 __builtin_vsx_le_be_v4sf, __builtin_vsx_le_be_v2df,
6107 __builtin_vsx_le_be_v16qi.
6108 * doc/extend.texi: Update the built-in documentation file for the
6109 new built-in functions.
6111 2017-07-31 Uros Bizjak <ubizjak@gmail.com>
6114 * config/i386/i386.c (ix86_allocate_stack_slots_for_args):
6116 (TARGET_ALLOCATE_STACK_SLOTS_FOR_ARGS): Define.
6118 2017-07-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
6120 * config.gcc: Add z14.
6121 * config/s390/driver-native.c (s390_host_detect_local_cpu): Add
6122 CPU model numbers for z13s and z14.
6123 * config/s390/s390-c.c (s390_resolve_overloaded_builtin): Replace
6125 * config/s390/s390-opts.h (enum processor_type): Rename
6126 PROCESSOR_ARCH12 to PROCESSOR_3906_Z14.
6127 * config/s390/s390.c (processor_table): Add field for CPU name to
6128 be passed to Binutils.
6129 (s390_asm_output_machine_for_arch): Use the new field in
6130 processor_table for Binutils.
6131 (s390_expand_builtin): Replace arch12 with z14.
6132 (s390_issue_rate): Rename PROCESSOR_ARCH12 to PROCESSOR_3906_Z14.
6133 (s390_get_sched_attrmask): Likewise.
6134 (s390_get_unit_mask): Likewise.
6135 * config/s390/s390.opt: Add z14 to processor_type enum.
6137 2017-07-31 Martin Jambor <mjambor@suse.cz>
6140 * ipa-fnsummary.c (ipa_fn_summary_generate): Set versionable
6141 regardless of optimization level.
6143 2017-07-31 Jan Hubicka <hubicka@ucw.cz>
6144 Martin Liska <mliska@suse.cz>
6146 * predict.def: Remove old comment and adjust probability.
6147 * gimplify.c (should_warn_for_implicit_fallthrough): Ignore
6150 2017-07-31 Uros Bizjak <ubizjak@gmail.com>
6153 * config/i386/i386.c (ix86_function_naked): New function.
6154 (ix86_can_use_return_insn_p): Return false for naked functions.
6155 (ix86_expand_prologue): Skip prologue for naked functions.
6156 (ix86_expand_epilogue): Skip epilogue for naked functions
6157 and emit trap instruction.
6158 (ix86_warn_func_return): New function.
6159 (ix86_attribute_table): Add "naked" attribute specification.
6160 (TARGET_WARN_FUNC_RETURN): Define.
6161 * doc/extend.texi (x86 Function Attributes) <naked>: Document it.
6163 2017-07-31 Martin Liska <mliska@suse.cz>
6165 * gimple-pretty-print.c (dump_gimple_label): Never dump BB info.
6166 (dump_gimple_bb_header): Always dump BB info.
6167 (pp_cfg_jump): Do not append info about BB when dumping a jump.
6169 2017-07-31 Martin Liska <mliska@suse.cz>
6172 * convert.c (convert_to_integer_1): Guard condition with flag_sanitize_p
6173 also with current_function_decl non-null equality.
6175 2017-07-31 Jakub Jelinek <jakub@redhat.com>
6178 * ubsan.c (ubsan_type_descriptor): For UBSAN_PRINT_ARRAY don't
6179 change type to the element type, instead add eltype variable and
6180 use it where we are interested in the element type.
6182 PR tree-optimization/81603
6183 * ipa-polymorphic-call.c
6184 (ipa_polymorphic_call_context::ipa_polymorphic_call_context): Perform
6185 offset arithmetic in offset_int, bail out if the resulting bit offset
6186 doesn't fit into shwi.
6188 2017-07-31 Martin Liska <mliska@suse.cz>
6190 * gimplify.c (mostly_copy_tree_r): Remove Java specific hunk.
6191 (gimplify_save_expr): Fix comment.
6193 2017-07-30 H.J. Lu <hongjiu.lu@intel.com>
6196 * config/i386/i386.c (ix86_function_arg): Update arguments for
6198 (ix86_compute_frame_layout): Set the initial stack offset to
6199 INCOMING_FRAME_SP_OFFSET. Update red-zone offset with
6200 INCOMING_FRAME_SP_OFFSET.
6201 (ix86_expand_epilogue): Don't pop the 'ERROR_CODE' off the
6202 stack before exception handler returns.
6203 * config/i386/i386.h (INCOMING_FRAME_SP_OFFSET): Add the
6204 the 'ERROR_CODE' for exception handler.
6206 2017-07-30 Uros Bizjak <ubizjak@gmail.com>
6208 * config/i386/i386.h (ASM_PRINTF_EXTENSIONS): New macro.
6209 (ASM_OUTPUT_REG_PUSH): Rewrite with new operand modifiers.
6210 (ASM_OUTPUT_REG_POP): Ditto.
6211 * config/i386/i386.c (ix86_asm_output_function_label): Use fputs
6212 instead of asm_fprintf to output pure string.
6214 2017-07-29 Jakub Jelinek <jakub@redhat.com>
6216 * debug.h (struct gcc_debug_hooks): Add IMPLICIT argument
6217 to imported_module_or_decl hook.
6218 (debug_nothing_tree_tree_tree_bool): Remove.
6219 (debug_nothing_tree_tree_tree_bool_bool): New declaration.
6220 * debug.c (do_nothing_debug_hooks): Use
6221 debug_nothing_tree_tree_tree_bool_bool instead of
6222 debug_nothing_tree_tree_tree_bool.
6223 * vmsdbgout.c (vmsdbg_debug_hooks): Likewise.
6224 * dbxout.c (dbx_debug_hooks, xcoff_debug_hooks): Likewise.
6225 * sdbout.c (sdb_debug_hooks): Likewise.
6226 * dwarf2out.c (dwarf2_lineno_debug_hooks): Likewise.
6227 (gen_namespace_die): Add DW_AT_export_symbols attribute if
6229 (dwarf2out_imported_module_or_decl): Add IMPLICIT argument,
6230 if true, -gdwarf-5 and decl will have DW_AT_export_symbols
6231 attribute, don't add anything.
6233 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6235 * fold-const.c (fold_build1_stat_loc): Remove _stat from name.
6236 (fold_build2_stat_loc): Likewise.
6237 (fold_build3_stat_loc): Likewise.
6238 * fold-const.h (fold_build1, fold_build2, fold_build3): Adjust.
6239 (fold_build1_loc): Remove macro.
6240 (fold_build2_loc): Likewise.
6241 (fold_build3_loc): Likewise.
6243 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6245 * gimple.c (gimple_build_debug_bind_stat): Remove _stat from name.
6246 (gimple_build_debug_bind_source_stat): Likewise.
6247 * gimple.h (gimple_build_debug_bind): Remove macro.
6248 (gimple_build_debug_bind_source): Likewise.
6250 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6252 * bitmap.c (bitmap_alloc): Adjust.
6253 (bitmap_gc_alloc): Likewise.
6254 * bitmap.h (bitmap_initialize_stat): Remove _stat from name.
6256 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6258 * bitmap.c (bitmap_obstack_alloc_stat): Rename to bitmap_alloc.
6259 (bitmap_gc_alloc_stat): Rename to bitmap_gc_alloc.
6260 * bitmap.h (bitmap_obstack_alloc_stat): Adjust prototype.
6261 (bitmap_gc_alloc_stat): Likewise.
6262 (BITMAP_ALLOC, BITMAP_GGC_ALLOC): Adjust.
6264 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6266 * rtl.c (shallow_copy_rtx_stat): Remove _stat from name.
6267 * rtl.h (shallow_copy_rtx): Remove macro.
6269 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6271 * emit-rtl.c (gen_raw_REG): Adjust.
6272 * gengenrtl.c (gendef): Likewise.
6273 * rtl.c (rtx_alloc_stat): Remove _stat from name.
6274 * rtl.h (rtx_alloc): Remove macro.
6276 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6278 * tree.c (build_tree_list_vec_stat): Remove _stat from name.
6279 (build_tree_list_stat): Likewise.
6280 * tree.h (build_tree_list): Remove macro.
6281 (build_tree_list_vec): Likewise.
6283 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6285 * tree.c (make_vector_stat): Remove _stat from name.
6286 (build_vector_stat): Likewise.
6287 * tree.h (make_vector_stat): Remove macro.
6288 (build_vector_stat): Likewise.
6290 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6292 * tree.h (build_var_debug_value): Remove prototype.
6294 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6296 * tree.c (tree_cons_stat): Remove _stat from name.
6297 * tree.h (tree_cons): Remove macro.
6299 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6301 * tree.c (build_vl_exp_stat): Remove _stat from name.
6302 * tree.h (build_vl_exp): Remove macro.
6304 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6306 * tree.c (build_decl_stat): Remove _stat from name.
6307 * tree.h (build_decl): Remove macro.
6309 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6311 * gimple.c (gimple_build_with_ops_stat): Adjust.
6312 (gimple_alloc_stat): Remove _stat from name.
6313 * gimple.h (gimple_alloc): Remove macro.
6315 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6317 * tree.c (make_tree_vec_stat): Remove _stat from name.
6318 (grow_tree_vec_stat): Likewise.
6319 * tree.h (make_tree_vec_stat): Adjust prototype.
6320 (grow_tree_vec_stat): Likewise.
6321 (make_tree_vec): Remove macro.
6322 (grow_tree_vec): Likewise.
6324 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6326 * fold-const.c (fold_build1_stat_loc): Adjust.
6327 (fold_build2_stat_loc): Likewise.
6328 (fold_build3_stat_loc): Likewise.
6329 * tree.c (build0_stat): Remove _stat from name.
6330 (build1_stat): Likewise.
6331 (build2_stat): Likewise.
6332 (build3_stat): Likewise.
6333 (build4_stat): Likewise.
6334 (build5_stat): Likewise.
6335 * tree.h (build1_loc): Remove macro, and rename _stat function
6337 (build2_loc): Likewise.
6338 (build3_loc): Likewise.
6339 (build4_loc): Likewise.
6340 (build5_loc): Likewise.
6342 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6344 * tree.c (make_int_cst_stat): Remove _stat from name.
6345 * tree.h (make_int_cst_stat): Adjust prototype.
6346 (make_int_cst): Remove macro.
6348 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6350 * tree.c (make_tre_binfo_stat): Remove _stat from name.
6351 * tree.h (make_tree_binfo_stat): Adjust prototype.
6352 (make_tree_binfo): Remove.
6354 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6356 * tree.c (copy_node_stat): Rename to copy_node.
6357 (build_distinct_type_copy): Adjust.
6358 * tree.h (copy_node_stat): Adjust prototype.
6359 (copy_node): Remove macro.
6361 2017-07-28 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
6363 * tree.c (make_node_stat): rename to make_node.
6364 (build_tree_list_stat): Adjust.
6365 (build0_stat): Likewise.
6366 (build2_stat): Likewise.
6367 (build3_stat): Likewise.
6368 (build4_stat): Likewise.
6369 (build5_stat): Likewise.
6370 (build_decl_stat): Likewise.
6371 * tree.h (make_node_stat): Adjust prototype.
6372 (make_node): remove macro.
6374 2017-07-28 Peter Bergner <bergner@vnet.ibm.com>
6376 * config/rs6000/ppc-auxv.h (PPC_FEATURE2_DARN): New define.
6377 (PPC_FEATURE2_SCV): Likewise.
6378 * config/rs6000/rs6000.c (cpu_supports_info): Use them.
6380 2017-07-28 Tamar Christina <tamar.christina@arm.com>
6382 * config/aarch64/aarch64.c
6383 (aarch64_internal_mov_immediate): Add new special pattern.
6384 * config/aarch64/aarch64.md (*movdi_aarch64):
6385 Add reg/32bit const mov case.
6387 2017-07-28 Tamar Christina <tamar.christina@arm.com>
6388 Richard Sandiford <richard.sandiford@linaro.org>
6390 * config/aarch64/aarch64.md (mov<mode>): Generalize.
6391 (*movhf_aarch64, *movsf_aarch64, *movdf_aarch64):
6392 Add integer and movi cases.
6393 (movi-split-hf-df-sf split, fp16): New.
6394 (enabled): Added TARGET_FP_F16INST.
6395 * config/aarch64/iterators.md (GPF_HF): New.
6396 * config/aarch64/predicates.md (aarch64_reg_or_fp_float): New.
6398 2017-07-28 Tamar Christina <tamar.christina@arm.com>
6400 * config/aarch64/aarch64.c
6401 (aarch64_simd_container_mode): Add prototype.
6402 (aarch64_expand_mov_immediate): Add HI support.
6403 (aarch64_reinterpret_float_as_int, aarch64_float_const_rtx_p: New.
6404 (aarch64_can_const_movi_rtx_p): New.
6405 (aarch64_preferred_reload_class):
6406 Remove restrictions of using FP registers for certain SIMD operations.
6407 (aarch64_rtx_costs): Added new cost for CONST_DOUBLE moves.
6408 (aarch64_valid_floating_const): Add integer move validation.
6409 (aarch64_simd_imm_scalar_p): Remove.
6410 (aarch64_output_scalar_simd_mov_immediate): Generalize function.
6411 (aarch64_legitimate_constant_p): Expand list of supported cases.
6412 * config/aarch64/aarch64-protos.h
6413 (aarch64_float_const_rtx_p, aarch64_can_const_movi_rtx_p): New.
6414 (aarch64_reinterpret_float_as_int): New.
6415 (aarch64_simd_imm_scalar_p): Remove.
6416 * config/aarch64/constraints.md (Uvi): New.
6417 (Dd): Split into Ds and new Dd.
6418 * config/aarch64/aarch64.md (*movsi_aarch64):
6420 (*movdi_aarch64): Add SIMD mov case.
6422 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6424 * tree-predcom.c: (struct chain): Handle store-store chain in which
6425 stores for elimination only store loop invariant values.
6426 (execute_pred_commoning_chain): Ditto.
6427 (prepare_initializers_chain_store_elim): Ditto.
6428 (prepare_finalizers): Ditto.
6429 (is_inv_store_elimination_chain): New function.
6430 (initialize_root_vars_store_elim_1): New function.
6432 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6434 * tree-predcom.c: Revise general description of the pass.
6435 (enum chain_type): New enum type for store elimination.
6436 (struct chain): New field supporting store elimination.
6437 (struct component): Ditto.
6438 (dump_chain): Dump store-stores chain.
6439 (release_chain): Release resources.
6440 (split_data_refs_to_components): Compute and create component
6441 contains only stores for elimination.
6442 (get_chain_last_ref_at): New function.
6443 (make_invariant_chain): Initialization.
6444 (make_rooted_chain): Specify chain type in parameter and record it.
6445 (add_looparound_copies): Skip for store-stores chain.
6446 (determine_roots_comp): Compute type of chain and pass it to
6448 (initialize_root_vars_store_elim_2): New function.
6449 (finalize_eliminated_stores): New function.
6450 (remove_stmt): Handle store for elimination.
6451 (execute_pred_commoning_chain): Execute predictive commoning on
6453 (determine_unroll_factor): Skip unroll for store-stores chain.
6454 (prepare_initializers_chain_store_elim): New function.
6455 (prepare_initializers_chain): Hanlde store-store chain.
6456 (prepare_finalizers_chain, prepare_finalizers): New function.
6457 (tree_predictive_commoning_loop): Return integer value indicating
6458 if loop is unrolled or lcssa form is corrupted.
6459 (tree_predictive_commoning): Rewrite for lcssa form if necessary.
6461 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6463 * tree-predcom.c (initialize_root): Delete.
6464 (execute_pred_commoning_chain): Initialize root vars and replace
6465 reference of non-combined chain directly, rather than call above
6468 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6470 * tree-predcom.c (ref_at_iteration): Add parameter NITERS. Compute
6471 memory reference to DR at (NITERS + ITERS)-th iteration of loop.
6473 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6475 * tree-predcom.c (struct chain): New field init_seq.
6476 (release_chain): Release init_seq.
6477 (prepare_initializers_chain): Record intialization stmts in above
6479 (insert_init_seqs): New function.
6480 (tree_predictive_commoning_loop): Call insert_init_seqs.
6482 2017-07-28 Bin Cheng <bin.cheng@arm.com>
6484 * tree-predcom.c (determine_roots_comp): Skip trivial components.
6486 2017-07-28 Richard Biener <rguenther@suse.de>
6488 * match.pd: Remove superfluous :c.
6489 * genmatch.c (simplify::id): Add member.
6490 (lower_commutative, lower_opt_convert, lower_cond, lower_for):
6492 (current_id): New global.
6493 (dt_node::parent): Move from ...
6494 (dt_operand::parent): ... here. Add for_id member.
6495 (is_a_helper <dt_operand *>::test): DT_TRUE is also a dt_operand.
6496 (decision_tree::find_node): Relax order requirement when
6497 merging DT_TRUE nodes to ones inbetween the current simplify
6498 and the one we try to merge with. Add diagnostic whenever
6499 we need to enforce pattern order by not merging.
6500 (decision_tree::insert): Set current_id.
6501 (decision_tree::print_node): Dump parent node and for_id.
6502 (parser::last_id): Add member.
6503 (parser::push_simplify): Assign unique id.
6504 (parser::parser): Initialize last_id.
6506 2017-07-28 Martin Liska <mliska@suse.cz>
6509 * sanopt.c (sanitize_rewrite_addressable_params): Set VALUE_EXPR after
6510 gimple_build_debug_bind.
6512 2017-07-28 Richard Biener <rguenther@suse.de>
6514 PR tree-optimization/81502
6515 * match.pd: Add pattern combining BIT_INSERT_EXPR with
6517 * tree-cfg.c (verify_expr): Verify types of BIT_FIELD_REF
6519 (verify_gimple_assign_ternary): Likewise for BIT_INSERT_EXPR pos.
6520 * gimple-fold.c (maybe_canonicalize_mem_ref_addr): Use bitsizetype
6521 for BIT_FIELD_REF args.
6522 * fold-const.c (make_bit_field_ref): Likewise.
6523 * tree-vect-stmts.c (vectorizable_simd_clone_call): Likewise.
6525 2017-07-28 Jakub Jelinek <jakub@redhat.com>
6528 * sanopt.c (pass_sanopt::execute): Handle IFN_UBSAN_PTR.
6529 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Likewise.
6530 * flag-types.h (enum sanitize_code): Add SANITIZER_POINTER_OVERFLOW.
6531 Or it into SANITIZER_UNDEFINED.
6532 * ubsan.c: Include gimple-fold.h and varasm.h.
6533 (ubsan_expand_ptr_ifn): New function.
6534 (instrument_pointer_overflow): New function.
6535 (maybe_instrument_pointer_overflow): New function.
6536 (instrument_object_size): Formatting fix.
6537 (pass_ubsan::execute): Call instrument_pointer_overflow
6538 and maybe_instrument_pointer_overflow.
6539 * internal-fn.c (expand_UBSAN_PTR): New function.
6540 * ubsan.h (ubsan_expand_ptr_ifn): Declare.
6541 * sanitizer.def (__ubsan_handle_pointer_overflow,
6542 __ubsan_handle_pointer_overflow_abort): New builtins.
6543 * tree-ssa-tail-merge.c (merge_stmts_p): Handle IFN_UBSAN_PTR.
6544 * internal-fn.def (UBSAN_PTR): New internal function.
6545 * opts.c (sanitizer_opts): Add pointer-overflow.
6546 * lto-streamer-in.c (input_function): Handle IFN_UBSAN_PTR.
6547 * fold-const.c (build_range_check): Compute pointer range check in
6548 integral type if pointer arithmetics would be needed. Formatting
6551 2017-07-28 Martin Liska <mliska@suse.cz>
6554 * sanopt.c (sanitize_rewrite_addressable_params): Do not rewrite
6555 parameters that are of a variable-length.
6557 2017-07-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
6559 * config.gcc (powerpc-*-rtems*): Remove rs6000/eabi.h. Add
6561 * config/rs6000/rtems.h (ASM_DECLARE_FUNCTION_SIZE): New macro.
6562 (ASM_OUTPUT_SPECIAL_POOL_ENTRY_P): Likewise.
6563 (CRT_CALL_STATIC_FUNCTION): Likewise.
6564 (ASM_DEFAULT_SPEC): New define.
6565 (ASM_SPEC32): Likewise.
6566 (ASM_SPEC64): Likewise.
6567 (ASM_SPEC_COMMON): Likewise.
6568 (ASM_SPEC): Likewise.
6569 (INVALID_64BIT): Likewise.
6570 (LINK_OS_DEFAULT_SPEC): Likewise.
6571 (LINK_OS_SPEC32): Likewise.
6572 (LINK_OS_SPEC64): Likewise.
6573 (POWERPC_LINUX): Likewise.
6574 (PTRDIFF_TYPE): Likewise.
6575 (RESTORE_FP_PREFIX): Likewise.
6576 (RESTORE_FP_SUFFIX): Likewise.
6577 (SAVE_FP_PREFIX): Likewise.
6578 (SAVE_FP_SUFFIX): Likewise.
6579 (SIZE_TYPE): Likewise.
6580 (SUBSUBTARGET_OVERRIDE_OPTIONS): Likewise.
6581 (TARGET_64BIT): Likewise.
6582 (TARGET_64BIT): Likewise.
6583 (TARGET_AIX): Likewise.
6584 (WCHAR_TYPE_SIZE): Likewise.
6585 (WCHAR_TYPE): Undefine.
6586 (TARGET_OS_CPP_BUILTINS): Add 64-bit PowerPC defines.
6587 (CPP_OS_DEFAULT_SPEC): Use previous CPP_OS_RTEMS_SPEC.
6588 (CPP_OS_RTEMS_SPEC): Delete.
6589 (SUBSUBTARGET_EXTRA_SPECS): Remove cpp_os_rtems. Add
6590 asm_spec_common, asm_spec32, asm_spec64, link_os_spec32, and
6592 * config/rs6000/t-rtems: Add mcpu=e6500/m64 multilibs.
6594 2017-07-28 Jakub Jelinek <jakub@redhat.com>
6596 PR tree-optimization/81578
6597 * tree-parloops.c (build_new_reduction): Bail out if
6598 reduction_code isn't one of the standard OpenMP reductions.
6599 Move the details printing after that decision.
6601 2017-07-27 Peter Bergner <bergner@vnet.ibm.com>
6603 * config/rs6000/predicates.md (volatile_mem_operand): Remove code
6604 related to reload_in_progress.
6605 (splat_input_operand): Likewise.
6606 * config/rs6000/rs6000-protos.h (rs6000_secondary_memory_needed_rtx):
6608 * config/rs6000/rs6000.c (machine_function): Remove sdmode_stack_slot
6610 (TARGET_EXPAND_TO_RTL_HOOK): Delete.
6611 (TARGET_INSTANTIATE_DECLS): Likewise.
6612 (legitimate_indexed_address_p): Delete reload_in_progress code.
6613 (rs6000_debug_legitimate_address_p): Likewise.
6614 (rs6000_eliminate_indexed_memrefs): Likewise.
6615 (rs6000_emit_le_vsx_store): Likewise.
6616 (rs6000_emit_move_si_sf_subreg): Likewise.
6617 (rs6000_emit_move): Likewise.
6618 (register_to_reg_type): Likewise.
6619 (rs6000_pre_atomic_barrier): Likewise.
6620 (rs6000_machopic_legitimize_pic_address): Likewise.
6621 (rs6000_allocate_stack_temp): Likewise.
6622 (rs6000_address_for_fpconvert): Likewise.
6623 (rs6000_address_for_altivec): Likewise.
6624 (rs6000_secondary_memory_needed_rtx): Delete function.
6625 (rs6000_check_sdmode): Likewise.
6626 (rs6000_alloc_sdmode_stack_slot): Likewise.
6627 (rs6000_instantiate_decls): Likewise.
6628 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_RTX): Delete.
6629 * config/rs6000/rs6000.md (splitter for *movsi_got_internal):
6630 Delete reload_in_progress.
6631 (*vec_reload_and_plus_<mptrsize>): Likewise.
6632 * config/rs6000/vsx.md (vsx_mul_v2di): Likewise.
6633 (vsx_div_v2di): Likewise.
6634 (vsx_udiv_v2di): Likewise.
6636 2017-07-27 Peter Bergner <bergner@vnet.ibm.com>
6638 * config/rs6000/rs6000.opt (mlra): Replace with stub.
6639 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Delete OPTION_MASK_LRA.
6640 * config/rs6000/rs6000.c (TARGET_LRA_P): Delete.
6641 (rs6000_debug_reg_global): Delete print of LRA status.
6642 (rs6000_option_override_internal): Delete dead LRA related code.
6643 (rs6000_lra_p): Delete function.
6644 * doc/invoke.texi (RS/6000 and PowerPC Options): Delete -mlra.
6646 2017-07-27 Sebastian Huber <sebastian.huber@embedded-brains.de>
6648 * config.gcc (riscv*-*-elf*): Add (riscv*-*-rtems*).
6649 * config/riscv/rtems.h: New file.
6651 2017-07-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6652 Sudakshina Das <sudi.das@arm.com>
6654 * config/aarch64/aarch64.md
6655 (define_split for and<mode>3nr_compare): Move
6656 non aarch64_logical_operand to a register.
6657 (define_split for and_<SHIFT:optab><mode>3nr_compare0): Move non
6658 register immediate operand to a register.
6659 * config/aarch64/predicates.md (aarch64_mov_imm_operand): New.
6661 2017-07-27 Peter Bergner <bergner@vnet.ibm.com>
6664 * tree-cfg.c (group_case_labels_stmt): Handle already deleted blocks.
6666 2017-07-27 Richard Biener <rguenther@suse.de>
6668 PR tree-optimization/81573
6669 PR tree-optimization/81494
6670 * tree-vect-loop.c (vect_create_epilog_for_reduction): Handle
6671 multi defuse cycle case.
6673 2017-07-27 Richard Biener <rguenther@suse.de>
6675 PR tree-optimization/81571
6676 * tree-vect-slp.c (vect_build_slp_tree): Properly verify reduction
6679 2017-07-27 Eric Botcazou <ebotcazou@adacore.com>
6681 * config/sparc/sparc.c (sparc_option_override): Set MASK_FSMULD flag
6682 earlier and only if MASK_FPU is set. Adjust formatting.
6684 2017-07-27 Martin Liska <mliska@suse.cz>
6686 * opt-functions.awk: Add validation of value of Init.
6687 * optc-gen.awk: Pass new argument.
6689 2017-07-27 Martin Liska <mliska@suse.cz>
6691 * auto-profile.c (autofdo_source_profile::update_inlined_ind_target):
6692 Fix wrong condition.
6694 2017-07-27 Martin Liska <mliska@suse.cz>
6696 * auto-profile.c (afdo_annotate_cfg): Assign zero counts to
6697 BBs and edges seen by autoFDO.
6699 2017-07-27 Richard Biener <rguenther@suse.de>
6701 PR tree-optimization/81502
6702 * tree-ssa.c (non_rewritable_lvalue_p): Handle BIT_INSERT_EXPR
6703 with incompatible but same sized type.
6704 (execute_update_addresses_taken): Likewise.
6706 2017-07-27 James Greenhalgh <james.greenhalgh@arm.com>
6708 * tree-ssa-loop-ch.c (pass_ch::process_loop_p): Guard on
6709 flag_tree_loop_vectorize rather than flag_tree_vectorize.
6711 2017-07-27 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
6714 * config/s390/s390.md ("*atomic_compare_and_swap<mode>_1")
6715 ("*atomic_compare_and_swapdi_2", "*atomic_compare_and_swapsi_3"):
6716 Change s_operand to memory_operand.
6718 2017-07-27 Richard Sandiford <richard.sandiford@linaro.org>
6720 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_permute): Declare.
6721 * config/rs6000/rs6000.c (rs6000_gen_le_vsx_permute): Replace with...
6722 (rs6000_emit_le_vsx_permute): ...this. Take the destination as input.
6723 Emit instructions rather than returning an expression. Handle TFmode
6724 and KFmode by casting to TImode.
6725 (rs6000_emit_le_vsx_load): Update to use rs6000_emit_le_vsx_permute.
6726 (rs6000_emit_le_vsx_store): Likewise.
6727 * config/rs6000/vsx.md (VSX_TI): New iterator.
6728 (*vsx_le_permute_<mode>): Use it instead of VSX_LE_128.
6729 (*vsx_le_undo_permute_<mode>): Likewise.
6730 (*vsx_le_perm_load_<mode>): Use rs6000_emit_le_vsx_permute to
6731 emit the split sequence.
6732 (*vsx_le_perm_store_<mode>): Likewise.
6734 2017-07-27 Jakub Jelinek <jakub@redhat.com>
6736 PR tree-optimization/81555
6737 PR tree-optimization/81556
6738 * tree-ssa-reassoc.c (rewrite_expr_tree): Add NEXT_CHANGED argument,
6739 if true, force CHANGED for the recursive invocation.
6740 (reassociate_bb): Remember original length of ops array, pass
6741 len != orig_len as NEXT_CHANGED in rewrite_expr_tree call.
6743 * attribs.c (decl_attributes): Imply noinline, noclone and no_icf
6744 attributes for noipa attribute. For naked attribute use
6745 lookup_attribute first before lookup_attribute_spec.
6746 * final.c (rest_of_handle_final): Disable IPA RA for functions with
6748 * ipa-visibility.c (non_local_p): Fix comment typos. Return true
6749 for functions with noipa attribute.
6750 (cgraph_externally_visible_p): Return true for functions with noipa
6752 * cgraph.c (cgraph_node::get_availability): Return AVAIL_INTERPOSABLE
6753 for functions with noipa attribute.
6754 * doc/extend.texi: Document noipa function attribute.
6755 * tree-ssa-structalias.c (refered_from_nonlocal_fn): Set *nonlocal_p
6756 also for functions with noipa attribute.
6757 (ipa_pta_execute): Set nonlocal_p also for nodes with noipa attribute.
6759 2017-07-26 Andrew Pinski <apinski@cavium.com>
6761 * config/aarch64/aarch64.c (thunderx_vector_cost): Decrease cost of
6762 vec_unalign_load_cost and vec_unalign_store_cost.
6764 2017-07-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6766 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Delete
6767 -mvsx-small-integer option.
6768 (ISA_3_0_MASKS_IEEE): Likewise.
6769 (OTHER_VSX_VECTOR_MASKS): Likewise.
6770 (POWERPC_MASKS): Likewise.
6771 * config/rs6000/rs6000.opt (-mvsx-small-integer): Likewise.
6772 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Simplify
6773 code, only testing for DImode being allowed in non-VSX floating
6775 (rs6000_init_hard_regno_mode_ok): Change TARGET_VSX_SMALL_INTEGER
6776 to TARGET_P8_VECTOR test. Remove redundant VSX test inside of
6778 (rs6000_option_override_internal): Delete -mvsx-small-integer.
6779 (rs6000_expand_vector_set): Change TARGET_VSX_SMALL_INTEGER to
6780 TARGET_P8_VECTOR test.
6781 (rs6000_secondary_reload_simple_move): Likewise.
6782 (rs6000_preferred_reload_class): Delete TARGET_VSX_SMALL_INTEGER,
6783 since TARGET_P9_VECTOR was already tested.
6784 (rs6000_opt_masks): Remove -mvsx-small-integer.
6785 * config/rs6000/vsx.md (vsx_extract_<mode>): Delete
6786 TARGET_VSX_SMALL_INTEGER, since a test for TARGET_P9_VECTOR was
6788 (vsx_extract_<mode>_p9): Delete TARGET_VSX_SMALL_INTEGER, since a
6789 test for TARGET_VEXTRACTUB was used, and that uses
6791 (p9 extract splitter): Likewise.
6792 (vsx_extract_<mode>_di_p9): Likewise.
6793 (vsx_extract_<mode>_store_p9): Likewise.
6794 (vsx_extract_si): Delete TARGET_VSX_SMALL_INTEGER, since a test
6795 for TARGET_P9_VECTOR was used. Delete code that is now dead with
6796 the elimination of TARGET_VSX_SMALL_INTEGER.
6797 (vsx_extract_<mode>_p8): Likewise.
6798 (vsx_ext_<VSX_EXTRACT_I:VS_scalar>_fl_<FL_CONV:mode>): Likewise.
6799 (vsx_ext_<VSX_EXTRACT_I:VS_scalar>_ufl_<FL_CONV:mode>): Likewise.
6800 (vsx_set_<mode>_p9): Likewise.
6801 (vsx_set_v4sf_p9): Likewise.
6802 (vsx_set_v4sf_p9_zero): Likewise.
6803 (vsx_insert_extract_v4sf_p9): Likewise.
6804 (vsx_insert_extract_v4sf_p9_2): Likewise.
6805 * config/rs6000/rs6000.md (sign extend splitter): Change
6806 TARGET_VSX_SMALL_INTEGER to TARGET_P8_VECTOR test.
6807 (floatsi<mode>2_lfiwax_mem): Likewise.
6808 (floatunssi<mode>2_lfiwzx_mem): Likewise.
6809 (float<QHI:mode><FP_ISA3:mode>2): Delete TARGET_VSX_SMALL_INTEGER,
6810 since a test for TARGET_P9_VECTOR was used.
6811 (float<QHI:mode><FP_ISA3:mode>2_internal): Likewise.
6812 (floatuns<QHI:mode><FP_ISA3:mode>2): Likewise.
6813 (floatuns<QHI:mode><FP_ISA3:mode>2_internal): Likewise.
6814 (fix_trunc<mode>si2): Change TARGET_VSX_SMALL_INTEGER to
6815 TARGET_P8_VECTOR test.
6816 (fix_trunc<mode>si2_stfiwx): Likewise.
6817 (fix_trunc<mode>si2_internal): Likewise.
6818 (fix_trunc<SFDF:mode><QHI:mode>2): Delete
6819 TARGET_VSX_SMALL_INTEGER, since a test for TARGET_P9_VECTOR was
6821 (fix_trunc<SFDF:mode><QHI:mode>2_internal): Likewise.
6822 (fixuns_trunc<mode>si2): Change TARGET_VSX_SMALL_INTEGER to
6823 TARGET_P8_VECTOR test.
6824 (fixuns_trunc<mode>si2_stfiwx): Likewise.
6825 (fixuns_trunc<SFDF:mode><QHI:mode>2): Delete
6826 TARGET_VSX_SMALL_INTEGER, since a test for TARGET_P9_VECTOR was
6828 (fixuns_trunc<SFDF:mode><QHI:mode>2_internal): Likewise.
6829 (fctiw<u>z_<mode>_smallint): Delete TARGET_VSX_SMALL_INTEGER,
6830 since a test for TARGET_P9_VECTOR was used.
6831 (splitter for loading small constants): Likewise.
6833 2017-07-26 Andrew Pinski <apinski@cavium.com>
6835 * config/aarch64/aarch64.c (thunderx_vector_cost): Fix
6838 2017-07-26 H.J. Lu <hongjiu.lu@intel.com>
6841 * config/i386/i386.c (sp_valid_at): Properly check CFA offset.
6842 (fp_valid_at): Likewise.
6844 2017-07-26 James Greenhalgh <james.greenhalgh@arm.com>
6846 * config/aarch64/aarch64.c (cortexa57_addrcost_table): Remove.
6847 (qdf24xx_addrcost_table): Likewise.
6848 (cortexa57_tunings): Update to use generic_branch_cost.
6849 (cortexa72_tunings): Likewise.
6850 (cortexa73_tunings): Likewise.
6851 (qdf24xx_tunings): Likewise.
6853 2017-07-26 James Greenhalgh <james.greenhalgh@arm.com>
6855 * config/aarch64/aarch64.c (cortexa57_branch_cost): Remove.
6856 (thunderx2t99_branch_cost): Likewise.
6857 (cortexa35_tunings): Update to use generic_branch_cost.
6858 (cortexa53_tunings): Likewise.
6859 (cortexa57_tunings): Likewise.
6860 (cortexa72_tunings): Likewise.
6861 (cortexa73_tunings): Likewise.
6862 (thunderx2t99_tunings): Likewise.
6864 2017-07-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
6866 * config/sparc/sparc.c (dump_target_flag_bits): Dump MASK_FSMULD.
6867 (sparc_option_override): Honour MASK_FSMULD.
6868 * config/sparc/sparc.h (MASK_FEATURES): Add MASK_FSMULD.
6869 * config/sparc/sparc.md (muldf3_extend): Use TARGET_FSMULD.
6870 * config/sparc/sparc.opt (mfsmuld): New option.
6871 * doc/invoke.texi (mfsmuld): Document option.
6873 2017-07-26 Marek Polacek <polacek@redhat.com>
6876 * tree.c (build2_stat): Don't set TREE_CONSTANT on divisions by zero.
6878 2017-07-26 Richard Biener <rguenther@suse.de>
6880 * gimple-match-head.c (do_valueize): Return OP if valueize
6882 (get_def): New helper to get at the def stmt of a SSA name
6884 * genmatch.c (dt_node::gen_kids_1): Use get_def instead of
6885 do_valueize to get at the def stmt.
6886 (dt_operand::gen_gimple_expr): Simplify do_valueize calls.
6888 2017-07-26 Wilco Dijkstra <wdijkstr@arm.com>
6891 * auto-inc-dec.c (parse_add_or_inc): Block autoinc on sfp.
6893 2017-07-26 Martin Liska <mliska@suse.cz>
6896 * function.c (expand_function_start): Make expansion of
6897 nonlocal_goto_save_area after parm_birth_insn.
6899 2017-07-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
6901 * config/sparc/sparc.c (sparc_option_override): Remove MASK_FPU
6902 from all CPU target flags enable members.
6904 2017-07-26 Richard Biener <rguenther@suse.de>
6906 * genmatch.c (dt_simplify::gen): Make iterator vars const.
6907 (decision_tree::gen): Make 'type' const.
6908 (write_predicate): Likewise.
6910 2017-07-24 Michael Meissner <meissner@linux.vnet.ibm.com>
6912 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok):
6913 Eliminate TARGET_UPPER_REGS_{DF,DI,SF} usage.
6914 (rs6000_option_override_internal): Likewise.
6915 (rs6000_expand_vector_set): Likewise.
6916 * config/rs6000/rs6000.h (TARGET_UPPER_REGS_DF): Delete.
6917 (TARGET_UPPER_REGS_SF): Likewise.
6918 (TARGET_UPPER_REGS_DI): Likewise.
6919 (TARGET_VEXTRACTUB): Eliminate TARGET_UPPER_REGS_{DF,DI,SF}.
6920 (TARGET_DIRECT_MOVE_64BIT): Likewise.
6921 * config/rs6000/rs6000.md (ALTIVEC_DFORM): Likewise.
6922 (float<QHI:mode><FP_ISA3:mode>2_internal): Likewise.
6923 (Splitters for DI constants in Altivec registers): Likewise.
6924 * config/rs6000/vsx.md (vsx_set_<mode>_p9): Likewise.
6925 (vsx_set_v4sf_p9): Likewise.
6926 (vsx_set_v4sf_p9_zero): Likewise.
6927 (vsx_insert_extract_v4sf_p9): Likewise.
6928 (vsx_insert_extract_v4sf_p9_2): Likewise.
6930 2017-07-25 Carl Love <cel@us.ibm.com>
6932 * doc/extend.texi: Update the built-in documentation file for the
6933 existing built-in functions
6934 vector signed char vec_cnttz (vector signed char);
6935 vector unsigned char vec_cnttz (vector unsigned char);
6936 vector signed short vec_cnttz (vector signed short);
6937 vector unsigned short vec_cnttz (vector unsigned short);
6938 vector signed int vec_cnttz (vector signed int);
6939 vector unsigned int vec_cnttz (vector unsigned int);
6940 vector signed long long vec_cnttz (vector signed long long);
6941 vector unsigned long long vec_cnttz (vector unsigned long long);
6943 2017-07-25 Andrew Pinski <apinski@cavium.com>
6945 * tree-ssa-uninit.c (warn_uninitialized_vars): Don't warn about memory
6946 accesses where the use is for the first operand of a BIT_INSERT.
6948 2017-07-25 Jim Wilson <jim.wilson@linaro.org>
6951 * config/i386/winnt-cxx.c (i386_pe_adjust_class_at_definition): Look
6952 for FUNCTION_DECLs in TYPE_FIELDS rather than TYPE_METHODS.
6954 2017-07-25 Jim Wilson <jim.wilson@linaro.org>
6956 * config/i386/gstabs.h: Delete.
6957 * config/i386/openbsd.h, config/i386/t-openbsd: Likewise.
6959 2017-07-25 Uros Bizjak <ubizjak@gmail.com>
6961 * config/i386/i386.c (ix86_decompose_address): Do not check for
6962 register RTX when looking at index_reg or base_reg.
6963 * config/i386/i386.h (INCOMING_RETURN_ADDR_RTX): Use stack_pointer_rtx.
6965 2017-07-25 Eric Botcazou <ebotcazou@adacore.com>
6967 * gimple.c (gimple_assign_set_rhs_with_ops): Do not ask gsi_replace
6968 to update EH info here.
6970 2017-07-25 Alexander Monakov <amonakov@ispras.ru>
6972 * match.pd ((X * CST1) * CST2): Simplify to X * (CST1 * CST2).
6974 2017-07-25 Alexander Monakov <amonakov@ispras.ru>
6976 * match.pd ((X * CST) * Y): Reassociate to (X * Y) * CST.
6978 2017-07-25 Torsten Duwe <duwe@suse.de>
6980 * common.opt: Introduce -fpatchable-function-entry
6981 command line option, and its variables function_entry_patch_area_size
6982 and function_entry_patch_area_start.
6983 * opts.c (common_handle_option): Add -fpatchable_function_entry_ case,
6984 including a two-value parser.
6985 * target.def (print_patchable_function_entry): New target hook.
6986 * targhooks.h (default_print_patchable_function_entry): New function.
6987 * targhooks.c (default_print_patchable_function_entry): Likewise.
6988 * toplev.c (process_options): Switch off IPA-RA if
6989 patchable function entries are being generated.
6990 * varasm.c (assemble_start_function): Look at the
6991 patchable-function-entry command line switch and current
6992 function attributes and maybe generate NOP instructions by
6993 calling the print_patchable_function_entry hook.
6994 * doc/extend.texi: Document patchable_function_entry attribute.
6995 * doc/invoke.texi: Document -fpatchable_function_entry
6996 command line option.
6997 * doc/tm.texi.in (TARGET_ASM_PRINT_PATCHABLE_FUNCTION_ENTRY):
6999 * doc/tm.texi: Re-generate.
7001 2017-07-25 Jakub Jelinek <jakub@redhat.com>
7004 * config/i386/constraints.md (Yd, Ye): Use ALL_SSE_REGS for
7005 TARGET_AVX512DQ rather than TARGET_AVX512BW.
7007 2017-07-25 Tamar Christina <tamar.christina@arm.com>
7009 * config/arm/parsecpu.awk (all_cores): Remove duplicates.
7011 2017-07-25 Richard Biener <rguenther@suse.de>
7013 PR tree-optimization/81455
7014 * tree-ssa-loop-unswitch.c (find_loop_guard): Make sure to
7015 not walk in cycles when looking for guards.
7017 2017-07-25 Richard Biener <rguenther@suse.de>
7019 PR tree-optimization/81529
7020 * tree-vect-stmts.c (process_use): Disregard live induction PHIs
7021 when optimizing backedge uses.
7023 2017-07-25 David Edelsohn <dje.gcc@gmail.com>
7025 * dwarf2asm.c (dw2_asm_output_nstring): Encode double quote
7027 * dwarf2out.c (output_macinfo): Copy debug_line_section_label
7028 to dl_section_ref. On AIX, append an expression to subtract
7029 the size of the section length to dl_section_ref.
7031 2017-07-25 Segher Boessenkool <segher@kernel.crashing.org>
7033 * configure.ac: If any of the config.* scripts fail, exit 1.
7034 * configure: Regenerate.
7036 2017-07-25 Richard Biener <rguenther@suse.de>
7039 * tree-ssa-operands.c (verify_imm_links): Remove cap on number
7040 of immediate uses, be more verbose on errors.
7042 2017-07-25 Richard Biener <rguenther@suse.de>
7044 PR tree-optimization/81510
7045 * tree-vect-loop.c (vect_is_simple_reduction): When the
7046 reduction stmt is not inside the loop bail out.
7048 2017-07-25 Richard Biener <rguenther@suse.de>
7050 PR tree-optimization/81303
7051 * tree-vect-loop-manip.c (vect_loop_versioning): Build
7052 profitability check against LOOP_VINFO_NITERSM1.
7054 2017-07-25 Alexander Monakov <amonakov@ispras.ru>
7056 * domwalk.c (cmp_bb_postorder): Simplify.
7057 (sort_bbs_postorder): New function. Use it...
7058 (dom_walker::walk): ...here to optimize common cases.
7060 2017-07-25 Martin Liska <mliska@suse.cz>
7063 * ipa-visibility.c (function_and_variable_visibility): Make the
7064 redirection just on target that supports aliasing.
7065 Fix GNU coding style.
7067 2017-07-25 Sebastian Huber <sebastian.huber@embedded-brains.de>
7070 * config/aarch64/rtems.h: Add GCC Runtime Library Exception.
7072 * config/arm/rtems.h: Likewise.
7073 * config/bfin/rtems.h: Likewise.
7074 * config/i386/rtemself.h: Likewise.
7075 * config/lm32/rtems.h: Likewise.
7076 * config/m32c/rtems.h: Likewise.
7077 * config/m68k/rtemself.h: Likewise.
7078 * config/microblaze/rtems.h: Likewise.
7079 * config/mips/rtems.h: Likewise.
7080 * config/moxie/rtems.h: Likewise.
7081 * config/nios2/rtems.h: Likewise.
7082 * config/powerpcspe/rtems.h: Likewise.
7083 * config/rs6000/rtems.h: Likewise.
7084 * config/rtems.h: Likewise.
7085 * config/sh/rtems.h: Likewise.
7086 * config/sh/rtemself.h: Likewise.
7087 * config/sparc/rtemself.h: Likewise.
7089 2017-07-25 Georg-Johann Lay <avr@gjlay.de>
7092 * hsa-brig.c (brig_init): Use xasprintf instead of asprintf.
7093 * gimple-pretty-print.c (dump_profile, dump_probability): Same.
7094 * tree-ssa-structalias.c (alias_get_name): Same.
7096 2017-07-25 Bin Cheng <bin.cheng@arm.com>
7099 * config/aarch64/cortex-a57-fma-steering.c (analyze): Skip fmul/fmac
7100 instructions if no du chain is found.
7102 2017-07-25 Georg-Johann Lay <avr@gjlay.de>
7104 * config/avr/avr-log.c (avr_log_vadump) ['T']: Print NULL-TREE.
7106 2017-07-25 Richard Biener <rguenther@suse.de>
7109 * fold-const.c (fold_negate_const): TREE_OVERFLOW should be
7112 2017-07-24 Michael Meissner <meissner@linux.vnet.ibm.com>
7114 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Delete
7116 (ISA_2_7_MASKS_SERVER): Likewise.
7117 (ISA_3_0_MASKS_IEEE): Likewise.
7118 (OTHER_P8_VECTOR_MASKS): Likewise.
7119 (OTHER_VSX_VECTOR_MASKS): Likewise.
7120 (POWERPC_MASKS): Likewise.
7121 (power7 cpu): Use ISA_2_6_MASKS_SERVER instead of using a
7122 duplicate list of options.
7123 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Remove
7124 explicit -mupper-regs options.
7125 * config/rs6000/rs6000.opt (-mvsx-scalar-memory): Delete
7126 -mupper-regs* options. Delete -mvsx-scalar-memory, which was an
7127 alias for -mupper-regs-df.
7128 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Likewise.
7129 (rs6000_init_hard_regno_mode_ok): Likewise.
7130 (rs6000_option_override_internal): Likewise.
7131 (rs6000_opt_masks): Likewise.
7132 * config/rs6000/rs6000.h (TARGET_UPPER_REGS_DF): Define upper regs
7133 options in terms of whether -mvsx or -mpower8-vector was used.
7134 (TARGET_UPPER_REGS_DI): Likewise.
7135 (TARGET_UPPER_REGS_SF): Likewise.
7136 * doc/invoke.texi (RS/6000 and PowerPC Options): Delete the
7137 -mupper-regs-* options.
7139 2017-07-24 Segher Boessenkool <segher@kernel.crashing.org>
7141 * passes.c (emergency_dump_function): Print some empty lines and a
7142 header before the RTL dump.
7144 2017-07-24 Segher Boessenkool <segher@kernel.crashing.org>
7146 * cfgrtl.c (rtl_dump_bb): Don't call NEXT_INSN on NULL.
7148 2017-07-24 Wilco Dijkstra <wdijkstr@arm.com>
7151 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7152 Avoid SYMBOL_SMALL_ABSOLUTE for literals with pc-relative literals.
7154 2017-07-24 Carl Love <cel@us.ibm.com>
7156 * config/rs6000/rs6000-c.c: Add support for built-in functions
7157 vector float vec_extract_fp32_from_shorth (vector unsigned short);
7158 vector float vec_extract_fp32_from_shortl (vector unsigned short);
7159 * config/rs6000/altivec.h (vec_extract_fp_from_shorth,
7160 vec_extract_fp_from_shortl): Add defines for the two builtins.
7161 * config/rs6000/rs6000-builtin.def (VEXTRACT_FP_FROM_SHORTH,
7162 VEXTRACT_FP_FROM_SHORTL): Add BU_P9V_OVERLOAD_1 and BU_P9V_VSX_1
7164 * config/rs6000/vsx.md vsx_xvcvhpsp): Add define_insn.
7165 (vextract_fp_from_shorth, vextract_fp_from_shortl): Add define_expands.
7166 * doc/extend.texi: Update the built-in documentation file for the
7167 new built-in function.
7169 2017-07-24 Jakub Jelinek <jakub@redhat.com>
7172 * tree.def: Remove TYPE_METHODS documentation, adjust TYPE_FIELDS
7174 * doc/generic.texi: Likewise.
7175 * config/i386/winnt-cxx.c (i386_pe_adjust_class_at_definition): Look
7176 for FUNCTION_DECLs in TYPE_FIELDS rather than TYPE_METHODS.
7178 2017-07-24 Jackson Woodruff <jackson.woodruff@arm.com>
7180 * config/aarch64/aarch64-simd.md (aarch64_mla_elt_merge<mode>): New.
7181 (aarch64_mls_elt_merge<mode>): Likewise.
7183 2017-07-23 Krister Walfridsson <krister.walfridsson@gmail.com>
7185 * config.gcc (*-*-netbsd*): Remove check for NetBSD versions not
7186 having __cxa_atexit.
7188 2017-07-23 Michael Collison <michael.collison@arm.com>
7190 * config/arm/arm.c (arm_option_override): Deprecate
7191 use of -mstructure-size-boundary.
7192 * config/arm/arm.opt: Deprecate -mstructure-size-boundary.
7193 * doc/invoke.texi: Deprecate -mstructure-size-boundary.
7195 2017-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7198 * config/rs6000/rs6000.c (rs6000_builtin_vectorization_cost):
7199 Reduce cost estimate for direct moves.
7201 2017-07-23 Uros Bizjak <ubizjak@gmail.com>
7204 * config/i386/i386.c (ix86_option_override_internal): Disable
7205 BMI, BMI2 and TBM instructions for -m16.
7207 2017-07-21 Carl Love <cel@us.ibm.com>
7209 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7210 ALTIVEC_BUILTIN_VMULESW, ALTIVEC_BUILTIN_VMULEUW,
7211 ALTIVEC_BUILTIN_VMULOSW, ALTIVEC_BUILTIN_VMULOUW entries.
7212 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin,
7213 builtin_function_type): Add ALTIVEC_BUILTIN_* case statements.
7214 * config/rs6000/altivec.md (MVULEUW, VMULESW, VMULOUW,
7215 VMULOSW): New enum "unspec" values.
7216 (altivec_vmuleuw, altivec_vmulesw, altivec_vmulouw,
7217 altivec_vmulosw): New patterns.
7218 * config/rs6000/rs6000-builtin.def (VMLEUW, VMULESW, VMULOUW,
7219 VMULOSW): Add definitions.
7221 2017-07-21 Jim Wilson <jim.wilson@linaro.org>
7223 * config/aarch64/aarch64-cores.def (falkor): Add AARCH64_FL_RDMA.
7224 (qdf24xx): Likewise.
7225 * config/aarch64/aarch64-options-extensions.def (rdma); New.
7226 * config/aarch64/aarch64.h (AARCH64_FL_RDMA): New.
7227 (AARCH64_FL_V8_1): Renumber.
7228 (AARCH64_FL_FOR_ARCH8_1): Add AARCH64_FL_RDMA.
7229 (AARCH64_ISA_RDMA): Use AARCH64_FL_RDMA.
7230 * config/aarch64/arm_neon.h: Use +rdma instead of arch=armv8.1-a.
7231 * doc/invoke.texi (AArch64 Options): Mention +rmda in -march docs. Add
7232 rdma to feature modifiers list.
7234 2017-07-21 Yury Gribov <tetra2005@gmail.com>
7237 * ipa-visibility (function_and_variable_visibility): Convert
7238 recursive PLT call to direct call if appropriate.
7240 2017-07-21 Andrew Pinski <apinski@cavium.com>
7242 * tree-ssa-sccvn.c (vn_nary_op_eq): Check BIT_INSERT_EXPR's
7243 operand 1 to see if the types precision matches.
7244 * fold-const.c (operand_equal_p): Likewise.
7246 2017-07-21 Richard Biener <rguenther@suse.de>
7248 PR tree-optimization/81303
7249 * tree-vect-data-refs.c (vect_get_peeling_costs_all_drs): Pass
7250 in datarefs vector. Allow NULL dr0 for no peeling cost estimate.
7251 (vect_peeling_hash_get_lowest_cost): Adjust.
7252 (vect_enhance_data_refs_alignment): Likewise. Use
7253 vect_get_peeling_costs_all_drs to compute the penalty for no
7254 peeling to match up costs.
7256 2017-07-21 Richard Biener <rguenther@suse.de>
7258 PR tree-optimization/81500
7259 * tree-vect-loop.c (vect_is_simple_reduction): Properly fail if
7260 we didn't identify a reduction path.
7262 2017-07-21 Tom de Vries <tom@codesourcery.com>
7263 Cesar Philippidis <cesar@codesourcery.com>
7265 PR gcov-profile/81442
7266 * config/nvptx/nvptx.c (nvptx_goacc_reduction_init): Add missing edge
7269 2017-07-21 Tom de Vries <tom@codesourcery.com>
7272 * config/nvptx/nvptx.c (nvptx_override_options_after_change): New
7274 (TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE): Define to
7275 nvptx_override_options_after_change.
7277 2017-07-21 Ulrich Drepper <drepper@redhat.com>
7279 * dwarf2out.c (output_file_names): Avoid double testing for
7282 2017-07-21 Georg-Johann Lay <avr@gjlay.de>
7284 * doc/invoke.texi (AVR Built-in Functions): Re-layout section.
7286 2016-07-21 Jan Hubicka <hubicka@ucw.cz>
7288 * cfgcleanup.c (flow_find_cross_jump): Do not crossjump across
7290 (try_crossjump_to_edge): Do not punt on partitioned functions.
7292 2016-07-21 Jan Hubicka <hubicka@ucw.cz>
7294 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
7295 Put all BBs reachable only via paths crossing cold region to cold
7297 * cfgrtl.c (find_bbs_reachable_by_hot_paths): New function.
7299 2016-07-21 Richard Biener <rguenther@suse.de>
7301 PR tree-optimization/81303
7302 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Take
7303 into account prologue and epilogue iterations when raising
7304 min_profitable_iters to sth at least covering one vector iteration.
7306 2017-07-21 Tamar Christina <tamar.christina@arm.com>
7308 * config/arm/arm.c (arm_test_cpu_arch_dat):
7311 2017-07-20 Nathan Sidwell <nathan@acm.org>
7313 Remove TYPE_METHODS.
7314 * tree.h (TYPE_METHODS): Delete.
7315 * dwarf2out.c (gen_member_die): Member fns are on TYPE_FIELDS.
7316 * dbxout.c (dbxout_type_fields): Ignore FUNCTION_DECLs.
7317 (dbxout_type_methods): Scan TYPE_FIELDS.
7318 (dbxout_type): Don't check TYPE_METHODS here.
7319 * function.c (use_register_for_decl): Always ignore register for
7320 class types when not optimizing.
7321 * ipa-devirt.c (odr_types_equivalent_p): Delete TYPE_METHODS scan.
7322 * tree.c (free_lang_data_in_type): Stitch out member functions and
7323 templates from TYPE_FIELDS.
7324 (build_distinct_type_copy, verify_type_variant,
7325 verify_type): Member fns are on TYPE_FIELDS.
7326 * tree-dump.c (dequeue_and_dump): No TYPE_METHODS.
7327 * tree-pretty-print.c (dump_generic_node): Likewise.
7329 2017-07-20 Jakub Jelinek <jakub@redhat.com>
7332 * config/i386/i386.c (ix86_expand_vector_init_general): Handle
7333 V2TImode and V4TImode.
7334 (ix86_expand_vector_extract): Likewise.
7335 * config/i386/sse.md (VMOVE): Enable V4TImode even for just
7336 TARGET_AVX512F, instead of only for TARGET_AVX512BW.
7337 (ssescalarmode): Handle V4TImode and V2TImode.
7338 (VEC_EXTRACT_MODE): Add V4TImode and V2TImode.
7339 (*vec_extractv2ti, *vec_extractv4ti): New insns.
7340 (VEXTRACTI128_MODE): New mode iterator.
7341 (splitter for *vec_extractv?ti first element): New.
7342 (VEC_INIT_MODE): New mode iterator.
7343 (vec_init<mode>): Consolidate 3 expanders into one using
7344 VEC_INIT_MODE mode iterator.
7346 2017-07-20 Alexander Monakov <amonakov@ispras.ru>
7348 * lra-assigns.c (pseudo_compare_func): Fix comparison step based on
7349 non_spilled_static_chain_regno_p.
7351 2017-07-20 Alexander Monakov <amonakov@ispras.ru>
7353 * gimple-ssa-store-merging.c (sort_by_bitpos): Return 0 on equal bitpos.
7355 2017-07-20 Jan Hubicka <hubicka@ucw.cz>
7357 * bb-reorder.c (connect_traces): Allow copying of blocks within
7360 2017-07-20 Richard Biener <rguenther@suse.de>
7362 * gimple.h (gimple_phi_result): Add gphi * overload.
7363 (gimple_phi_result_ptr): Likewise.
7364 (gimple_phi_arg): Likewise. Adjust index assert to only
7365 allow actual argument accesses rather than all slots available
7367 (gimple_phi_arg_def): Add gphi * overload.
7368 * tree-phinodes.c (make_phi_node): Initialize only actual
7370 (resize_phi_node): Clear memory not covered by old node,
7371 do not initialize excess argument slots.
7372 (reserve_phi_args_for_new_edge): Initialize new argument slot
7375 2017-07-20 Bin Cheng <bin.cheng@arm.com>
7377 PR tree-optimization/81388
7379 2016-07-21 Bin Cheng <bin.cheng@arm.com>
7381 * tree-ssa-loop-niter.c (number_of_iterations_lt_to_ne): Clean up
7382 by removing computation of may_be_zero.
7384 2017-07-18 Jan Hubicka <hubicka@ucw.cz>
7385 Tom de Vries <tom@codesourcery.com>
7388 * cfgbuild.c (find_many_sub_basic_blocks): Update REG_BR_PROB note
7389 when gimple level profile disagrees with what RTL expander did.
7391 2017-07-20 Richard Biener <rguenther@suse.de>
7393 PR tree-optimization/61171
7394 * tree-vectorizer.h (slp_instance): Add reduc_phis member.
7395 (vect_analyze_stmt): Add slp instance parameter.
7396 (vectorizable_reduction): Likewise.
7397 * tree-vect-loop.c (vect_analyze_loop_operations): Adjust.
7398 (vect_is_simple_reduction): Deal with chains not detected
7399 as SLP reduction chain, specifically not properly associated
7400 chains containing a mix of plus/minus.
7401 (get_reduction_op): Remove.
7402 (get_initial_defs_for_reduction): Simplify, pass in whether
7403 this is a reduction chain, pass in the SLP node for the PHIs.
7404 (vect_create_epilog_for_reduction): Get the SLP instance as
7406 (vectorizable_reduction): Get the SLP instance as arg.
7407 During analysis remember the SLP node with the PHIs in the
7408 instance. Simplify getting at the vectorized reduction PHIs.
7409 * tree-vect-slp.c (vect_slp_analyze_node_operations): Pass
7410 through SLP instance.
7411 (vect_slp_analyze_operations): Likewise.
7412 * tree-vect-stms.c (vect_analyze_stmt): Likewise.
7413 (vect_transform_stmt): Likewise.
7415 2017-07-20 Tom de Vries <tom@codesourcery.com>
7417 PR tree-optimization/81489
7418 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behavior): Move
7419 read of phi arg location to before loop that modifies phi.
7421 2017-07-20 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
7423 * match.pd (((m1 >/</>=/<= m2) * d -> (m1 >/</>=/<= m2) ? d : 0):
7426 2017-07-19 Jan Hubicka <hubicka@ucw.cz>
7429 * except.c (execute): Fix ordering issue.
7431 2018-07-19 Segher Boessenkool <segher@kernel.crashing.org>
7433 PR rtl-optimization/81423
7434 * combine.c (make_compound_operation_int): Don't try to optimize
7435 the AND of a SUBREG of an LSHIFTRT if that SUBREG is paradoxical.
7437 2017-07-19 Segher Boessenkool <segher@kernel.crashing.org>
7439 PR rtl-optimization/81423
7440 * simplify-rtx.c (simplify_truncation): Handle truncating an IOR
7441 with a constant that is -1 in the truncated to mode.
7443 2017-07-19 Jan Hubicka <hubicka@ucw.cz>
7445 * predict.c (propagate_unlikely_bbs_forward): Break out from ...
7446 (determine_unlikely_bbs): ... here.
7447 * predict.h (propagate_unlikely_bbs_forward): Declare.
7448 * cfgexpand.c (pass_expand::execute): Use it.
7449 * bb-reorder.c (sanitize_hot_paths): Do not consider known to be
7451 (find_rarely_executed_basic_blocks_and_crossing_edges): Use
7452 propagate_unlikely_bbs_forward.
7454 2017-07-19 Jan Hubicka <hubicka@ucw.cz>
7457 * except.c (maybe_add_nop_after_section_switch): New function.
7460 2017-07-19 Tom de Vries <tom@codesourcery.com>
7462 * gimple.h (gimple_phi_set_arg): Make assert more strict.
7464 2017-07-19 Tom de Vries <tom@codesourcery.com>
7466 * gimple.h (gimple_phi_arg): Make assert more strict.
7468 2017-07-19 Steven Munroe <munroesj@gcc.gnu.org>
7470 * config.gcc (powerpc*-*-*): Add mmintrin.h.
7471 * config/rs6000/mmintrin.h: New file.
7472 * config/rs6000/x86intrin.h [__ALTIVEC__]: Include mmintrin.h.
7474 2017-07-19 Jakub Jelinek <jakub@redhat.com>
7476 PR tree-optimization/81346
7477 * match.pd: Optimize (X - 1U) <= INT_MAX-1U into (int) X > 0.
7479 2017-07-19 Tom de Vries <tom@codesourcery.com>
7481 * config/nvptx/nvptx.md (VECIM): Add V2DI.
7483 2017-07-19 Tom de Vries <tom@codesourcery.com>
7485 * config/nvptx/nvptx-modes.def: Add V2DImode.
7486 * config/nvptx/nvptx-protos.h (nvptx_data_alignment): Declare.
7487 * config/nvptx/nvptx.c (nvptx_ptx_type_from_mode): Handle V2DImode.
7488 (nvptx_output_mov_insn): Handle lack of mov.b128.
7489 (nvptx_print_operand): Handle 'H' and 'L' codes.
7490 (nvptx_vector_mode_supported): Allow V2DImode.
7491 (nvptx_preferred_simd_mode): New function.
7492 (nvptx_data_alignment): New function.
7493 (TARGET_VECTORIZE_PREFERRED_SIMD_MODE): Redefine to
7494 nvptx_preferred_simd_mode.
7495 * config/nvptx/nvptx.h (STACK_BOUNDARY, BIGGEST_ALIGNMENT): Change from
7497 (DATA_ALIGNMENT): Define. Set to nvptx_data_alignment.
7499 2017-07-19 Tom de Vries <tom@codesourcery.com>
7501 * config/nvptx/nvptx-modes.def: New file. Add V2SImode.
7502 * config/nvptx/nvptx.c (nvptx_ptx_type_from_mode): Handle V2SImode.
7503 (nvptx_vector_mode_supported): New function. Allow V2SImode.
7504 (TARGET_VECTOR_MODE_SUPPORTED_P): Redefine to nvptx_vector_mode_supported.
7505 * config/nvptx/nvptx.md (VECIM): New mode iterator. Add V2SI.
7506 (mov<VECIM>_insn): New define_insn.
7507 (define_expand "mov<VECIM>): New define_expand.
7509 2017-07-19 Tom de Vries <tom@codesourcery.com>
7511 * config/nvptx/nvptx.c (nvptx_print_operand): Handle v2 vector mode.
7513 2017-07-19 Jakub Jelinek <jakub@redhat.com>
7515 PR tree-optimization/81346
7516 * fold-const.h (fold_div_compare, range_check_type): Declare.
7517 * fold-const.c (range_check_type): New function.
7518 (build_range_check): Use range_check_type.
7519 (fold_div_compare): No longer static, rewritten into
7520 a match.pd helper function.
7521 (fold_comparison): Don't call fold_div_compare here.
7522 * match.pd (X / C1 op C2): New optimization using fold_div_compare
7525 2017-07-19 Nathan Sidwell <nathan@acm.org>
7527 * tree.h (TYPE_MINVAL, TYPE_MAXVAL): Rename to ...
7528 (TYPE_MIN_VALUE_RAW, TYPE_MAX_VALUE_RAW): ... these.
7529 * tree.c (find_decls_types_r, verify_type): Use
7530 TYPE_{MIN,MAX}_VALUE_RAW.
7531 * lto-streamer-out.c (DFS::DFS_write_tree_body): Likewise.
7532 (hash_tree): Likewise.
7533 * tree-streamer-in.c (lto_input_ts_type_non_common_tree_pointers):
7535 * tree-streamer-out.c (write_ts_type_non_common_tree_pointers):
7538 2017-07-18 Tom de Vries <tom@codesourcery.com>
7541 * omp-expand.c (expand_omp_for_static_chunk): Handle
7542 equal-argument loop exit phi.
7544 2017-07-18 Uros Bizjak <ubizjak@gmail.com>
7547 * config/i386/i386.md (rorx_immediate_operand): New mode attribute.
7548 (*bmi2_rorx<mode>3_1): Use rorx_immediate_operand as
7549 operand 2 predicate.
7550 (*bmi2_rorxsi3_1_zext): Use const_0_to_31_operand as
7551 operand 2 predicate.
7552 (ror,rol -> rorx splitters): Use const_int_operand as
7553 operand 2 predicate.
7555 2017-06-18 Richard Biener <rguenther@suse.de>
7557 PR tree-optimization/81410
7558 * tree-vect-stmts.c (vectorizable_load): Properly adjust for
7559 the gap in the ! slp_perm SLP case after each group.
7561 2017-07-18 Jan Hubicka <hubicka@ucw.cz>
7564 * cfgloopmanip.c (scale_loop_profile): Watch out for zero frequency
7567 2017-07-18 Jan Hubicka <hubicka@ucw.cz>
7570 * predict.c (set_even_probabilities): Cleanup; do not affect
7571 probabilities that are already known.
7572 (combine_predictions_for_bb): Call even when count is set.
7574 2017-07-18 Nathan Sidwell <nathan@acm.org>
7576 * tree-parloops.c (try_transform_to_exit_first_loop_alt): Use
7579 2017-07-18 Bin Cheng <bin.cheng@arm.com>
7582 * tree-ssa-loop-niter.c (number_of_iterations_exit): Dump missed
7583 optimization for loop niter analysis.
7585 2017-07-18 Georg-Johann Lay <avr@gjlay.de>
7588 * config/avr/avr.c (avr_optimize_casesi): Don't use
7589 INT8_MIN, INT8_MAX, UINT8_MAX, INT16_MIN, INT16_MAX, UINT16_MAX.
7591 2017-07-18 Robin Dapp <rdapp@linux.vnet.ibm.com>
7593 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Remove
7594 body_cost_vec from _vect_peel_extended_info.
7595 (vect_peeling_hash_get_lowest_cost): Do not set body_cost_vec.
7596 (vect_peeling_hash_choose_best_peeling): Remove body_cost_vec and
7599 2017-07-18 Bin Cheng <bin.cheng@arm.com>
7601 * config/arm/arm.c (emit_unlikely_jump): Remove unused var.
7603 2017-07-18 Richard Biener <rguenther@suse.de>
7605 PR tree-optimization/80620
7606 PR tree-optimization/81403
7607 * tree-ssa-pre.c (phi_translate_1): Clear range and points-to
7608 info when re-using a VN table entry.
7610 2017-07-18 Richard Biener <rguenther@suse.de>
7612 PR tree-optimization/81418
7613 * tree-vect-loop.c (vectorizable_reduction): Properly compute
7614 vectype_in. Verify that with lane-reducing reduction operations
7615 we have a single def-use cycle.
7617 2017-07-17 Carl Love <cel@us.ibm.com>
7619 Revert commit r249424 2017-06-20 Carl Love <cel@us.ibm.com>
7621 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7622 ALTIVEC_BUILTIN_VMULESW, ALTIVEC_BUILTIN_VMULEUW,
7623 ALTIVEC_BUILTIN_VMULOSW, ALTIVEC_BUILTIN_VMULOUW entries.
7624 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin,
7625 builtin_function_type): Add ALTIVEC_BUILTIN_* case statements.
7626 * config/rs6000/altivec.md (MVULEUW, VMULESW, VMULOUW,
7627 VMULOSW): New enum "unspec" values.
7628 (vec_widen_umult_even_v4si, vec_widen_smult_even_v4si,
7629 vec_widen_umult_odd_v4si, vec_widen_smult_odd_v4si,
7630 altivec_vmuleuw, altivec_vmulesw, altivec_vmulouw,
7631 altivec_vmulosw): New patterns.
7632 * config/rs6000/rs6000-builtin.def (VMLEUW, VMULESW, VMULOUW,
7633 VMULOSW): Add definitions.
7634 2017-07-17 Uros Bizjak <ubizjak@gmail.com>
7636 * config/alpha/alpha.c: Include predict.h.
7638 2017-07-17 Yury Gribov <tetra2005@gmail.com>
7640 * tree-vrp.c (compare_assert_loc): Fix comparison function
7641 to return predictable results.
7643 2017-07-17 Claudiu Zissulescu <claziss@synopsys.com>
7645 * config/arc/arc.md (adddi3): Remove support for mexpand-adddi
7648 * config/arc/arc.opt (mexpand-adddi): Deprecate it.
7649 * doc/invoke.texi (mexpand-adddi): Update text.
7651 2017-07-17 Claudiu Zissulescu <claziss@synopsys.com>
7653 * config/arc/arc.md (clzsi2): Expand to an arc_clzsi2 instruction
7654 that also clobbers the CC register. The old expand code is moved
7656 (*arc_clzsi2): ... here.
7657 (ctzsi2): Expand to an arc_ctzsi2 instruction that also clobbers
7658 the CC register. The old expand code is moved to ...
7659 (arc_ctzsi2): ... here.
7661 2017-07-17 Claudiu Zissulescu <claziss@synopsys.com>
7663 * config/arc/arc.opt (mindexed-loads): Use initial value
7664 TARGET_INDEXED_LOADS_DEFAULT.
7665 (mauto-modify-reg): Use initial value
7666 TARGET_AUTO_MODIFY_REG_DEFAULT.
7667 * config/arc/elf.h (TARGET_INDEXED_LOADS_DEFAULT): Define.
7668 (TARGET_AUTO_MODIFY_REG_DEFAULT): Likewise.
7669 * config/arc/linux.h (TARGET_INDEXED_LOADS_DEFAULT): Define.
7670 (TARGET_AUTO_MODIFY_REG_DEFAULT): Likewise.
7672 2017-07-17 Martin Liska <mliska@suse.cz>
7675 * opts.c (finish_options): Do not allow -fgnu-tm
7676 w/ -fsanitize={kernel-,}address. Say sorry.
7678 2017-07-17 Bin Cheng <bin.cheng@arm.com>
7681 * tree-loop-distribution.c (classify_partition): Only assert on
7682 numer of iterations.
7683 (merge_dep_scc_partitions): Delete prameter. Update function call.
7684 (distribute_loop): Remove code handling loop with unknown niters.
7685 (pass_loop_distribution::execute): Skip loop with unknown niters.
7687 2017-07-17 Bin Cheng <bin.cheng@arm.com>
7690 * tree-loop-distribution.c (merge_dep_scc_partitions): Sink call to
7691 function sort_partitions_by_post_order.
7693 2017-07-17 Bin Cheng <bin.cheng@arm.com>
7695 PR tree-optimization/81374
7696 * tree-loop-distribution.c (pass_loop_distribution::execute): Record
7697 the max index of basic blocks, rather than number of basic blocks.
7699 2017-07-17 Claudiu Zissulescu <claziss@synopsys.com>
7701 * config/arc/arc-protos.h (arc_legitimate_pc_offset_p): Remove
7703 (arc_legitimate_pic_operand_p): Likewise.
7704 * config/arc/arc.c (arc_legitimate_pic_operand_p): Remove
7706 (arc_needs_pcl_p): Likewise.
7707 (arc_legitimate_pc_offset_p): Likewise.
7708 (arc_legitimate_pic_addr_p): Remove LABEL_REF case, as this
7709 function is also used in constrains.md.
7710 (arc_legitimate_constant_p): Use arc_legitimate_pic_addr_p to
7711 validate pic constants. Handle CONST_INT, CONST_DOUBLE, MINUS and
7712 PLUS. Only return true/false in known cases, otherwise assert.
7713 (arc_legitimate_address_p): Remove arc_legitimate_pic_addr_p as it
7714 is already called in arc_legitimate_constant_p.
7715 * config/arc/arc.h (CONSTANT_ADDRESS_P): Consider also LABEL for
7717 (LEGITIMATE_PIC_OPERAND_P): Use
7718 arc_raw_symbolic_reference_mentioned_p function.
7719 * config/arc/constraints.md (Cpc): Use arc_legitimate_pic_addr_p
7724 2017-07-17 Claudiu Zissulescu <claziss@synopsys.com>
7725 Andrew Burgess <andrew.burgess@embecosm.com>
7727 * config/arc/arc-protos.h (arc_compute_function_type): Change prototype.
7728 (arc_return_address_register): New function.
7729 * config/arc/arc.c (arc_handle_fndecl_attribute): New function.
7730 (arc_handle_fndecl_attribute): Add naked attribute.
7731 (TARGET_ALLOCATE_STACK_SLOTS_FOR_ARGS): Define.
7732 (TARGET_WARN_FUNC_RETURN): Likewise.
7733 (arc_allocate_stack_slots_for_args): New function.
7734 (arc_warn_func_return): Likewise.
7735 (machine_function): Change type fn_type.
7736 (arc_compute_function_type): Consider new naked function type,
7737 change function return type.
7738 (arc_must_save_register): Adapt to handle new
7739 arc_compute_function_type's return type.
7740 (arc_expand_prologue): Likewise.
7741 (arc_expand_epilogue): Likewise.
7742 (arc_return_address_regs): Delete.
7743 (arc_return_address_register): New function.
7744 (arc_epilogue_uses): Use above function.
7745 * config/arc/arc.h (arc_return_address_regs): Delete prototype.
7746 (arc_function_type): Change encoding, add naked type.
7747 (ARC_INTERRUPT_P): Change to handle the new encoding.
7748 (ARC_FAST_INTERRUPT_P): Likewise.
7749 (ARC_NORMAL_P): Define.
7750 (ARC_NAKED_P): Likewise.
7751 (arc_compute_function_type): Delete prototype.
7752 * config/arc/arc.md (in_ret_delay_slot): Use
7753 arc_return_address_register function.
7754 (simple_return): Likewise.
7755 (p_return_i): Likewise.
7757 2017-07-17 Jakub Jelinek <jakub@redhat.com>
7759 PR tree-optimization/81428
7760 * match.pd (X / X -> one): Don't optimize _Fract divisions, as 1
7761 can't be built for those types.
7763 2017-07-17 Georg-Johann Lay <avr@gjlay.de>
7765 Remove stuff dead since r239246.
7767 * config/avr/avr-arch.h (avr_inform_devices): Remove dead proto.
7768 * config/avr/avr-devices.c (mcu_name, comparator, avr_mcus_str)
7769 (avr_inform_devices): Remove dead stuff.
7771 2017-07-17 Tamar Christina <tamar.christina@arm.com>
7773 * config/arm/arm_neon.h: Fix softp typo.
7775 2017-07-17 Jakub Jelinek <jakub@redhat.com>
7777 PR tree-optimization/81365
7778 * tree-ssa-phiprop.c (propagate_with_phi): When considering hoisting
7779 aggregate moves onto bb predecessor edges, make sure there are no
7780 loads that could alias the lhs in between the start of bb and the
7783 2017-07-17 Georg-Johann Lay <avr@gjlay.de>
7786 * config/avr/avr.c (avr_mul_highpart_cost): New static function.
7787 (avr_rtx_costs_1) [TRUNCATE]: Use it to compute mul_highpart cost.
7788 [LSHIFTRT, outer_code = TRUNCATE]: Same.
7790 2017-07-17 Jakub Jelinek <jakub@redhat.com>
7792 PR tree-optimization/81396
7793 * tree-ssa-math-opts.c (struct symbolic_number): Add n_ops field.
7794 (init_symbolic_number): Initialize it to 1.
7795 (perform_symbolic_merge): Add n_ops from both operands into the new
7797 (find_bswap_or_nop): Don't consider n->n == cmpnop computations
7798 without base_addr as useless if they need more than one operation.
7799 (bswap_replace): Handle !bswap case for NULL base_addr.
7801 2017-07-17 Tom de Vries <tom@codesourcery.com>
7804 * config/nvptx/nvptx.c (nvptx_single): Insert diverging branch as late
7807 2017-07-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7809 * config/sparc/rtemself.h (TARGET_OS_CPP_BUILTINS): Add
7810 conditional builtin define __FIX_LEON3FT_B2BST.
7812 2017-07-17 Daniel Cederman <cederman@gaisler.com>
7814 * config/sparc/t-rtems: Add mfix-gr712rc multilibs. Replace
7815 MULTILIB_EXCEPTIONS with MULTILIB_REQUIRED. Match -mfix-gr712rc
7818 2017-07-16 Eric Botcazou <ebotcazou@adacore.com>
7820 PR rtl-optimization/81424
7821 * optabs.c (prepare_cmp_insn): Use copy_to_reg instead of force_reg
7822 to remove potential trapping from operands if -fnon-call-exceptions.
7824 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7826 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Use
7827 profile_proability for scalling.
7828 * scale_profile_for_vect_loop.c (scale_profile_for_vect_loop): Likewise.
7830 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7832 * cgraph.c (cgraph_edge::redirect_call_stmt_to_caller): Cleanup.
7834 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7836 * cfgloopmanip.c (scale_loop_profile): Avoid use of REG_BR_PROB_BASE
7837 fixpoint arithmetics.
7839 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7841 * tree-ssa-loop-unswitch.c (hoist_guard): Avoid use of REG_BR_PROB_BASE
7842 fixpoint arithmetics.
7844 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7846 * asan.c (create_cond_insert_point): Avoid use of REG_BR_PROB_BASE
7847 fixpoint arithmetics.
7849 2017-07-16 Jan Hubicka <hubicka@ucw.cz>
7851 * profile-count.h (profile_probability::from_reg_br_prob_note,
7852 profile_probability::to_reg_br_prob_note): New functions.
7853 * doc/rtl.texi (REG_BR_PROB_NOTE): Update documentation.
7854 * reg-notes.h (REG_BR_PROB, REG_BR_PRED): Update docs.
7855 * predict.c (probability_reliable_p): Update.
7856 (edge_probability_reliable_p): Update.
7857 (br_prob_note_reliable_p): Update.
7858 (invert_br_probabilities): Update.
7859 (add_reg_br_prob_note): New function.
7860 (combine_predictions_for_insn): Update.
7861 * asan.c (asan_clear_shadow): Update.
7862 * cfgbuild.c (compute_outgoing_frequencies): Update.
7863 * cfgrtl.c (force_nonfallthru_and_redirect): Update.
7864 (update_br_prob_note): Update.
7865 (rtl_verify_edges): Update.
7866 (purge_dead_edges): Update.
7867 (fixup_reorder_chain): Update.
7868 * emit-rtl.c (try_split): Update.
7869 * ifcvt.c (cond_exec_process_insns): Update.
7870 (cond_exec_process_if_block): Update.
7871 (dead_or_predicable): Update.
7872 * internal-fn.c (expand_addsub_overflow): Update.
7873 (expand_neg_overflow): Update.
7874 (expand_mul_overflow): Update.
7875 * loop-doloop.c (doloop_modify): Update.
7876 * loop-unroll.c (compare_and_jump_seq): Update.
7877 * optabs.c (emit_cmp_and_jump_insn_1): Update.
7878 * predict.h: Update.
7879 * reorg.c (mostly_true_jump): Update.
7881 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Update.
7882 * config/alpha/alpha.c (emit_unlikely_jump): Update.
7883 * config/arc/arc.c: (emit_unlikely_jump): Update.
7884 * config/arm/arm.c: (emit_unlikely_jump): Update.
7885 * config/bfin/bfin.c (cbranch_predicted_taken_p): Update.
7886 * config/frv/frv.c (frv_print_operand_jump_hint): Update.
7887 * config/i386/i386.c (ix86_expand_split_stack_prologue): Update.
7888 (ix86_print_operand): Update.
7889 (ix86_split_fp_branch): Update.
7890 (predict_jump): Update.
7891 * config/ia64/ia64.c (ia64_print_operand): Update.
7892 * config/mmix/mmix.c (mmix_print_operand): Update.
7893 * config/powerpcspe/powerpcspe.c (output_cbranch): Update.
7894 (rs6000_expand_split_stack_prologue): Update.
7895 * config/rs6000/rs6000.c: Update.
7896 * config/s390/s390.c (s390_expand_vec_strlen): Update.
7897 (s390_expand_vec_movstr): Update.
7898 (s390_expand_cs_tdsi): Update.
7899 (s390_expand_split_stack_prologue): Update.
7900 * config/sh/sh.c (sh_print_operand): Update.
7901 (expand_cbranchsi4): Update.
7902 (expand_cbranchdi4): Update.
7903 * config/sparc/sparc.c (output_v9branch): Update.
7904 * config/spu/spu.c (get_branch_target): Update.
7905 (ea_load_store_inline): Update.
7906 * config/tilegx/tilegx.c (cbranch_predicted_p): Update.
7907 * config/tilepro/tilepro.c: Update.
7909 2017-07-16 Eric Botcazou <ebotcazou@adacore.com>
7911 * gimplify.c (mostly_copy_tree_r): Revert latest change.
7912 (gimplify_save_expr): Likewise.
7914 2017-07-07 Jan Hubicka <hubicka@ucw.cz>
7916 * ipa-visibility.c (function_and_variable_visibility): Fix pasto.
7918 2017-07-07 Jan Hubicka <hubicka@ucw.cz>
7920 * ipa-fnsummary.c (pass_data_ipa_fn_summary): Use
7922 * timevar.def (TV_IPA_FNSUMMARY): Define.
7924 2017-07-16 Daniel Cederman <cederman@gaisler.com>
7926 * config/sparc/sparc.md (divdf3_fix): Add NOP to prevent back
7927 to back store errata sensitive sequence from being generated.
7928 (sqrtdf2_fix): Likewise.
7930 2017-07-07 Jan Hubicka <hubicka@ucw.cz>
7932 * tree-ssa-threadupdate.c (compute_path_counts,
7933 update_joiner_offpath_counts): Use profile_probability.
7935 2017-07-15 Thomas Preud'homme <thomas.preudhomme@arm.com>
7938 2017-07-14 Thomas Preud'homme <thomas.preudhomme@arm.com>
7940 * config/arm/arm-c.c (arm_cpu_builtins): Define
7941 __ARM_FEATURE_NUMERIC_MAXMIN solely based on TARGET_VFP5.
7943 2017-07-14 Kelvin Nilsen <kelvin@gcc.gnu.org>
7945 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7946 array entries to represent __ieee128 versions of the
7947 scalar_test_data_class, scalar_test_neg, scalar_extract_exp,
7948 scalar_extract_sig, and scalar_insert_exp built-in functions.
7949 (altivec_resolve_overloaded_builtin): Add special case handling
7950 for the __builtin_scalar_insert_exp function, as represented by
7951 the P9V_BUILTIN_VEC_VSIEDP constant.
7952 * config/rs6000/rs6000-builtin.def (VSEEQP): Add scalar extract
7953 exponent support for __ieee128 argument.
7954 (VSESQP): Add scalar extract signature support for __ieee128
7956 (VSTDCNQP): Add scalar test negative support for __ieee128
7958 (VSIEQP): Add scalar insert exponent support for __int128 argument
7959 with __ieee128 result.
7960 (VSIEQPF): Add scalar insert exponent support for __ieee128
7961 argument with __ieee128 result.
7962 (VSTDCQP): Add scalar test data class support for __ieee128
7964 (VSTDCNQP): Add overload support for scalar test negative with
7966 (VSTDCQP): Add overload support for scalar test data class
7968 * config/rs6000/vsx.md (UNSPEC_VSX_SXSIG) Replace
7970 (UNSPEC_VSX_SIEXPQP): New constant.
7971 (xsxexpqp): New insn for VSX scalar extract exponent quad
7973 (xsxsigqp): New insn for VSX scalar extract significand quad
7975 (xsiexpqpf): New insn for VSX scalar insert exponent quad
7976 precision with floating point argument.
7977 (xststdcqp): New expand for VSX scalar test data class quad
7979 (xststdcnegqp): New expand for VSX scalar test negative quad
7981 (xststdcqp): New insn to match expansions for VSX scalar test data
7982 class quad precision and VSX scalar test negative quad precision.
7983 * config/rs6000/rs6000.c (rs6000_expand_binop_builtin): Add
7984 special case operand checking to enforce that second operand of
7985 VSX scalar test data class with quad precision argument is a 7-bit
7987 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Add
7988 prototypes and descriptions of __ieee128 versions of
7989 scalar_extract_exp, scalar_extract_sig, scalar_insert_exp,
7990 scalar_test_data_class, and scalar_test_neg built-in functions.
7992 2016-07-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7994 PR tree-optimization/81162
7995 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Don't
7996 replace a negate with an add.
7998 2017-07-14 James Greenhalgh <james.greenhalgh@arm.com>
8000 * doc/invoke.texi (arm/-mcpu): Document +crypto.
8002 2017-07-14 Thomas Preud'homme <thomas.preudhomme@arm.com>
8004 * config/arm/arm-c.c (arm_cpu_builtins): Define
8005 __ARM_FEATURE_NUMERIC_MAXMIN solely based on TARGET_VFP5.
8007 2017-07-14 Thomas Preud'homme <thomas.preudhomme@arm.com>
8009 * config/arm/arm-cpus.in (cortex-r52): Add new entry.
8010 (armv8-r): Set ARM Cortex-R52 as default CPU.
8011 * config/arm/arm-tables.opt: Regenerate.
8012 * config/arm/arm-tune.md: Regenerate.
8013 * config/arm/driver-arm.c (arm_cpu_table): Add entry for ARM
8015 * doc/invoke.texi: Mention -mtune=cortex-r52 and availability of fp.dp
8016 extension for -mcpu=cortex-r52.
8018 2017-07-14 Thomas Preud'homme <thomas.preudhomme@arm.com>
8020 * config/arm/arm-isa.h (isa_bit_FP_ARMv8): Delete enumerator.
8021 (ISA_FP_ARMv8): Define as ISA_FPv5 and ISA_FP_D32.
8022 * config/arm/arm-cpus.in (armv8-r): Define fp.sp as enabling FPv5.
8023 (fp-armv8): Define it as FP_ARMv8 only.
8024 config/arm/arm.h (TARGET_FPU_ARMV8): Delete.
8025 (TARGET_VFP_FP16INST): Define using TARGET_VFP5 rather than
8027 config/arm/arm.c (arm_rtx_costs_internal): Replace checks against
8028 TARGET_FPU_ARMV8 by checks against TARGET_VFP5.
8029 * config/arm/arm-builtins.c (arm_builtin_vectorized_function): Define
8030 first ARM_CHECK_BUILTIN_MODE definition using TARGET_VFP5 rather
8031 than TARGET_FPU_ARMV8.
8032 * config/arm/arm-c.c (arm_cpu_builtins): Likewise for
8033 __ARM_FEATURE_NUMERIC_MAXMIN macro definition.
8034 * config/arm/arm.md (cmov<mode>): Condition on TARGET_VFP5 rather than
8036 * config/arm/neon.md (neon_vrint): Likewise.
8037 (neon_vcvt): Likewise.
8038 (neon_<fmaxmin_op><mode>): Likewise.
8039 (<fmaxmin><mode>3): Likewise.
8040 * config/arm/vfp.md (l<vrint_pattern><su_optab><mode>si2): Likewise.
8041 * config/arm/predicates.md (arm_cond_move_operator): Check against
8042 TARGET_VFP5 rather than TARGET_FPU_ARMV8 and fix spacing.
8044 2017-07-14 Jackson Woodruff <jackson.woodruff@arm.com>
8046 * config/aarch64/aarch64.c (aarch64_print_operand): Move comments
8049 2017-07-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8051 * gimple-ssa-store-merging.c (clear_bit_region): Replace reference to
8052 loop in comment with memset.
8054 2017-07-14 Martin Liska <mliska@suse.cz>
8056 * cfgexpand.c (expand_gimple_basic_block): Remove dead comment.
8057 * dwarf2out.c (is_java): Remove the function.
8058 (output_pubname): Remove usage of the function.
8059 (lower_bound_default): Remove usage of DW_LANG_Java.
8060 (gen_compile_unit_die): Likewise.
8061 * gcc.c: Remove compiler defaults for .java and .zip files.
8062 * gimple-expr.c (remove_suffix): Change as there's no longer
8063 extension than 4-letter one.
8064 * gimplify.c (mostly_copy_tree_r): Remove Java-special part.
8065 (gimplify_save_expr): Likewise.
8066 * ipa-utils.h (polymorphic_type_binfo_p): Remove the comment
8067 as it's possible even for other languages than Java.
8068 * langhooks.h (struct lang_hooks): Remove Java from a comment.
8069 * lto-opts.c (lto_write_options): Remove reference to Java.
8070 * opts.c (strip_off_ending): Update file extension handling.
8071 * tree-cfg.c (verify_gimple_call): Remove comment with Java.
8072 * tree-eh.c (lower_resx): Likewise.
8073 * tree.c (free_lang_data_in_type): Remove dead code.
8074 (find_decls_types_r): Likewise.
8075 (build_common_builtin_nodes): Remove Java from a comment.
8076 (verify_type): Remove dead code.
8077 * varasm.c (assemble_external): Remove Java from a comment.
8079 2017-07-14 Martin Liska <mliska@suse.cz>
8081 * opts.c (finish_options): Add quotes.
8082 (common_handle_option): Likewise.
8084 2017-07-14 Martin Liska <mliska@suse.cz>
8086 * dbxout.c (get_lang_number): Do not handle GNU Pascal.
8087 * dbxout.h (extern void dbxout_stab_value_internal_label_diff):
8089 * dwarf2out.c (lower_bound_default): Do not handle
8091 (gen_compile_unit_die): Likewise.
8092 * gcc.c: Remove default extension binding for GNU Pascal.
8093 * stmt.c: Remove Pascal language from a comment.
8094 * xcoffout.c: Likewise.
8096 2017-07-13 David Malcolm <dmalcolm@redhat.com>
8099 * diagnostic-show-locus.c (fixit_cmp): New function.
8100 (layout::layout): Sort m_fixit_hints.
8101 (column_range::column_range): Assert that the values are valid.
8102 (struct char_span): New struct.
8103 (correction::overwrite): New method.
8104 (struct source_line): New struct.
8105 (line_corrections::add_hint): Add assertions. Reimplement memcpy
8106 calls in terms of classes source_line and char_span, and
8107 correction::overwrite.
8108 (selftest::test_overlapped_fixit_printing_2): New function.
8109 (selftest::diagnostic_show_locus_c_tests): Call it.
8111 2017-07-13 Will Schmidt <will_schmidt@vnet.ibm.com>
8113 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Return
8114 early if there is no lhs.
8116 2017-07-13 Martin Liska <mliska@suse.cz>
8118 * dwarf2out.c (gen_pointer_type_die): Remove dead code.
8119 (gen_reference_type_die): Likewise.
8120 * stor-layout.c: Remove Pascal-related comment.
8122 2017-07-13 Martin Liska <mliska@suse.cz>
8124 * opts.c (finish_options): Add quotes to error messages.
8125 (parse_sanitizer_options): Likewise.
8127 2017-07-13 Thomas Preud'homme <thomas.preudhomme@arm.com>
8129 * doc/invoke.texi (armv8-r): Document +fp.sp ARMv8-R extension.
8131 2017-07-13 Richard Earnshaw <rearnsha@arm.com>
8133 * config/arm/vxworks.h (TARGET_ENDIAN_DEFAULT): Define.
8135 2017-07-13 Maxim Ostapenko <m.ostapenko@samsung.com>
8137 * asan.c (asan_emit_allocas_unpoison): Use ptr_mode for arguments
8139 * builtins.c (expand_asan_emit_allocas_unpoison): Likewise.
8141 2017-07-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8144 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): If GLIBC
8145 provides the hardware capability bits, define the macro
8146 __BUILTIN_CPU_SUPPORTS__.
8147 * config/rs6000/rs6000.c (cpu_expand_builtin): Generate a warning
8148 if GLIBC does not provide the hardware capability bits. Add a
8149 gcc_unreachable call if the built-in cpu function is neither
8150 __builtin_cpu_is nor __builtin_cpu_supports.
8151 (rs6000_get_function_versions_dispatcher): Change the warning
8152 that an old GLIBC is used which does not export the capability
8153 bits to be an error.
8154 * doc/extend.texi (target_clones attribute): Document the
8155 restriction that GLIBC 2.23 or newer is needed on the PowerPC.
8156 (PowerPC built-in functions): Document that GLIBC 2.23 or newer is
8157 needed by __builtin_cpu_is and __builtin_cpu_supports. Document
8158 the macros defined by GCC if the newer GLIBC is available.
8160 2017-07-12 Jeff Law <law@redhat.com>
8162 * config/riscv/riscv.c: Remove unnecessary includes. Reorder
8163 remaining includes slightly.
8164 * config/riscv/riscv-builtins.c: Include profile-count.h.
8166 2017-07-12 Georg-Johann Lay <avr@gjlay.de>
8169 * config/avr/avr.c (avr_set_current_function): In diagnostic
8170 messages: Quote keywords and (parts of) identifiers.
8171 [WITH_AVRLIBC]: Warn for functions named "ISR", "SIGNAL" or
8174 2017-07-12 Carl Love <cel@us.ibm.com>
8176 * config/rs6000/rs6000-c.c: Add support for built-in functions
8177 vector bool char vec_revb (vector bool char);
8178 vector bool short vec_revb (vector short char);
8179 vector bool int vec_revb (vector bool int);
8180 vector bool long long vec_revb (vector bool long long);
8181 * doc/extend.texi: Update the built-in documentation file for the
8182 new built-in functions.
8184 2017-07-12 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
8186 * config/s390/s390.md: Remove movcc splitter.
8188 2017-07-12 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
8190 * config/s390/s390.c (s390_rtx_costs): Return proper costs for
8191 load/store on condition.
8193 2017-07-12 Georg-Johann Lay <avr@gjlay.de>
8196 * config/avr/avr.c (avr_encode_section_info)
8197 [progmem && !TREE_READONLY]: Error if progmem object needs
8200 2017-07-11 Michael Collison <michael.collison@arm.com>
8202 * config/aarch64/aarch64-simd.md (aarch64_sub<mode>_compare0):
8205 2017-07-11 Carl Love <cel@us.ibm.com>
8207 * config/rs6000/rs6000-c.c: Add support for builtins
8208 vector unsigned int vec_parity_lsbb (vector signed int);
8209 vector unsigned int vec_parity_lsbb (vector unsigned int);
8210 vector unsigned __int128 vec_parity_lsbb (vector signed __int128);
8211 vector unsigned __int128 vec_parity_lsbb (vector unsigned __int128);
8212 vector unsigned long long vec_parity_lsbb (vector signed long long);
8213 vector unsigned long long vec_parity_lsbb (vector unsigned long long);
8214 * config/rs6000/rs6000-builtin.def (VPARITY_LSBB): Add BU_P9V_OVERLOAD1.
8215 * config/rs6000/altivec.h (vec_parity_lsbb): Add define.
8216 * doc/extend.texi: Update the built-in documentation file for the
8217 new built-in functions.
8219 2017-07-11 David Malcolm <dmalcolm@redhat.com>
8221 * diagnostic-show-locus.c: Include "gcc-rich-location.h".
8222 (layout::m_primary_loc): New field.
8223 (layout::layout): Initialize new field. Move location filtering
8224 logic from here to...
8225 (layout::maybe_add_location_range): ...this new method. Add
8226 support for filtering to just the lines already specified by other
8228 (layout::will_show_line_p): New method.
8229 (gcc_rich_location::add_location_if_nearby): New method.
8230 (selftest::test_add_location_if_nearby): New test function.
8231 (selftest::diagnostic_show_locus_c_tests): Call it.
8232 * gcc-rich-location.h (gcc_rich_location::add_location_if_nearby):
8235 2017-07-11 Tom de Vries <tom@codesourcery.com>
8237 * config/nvptx/nvptx.c (WORKAROUND_PTXJIT_BUG): New macro.
8238 (bb_first_real_insn): New function.
8239 (nvptx_single): Add extra initialization of broadcasted condition
8242 2017-07-11 Nathan Sidwell <nathan@acm.org>
8244 * dwarf2out.c (gen_member_die): Remove useless check for anon ctors.
8246 2017-07-11 Georg-Johann Lay <avr@gjlay.de>
8248 * doc/extend.texi (AVR Function Attributes): Remove weblink to
8249 Binutils doc as TEXI will mess them up.
8250 * doc/invoke.texi (AVR Options): Same here.
8252 2017-07-11 Daniel Cederman <cederman@gaisler.com>
8254 * config/sparc/sparc.opt (mfix-ut700): New option.
8255 (mfix-gr712rc): Likewise.
8256 (sparc_fix_b2bst): New variable.
8257 * doc/invoke.texi (SPARC options): Document them.
8258 (ARM options): Fix warnings.
8259 * config/sparc/sparc.c (sparc_do_work_around_errata): Insert NOP
8260 instructions to prevent sequences that can trigger the store-store
8261 errata for certain LEON3FT processors.
8262 (pass_work_around_errata::gate): Also test sparc_fix_b2bst.
8263 (sparc_option_override): Set sparc_fix_b2bst appropriately.
8264 * config/sparc/sparc.md (fix_b2bst): New attribute.
8265 (in_branch_delay): Prevent stores in delay slot if fix_b2bst.
8267 2017-07-10 Uros Bizjak <ubizjak@gmail.com>
8270 * config/i386/i386.md (divsf3): Add TARGET_SSE to TARGET_SSE_MATH.
8272 (*rsqrtsf2_sse): Ditto.
8274 (div<mode>3): Macroize insn from divdf3 and divsf3
8275 using MODEF mode iterator.
8277 2017-07-10 Martin Sebor <msebor@redhat.com>
8279 PR tree-optimization/80397
8280 * gimple-ssa-sprintf.c (format_integer): Use INTEGRAL_TYPE_P()
8281 instead of testing for equality to INTEGER_TYPE.
8283 2017-07-10 Vineet Gupta <vgupta@synopsys.com>
8285 * config.gcc: Remove uclibc from arc target spec.
8287 2017-07-10 Claudiu Zissulescu <claziss@synopsys.com>
8289 * config/arc/arc.h (ADDITIONAL_REGISTER_NAMES): Define.
8291 2017-07-07 Jan Hubicka <hubicka@ucw.cz>
8294 * lto-wrapper.c (remove_option): New function.
8295 (merge_and_complain): Merge PIC/PIE options more realistically.
8297 2017-07-10 Georg-Johann Lay <avr@gjlay.de>
8299 Better ISR prologues by supporting GASes __gcc_isr pseudo insn.
8303 * configure.ac [target=avr]: Add GAS check for -mgcc-isr.
8304 (HAVE_AS_AVR_MGCCISR_OPTION): If so, AC_DEFINE it.
8305 * config.in: Regenerate.
8306 * configure: Regenerate.
8307 * doc/extend.texi (AVR Function Attributes) <no_gccisr>: Document it.
8308 * doc/invoke.texi (AVR Options) <-mgas-isr-prologues>: Document it.
8309 * config/avr/avr.opt (-mgas-isr-prologues): New option and...
8310 (TARGET_GASISR_PROLOGUES): ...target mask.
8311 * common/config/avr/avr-common.c
8312 (avr_option_optimization_table) [OPT_LEVELS_1_PLUS_NOT_DEBUG]:
8313 Set -mgas-isr-prologues.
8314 * config/avr/avr-passes.def (avr_pass_pre_proep): Add
8315 INSERT_PASS_BEFORE for it.
8316 * config/avr/avr-protos.h (make_avr_pass_pre_proep): New proto.
8317 * config/avr/avr.c (avr_option_override)
8318 [!HAVE_AS_AVR_MGCCISR_OPTION]: Unset TARGET_GASISR_PROLOGUES.
8319 (avr_no_gccisr_function_p, avr_hregs_split_reg): New static functions.
8320 (avr_attribute_table) <no_gccisr>: Add new function attribute.
8321 (avr_set_current_function) <is_no_gccisr>: Init machine field.
8322 (avr_pass_data_pre_proep, avr_pass_pre_proep): New pass data
8324 (make_avr_pass_pre_proep): New function.
8325 (emit_push_sfr) <treg>: Add argument to function and use it
8327 (avr_expand_prologue) [machine->gasisr.maybe]: Emit gasisr insn
8328 and set machine->gasisr.yes.
8329 (avr_expand_epilogue) [machine->gasisr.yes]: Similar.
8330 (avr_asm_function_end_prologue) [machine->gasisr.yes]: Add
8331 __gcc_isr.n_pushed to .L__stack_usage.
8332 (TARGET_ASM_FINAL_POSTSCAN_INSN): Define to...
8333 (avr_asm_final_postscan_insn): ...this new static function.
8334 * config/avr/avr.h (machine_function)
8335 <is_no_gccisr, use_L__stack_usage>: New fields.
8336 <gasisr, gasisr.yes, gasisr.maybe, gasisr.regno>: New fields.
8337 * config/avr/avr.md (UNSPECV_GASISR): Add unspecv enum.
8338 (GASISR_Prologue, GASISR_Epilogue, GASISR_Done): New define_constants.
8339 (gasisr, *gasisr): New expander and insn.
8340 * config/avr/gen-avr-mmcu-specs.c (print_mcu)
8341 [HAVE_AS_AVR_MGCCISR_OPTION]: Print asm_gccisr spec.
8342 * config/avr/specs.h (ASM_SPEC) <asm_gccisr>: Add sub spec.
8344 2017-07-10 Richard Earnshaw <rearnsha@arm.com>
8346 * config/arm/parsecpu.awk (gen_comm_data): Do not escape single quotes
8349 2017-07-10 Georg-Johann Lay <avr@gjlay.de>
8351 Move jump-tables out of .text again.
8354 * config/avr/avr.c (ASM_OUTPUT_ADDR_VEC_ELT): Remove function.
8355 (ASM_OUTPUT_ADDR_VEC): New function.
8356 (avr_adjust_insn_length) [JUMP_TABLE_DATA_P]: Return 0.
8357 (avr_final_prescan_insn) [avr_log.insn_addresses]: Dump
8358 INSN_ADDRESSes as asm comment.
8359 * config/avr/avr.h (JUMP_TABLES_IN_TEXT_SECTION): Adjust comment.
8360 (ASM_OUTPUT_ADDR_VEC_ELT): Remove define.
8361 (ASM_OUTPUT_ADDR_VEC): Define to avr_output_addr_vec.
8362 * config/avr/avr.md (*tablejump): Adjust comment.
8363 * config/avr/elf.h (ASM_OUTPUT_BEFORE_CASE_LABEL): Remove.
8364 * config/avr/avr-log.c (avr_log_set_avr_log) <insn_addresses>:
8366 * config/avr/avr-protos.h (avr_output_addr_vec_elt): Remove proto.
8367 (avr_output_addr_vec): New proto.
8368 (avr_log_t) <insn_addresses>: New field.
8370 2017-07-09 H.J. Lu <hongjiu.lu@intel.com>
8373 * config/i386/i386.c (ix86_function_arg_advance): Set
8374 outgoing_args_on_stack to true if there are outgoing arguments
8376 (ix86_function_arg): Likewise.
8377 (ix86_get_drap_rtx): Use DRAP only if there are outgoing
8378 arguments on stack and ACCUMULATE_OUTGOING_ARGS is false.
8379 * config/i386/i386.h (machine_function): Add
8380 outgoing_args_on_stack.
8382 2017-07-09 Krister Walfridsson <krister.walfridsson@gmail.com>
8384 * config.gcc (*-*-netbsd*): Remove check for NetBSD versions not
8386 * config/netbsd.h (NETBSD_LIBGCC_SPEC): Always enable pthreads.
8388 2017-07-08 Richard Sandiford <richard.sandiford@linaro.org>
8390 * Makefile.in (HOOKS_H, RTL_BASE_H, FUNCTION_H, EXPR_H, REGS_H)
8391 (REAL_H): Remove $(MACHMODE_H).
8392 (FIXED_VALUE_H, TREE_CORE_H, CFGLOOP_H): Remove $(MACHMODE_H) and
8394 (CORETYPES_H): Add signop.h, wide-int.h, wide-int-print.h,
8395 $(MACHMODE_H) and double-int.h.
8396 (build/min-insn-modes.o): Depend on $(CORETYPES_H) rather than
8398 (gengtype-state.o, gengtype.o, build/gengtype.o): Don't depend on
8401 2017-07-07 Andrew Pinski <apinski@cavium.com>
8403 * config/aarch64/aarch64.c (aarch_macro_fusion_pair_p): Check
8404 prev_set and curr_set for AARCH64_FUSE_ALU_BRANCH.
8406 2017-07-07 Michael Meissner <meissner@linux.vnet.ibm.com>
8408 * config/rs6000/rs6000.c (rs6000_get_function_versions_dispatcher):
8409 Add warning if GCC was not configured to link against a GLIBC that
8410 exports the hardware capability bits.
8411 (make_resolver_func): Make resolver function private and not a
8412 COMDAT function. Create the name with clone_function_name instead
8413 of make_unique_name.
8416 * config/rs6000/rs6000.md (HI sign_extend splitter): Use the
8417 correct operand in doing the split.
8419 2017-07-07 Carl Love <cel@us.ibm.com>
8421 * config/rs6000/rs6000-c: Add support for built-in function
8422 vector unsigned short vec_pack_to_short_fp32 (vector float,
8424 * config/rs6000/rs6000-builtin.def (CONVERT_4F32_8I16): Add
8425 BU_P9V_AV_2 and BU_P9V_OVERLOAD_2 definitions.
8426 * config/rs6000/altivec.h (vec_pack_to_short_fp32): Add define.
8427 * config/rs6000/altivec.md(UNSPEC_CONVERT_4F32_8I16): Add UNSPEC.
8428 (convert_4f32_8i16): Add define_expand.
8429 * doc/extend.texi: Update the built-in documentation file for the
8430 new built-in function.
8432 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8434 * config/sparc/m8.md: New file.
8435 * config/sparc/sparc.md: Include m8.md.
8437 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8439 * config/sparc/sparc.opt: New option -mvis4b.
8440 * config/sparc/sparc.c (dump_target_flag_bits): Handle MASK_VIS4B.
8441 (sparc_option_override): Handle VIS4B.
8442 (enum sparc_builtins): Define
8443 SPARC_BUILTIN_DICTUNPACK{8,16,32},
8444 SPARC_BUILTIN_FPCMP{LE,GT,EQ,NE}{8,16,32}SHL,
8445 SPARC_BUILTIN_FPCMPU{LE,GT}{8,16,32}SHL,
8446 SPARC_BUILTIN_FPCMPDE{8,16,32}SHL and
8447 SPARC_BUILTIN_FPCMPUR{8,16,32}SHL.
8448 (check_constant_argument): New function.
8449 (sparc_vis_init_builtins): Define builtins
8450 __builtin_vis_dictunpack{8,16,32},
8451 __builtin_vis_fpcmp{le,gt,eq,ne}{8,16,32}shl,
8452 __builtin_vis_fpcmpu{le,gt}{8,16,32}shl,
8453 __builtin_vis_fpcmpde{8,16,32}shl and
8454 __builtin_vis_fpcmpur{8,16,32}shl.
8455 (sparc_expand_builtin): Check that the constant operands to
8456 __builtin_vis_fpcmp*shl and _builtin_vis_dictunpack* are indeed
8457 constant and in range.
8458 * config/sparc/sparc-c.c (sparc_target_macros): Handle
8460 * config/sparc/sparc.h (SPARC_IMM2_P): Define.
8461 (SPARC_IMM5_P): Likewise.
8462 * config/sparc/sparc.md (cpu_feature): Add new feagure "vis4b".
8463 (enabled): Handle vis4b.
8464 (UNSPEC_DICTUNPACK): New unspec.
8465 (UNSPEC_FPCMPSHL): Likewise.
8466 (UNSPEC_FPUCMPSHL): Likewise.
8467 (UNSPEC_FPCMPDESHL): Likewise.
8468 (UNSPEC_FPCMPURSHL): Likewise.
8469 (cpu_feature): New CPU feature `vis4b'.
8470 (dictunpack{8,16,32}): New insns.
8471 (FPCSMODE): New mode iterator.
8472 (fpcscond): New code iterator.
8473 (fpcsucond): Likewise.
8474 (fpcmp{le,gt,eq,ne}{8,16,32}{si,di}shl): New insns.
8475 (fpcmpu{le,gt}{8,16,32}{si,di}shl): Likewise.
8476 (fpcmpde{8,16,32}{si,di}shl): Likewise.
8477 (fpcmpur{8,16,32}{si,di}shl): Likewise.
8478 * config/sparc/constraints.md: Define constraints `q' for unsigned
8479 2-bit integer constants and `t' for unsigned 5-bit integer
8481 * config/sparc/predicates.md (imm5_operand_dictunpack8): New
8483 (imm5_operand_dictunpack16): Likewise.
8484 (imm5_operand_dictunpack32): Likewise.
8485 (imm2_operand): Likewise.
8486 * doc/invoke.texi (SPARC Options): Document -mvis4b.
8487 * doc/extend.texi (SPARC VIS Built-in Functions): Document the
8488 ditunpack* and fpcmp*shl builtins.
8490 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8492 * config.gcc: Handle m8 in --with-{cpu,tune} options.
8493 * config.in: Add HAVE_AS_SPARC6 define.
8494 * config/sparc/driver-sparc.c (cpu_names): Add entry for the SPARC
8496 * config/sparc/sol2.h (CPP_CPU64_DEFAULT_SPEC): Define for
8498 (ASM_CPU32_DEFAUILT_SPEC): Likewise.
8499 (CPP_CPU_SPEC): Handle m8.
8500 (ASM_CPU_SPEC): Likewise.
8501 * config/sparc/sparc-opts.h (enum processor_type): Add
8503 * config/sparc/sparc.c (m8_costs): New struct.
8504 (sparc_option_override): Handle TARGET_CPU_m8.
8505 (sparc32_initialize_trampoline): Likewise.
8506 (sparc64_initialize_trampoline): Likewise.
8507 (sparc_issue_rate): Likewise.
8508 (sparc_register_move_cost): Likewise.
8509 * config/sparc/sparc.h (TARGET_CPU_m8): Define.
8510 (CPP_CPU64_DEFAULT_SPEC): Define for M8.
8511 (ASM_CPU64_DEFAULT_SPEC): Likewise.
8512 (CPP_CPU_SPEC): Handle M8.
8513 (ASM_CPU_SPEC): Likewise.
8514 (AS_M8_FLAG): Define.
8515 * config/sparc/sparc.md: Add m8 to the cpu attribute.
8516 * config/sparc/sparc.opt: New option -mcpu=m8 for sparc targets.
8517 * configure.ac (HAVE_AS_SPARC6): Check for assembler support for
8519 * configure: Regenerate.
8520 * doc/invoke.texi (SPARC Options): Document -mcpu=m8 and
8523 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8525 * config/sparc/niagara7.md: Rework the DFA scheduler to use insn
8527 * config/sparc/sparc.md: Remove the `v3pipe' insn attribute.
8528 ("*movdi_insn_sp32"): Do not set v3pipe.
8529 ("*movsi_insn"): Likewise.
8530 ("*movdi_insn_sp64"): Likewise.
8531 ("*movsf_insn"): Likewise.
8532 ("*movdf_insn_sp32"): Likewise.
8533 ("*movdf_insn_sp64"): Likewise.
8534 ("*zero_extendsidi2_insn_sp64"): Likewise.
8535 ("*sign_extendsidi2_insn"): Likewise.
8536 ("*mov<VM32:mode>_insn"): Likewise.
8537 ("*mov<VM64:mode>_insn_sp64"): Likewise.
8538 ("*mov<VM64:mode>_insn_sp32"): Likewise.
8539 ("<plusminus_insn><VADDSUB:mode>3"): Likewise.
8540 ("<vlop:code><VL:mode>3"): Likewise.
8541 ("*not_<vlop:code><VL:mode>3"): Likewise.
8542 ("*nand<VL:mode>_vis"): Likewise.
8543 ("*<vlnotop:code>_not1<VL:mode>_vis"): Likewise.
8544 ("*<vlnotop:code>_not2<VL:mode>_vis"): Likewise.
8545 ("one_cmpl<VL:mode>2"): Likewise.
8546 ("faligndata<VM64:mode>_vis"): Likewise.
8547 ("alignaddrsi_vis"): Likewise.
8548 ("alignaddrdi_vis"): Likweise.
8549 ("alignaddrlsi_vis"): Likewise.
8550 ("alignaddrldi_vis"): Likewise.
8551 ("fcmp<gcond:code><GCM:gcm_name><P:mode>_vis"): Likewise.
8552 ("bmaskdi_vis"): Likewise.
8553 ("bmasksi_vis"): Likewise.
8554 ("bshuffle<VM64:mode>_vis"): Likewise.
8555 ("cmask8<P:mode>_vis"): Likewise.
8556 ("cmask16<P:mode>_vis"): Likewise.
8557 ("cmask32<P:mode>_vis"): Likewise.
8558 ("pdistn<P:mode>_vis"): Likewise.
8559 ("<vis3_addsub_ss_patname><VASS:mode>3"): Likewise.
8561 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8563 * config/sparc/sparc.md ("subtype"): New insn attribute.
8564 ("*wrgsr_sp64"): Set insn subtype.
8565 ("*rdgsr_sp64"): Likewise.
8566 ("alignaddrsi_vis"): Likewise.
8567 ("alignaddrdi_vis"): Likewise.
8568 ("alignaddrlsi_vis"): Likewise.
8569 ("alignaddrldi_vis"): Likewise.
8570 ("<plusminus_insn><VADDSUB:mode>3"): Likewise.
8571 ("fexpand_vis"): Likewise.
8572 ("fpmerge_vis"): Likewise.
8573 ("faligndata<VM64:mode>_vis"): Likewise.
8574 ("bshuffle<VM64:mode>_vis"): Likewise.
8575 ("cmask8<P:mode>_vis"): Likewise.
8576 ("cmask16<P:mode>_vis"): Likewise.
8577 ("cmask32<P:mode>_vis"): Likewise.
8578 ("fchksm16_vis"): Likewise.
8579 ("v<vis3_shift_patname><GCM:mode>3"): Likewise.
8580 ("fmean16_vis"): Likewise.
8581 ("fp<plusminus_insn>64_vis"): Likewise.
8582 ("<plusminus_insn>v8qi3"): Likewise.
8583 ("<vis3_addsub_ss_patname><VASS:mode>3"): Likewise.
8584 ("<vis4_minmax_patname><VMMAX:mode>3"): Likewise.
8585 ("<vis4_uminmax_patname><VMMAX:mode>3"): Likewise.
8586 ("<vis3_addsub_ss_patname>v8qi3"): Likewise.
8587 ("<vis4_addsub_us_patname><VAUS:mode>3"): Likewise.
8588 ("*movqi_insn"): Likewise.
8589 ("*movhi_insn"): Likewise.
8590 ("*movsi_insn"): Likewise.
8591 ("movsi_pic_gotdata_op"): Likewise.
8592 ("*movdi_insn_sp32"): Likewise.
8593 ("*movdi_insn_sp64"): Likewise.
8594 ("movdi_pic_gotdata_op"): Likewise.
8595 ("*movsf_insn"): Likewise.
8596 ("*movdf_insn_sp32"): Likewise.
8597 ("*movdf_insn_sp64"): Likewise.
8598 ("*zero_extendhisi2_insn"): Likewise.
8599 ("*zero_extendqihi2_insn"): Likewise.
8600 ("*zero_extendqisi2_insn"): Likewise.
8601 ("*zero_extendqidi2_insn"): Likewise.
8602 ("*zero_extendhidi2_insn"): Likewise.
8603 ("*zero_extendsidi2_insn_sp64"): Likewise.
8604 ("ldfsr"): Likewise.
8605 ("prefetch_64"): Likewise.
8606 ("prefetch_32"): Likewise.
8607 ("tie_ld32"): Likewise.
8608 ("tie_ld64"): Likewise.
8609 ("*tldo_ldub_sp32"): Likewise.
8610 ("*tldo_ldub1_sp32"): Likewise.
8611 ("*tldo_ldub2_sp32"): Likewise.
8612 ("*tldo_ldub_sp64"): Likewise.
8613 ("*tldo_ldub1_sp64"): Likewise.
8614 ("*tldo_ldub2_sp64"): Likewise.
8615 ("*tldo_ldub3_sp64"): Likewise.
8616 ("*tldo_lduh_sp32"): Likewise.
8617 ("*tldo_lduh1_sp32"): Likewise.
8618 ("*tldo_lduh_sp64"): Likewise.
8619 ("*tldo_lduh1_sp64"): Likewise.
8620 ("*tldo_lduh2_sp64"): Likewise.
8621 ("*tldo_lduw_sp32"): Likewise.
8622 ("*tldo_lduw_sp64"): Likewise.
8623 ("*tldo_lduw1_sp64"): Likewise.
8624 ("*tldo_ldx_sp64"): Likewise.
8625 ("*mov<VM32:mode>_insn"): Likewise.
8626 ("*mov<VM64:mode>_insn_sp64"): Likewise.
8627 ("*mov<VM64:mode>_insn_sp32"): Likewise.
8629 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8631 * config/sparc/sparc.md ("type"): New insn type viscmp.
8632 ("fcmp<gcond:code><GCM:gcm_name><P:mode>_vis"): Set insn type to
8634 ("fpcmp<gcond:code>8<P:mode>_vis"): Likewise.
8635 ("fucmp<gcond:code>8<P:mode>_vis"): Likewise.
8636 ("fpcmpu<gcond:code><GCM:gcm_name><P:mode>_vis"): Likewise.
8637 * config/sparc/niagara7.md ("n7_vis_logical_v3pipe"): Handle
8639 ("n7_vis_logical_11cycle"): Likewise.
8640 * config/sparc/niagara4.md ("n4_vis_logical"): Likewise.
8641 * config/sparc/niagara2.md ("niag3_vis": Likewise.
8642 * config/sparc/niagara.md ("niag_vis"): Likewise.
8643 * config/sparc/ultra3.md ("us3_fga"): Likewise.
8644 * config/sparc/ultra1_2.md ("us1_fga_double"): Likewise.
8646 2017-07-07 Jose E. Marchesi <jose.marchesi@oracle.com>
8648 * config/sparc/sparc.md: New instruction type `bmask'.
8649 (bmaskdi_vis): Use the `bmask' type.
8650 (bmasksi_vis): Likewise.
8651 * config/sparc/ultra3.md (us3_array): Likewise.
8652 * config/sparc/niagara7.md (n7_array): Likewise.
8653 * config/sparc/niagara4.md (n4_array): Likewise.
8654 * config/sparc/niagara2.md (niag2_vis): Likewise.
8655 (niag3_vis): Likewise.
8656 * config/sparc/niagara.md (niag_vis): Likewise.
8658 2017-07-06 Jan Hubicka <hubicka@ucw.cz>
8660 * ipa-comdats.c: Remove optimize check from gate.
8661 * ipa-fnsummary.c (ipa_fn_summary_generate): do not generate summary
8662 for functions not optimized.
8663 (ipa_fn_summary_read): Skip optimize check.
8664 (ipa_fn_summary_write): Likewise.
8665 * ipa-inline-analysis.c (do_estimate_growth_1): Check that caller
8667 * ipa-inline.c (can_inline_edge_p): Not optimized functions are
8669 (can_inline_edge_p): Check flag_pcc_struct_return for match.
8670 (check_callers): Give up on caller which is not optimized.
8671 (inline_small_functions): Likewise.
8672 (ipa_inline): Do not give up when not optimizing.
8673 * ipa-visbility.c (function_and_variable_visibility): Do not optimize
8674 away unoptimizes cdtors.
8675 (whole_program_function_and_variable_visibility): Do
8676 ipa_discover_readonly_nonaddressable_vars in LTO mode.
8677 * ipa.c (process_references): Do not check optimize.
8678 (symbol_table::remove_unreachable_nodes): Update optimize check.
8679 (set_writeonly_bit): Update optimize check.
8680 (pass_ipa_cdtor_merge::gate): Do not check optimize.
8681 (pass_ipa_single_use::gate): Remove.
8683 2017-07-06 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
8685 * config/rs6000/rs6000.c (union_defs, union_uses, insn_is_load_p,
8686 insn_is_store_p, insn_is_swap_p, const_load_sequence_p, v2df_reduction_p,
8687 rtx_is_swappable_p, insn_is_swappable_p, chain_contains_only_swaps,
8688 mark_swaps_for_removal, swap_const_vector_halves, adjust_subreg_index,
8689 permute_load, permute_store, adjust_extract, adjust_splat,
8690 adjust_xxpermdi, adjust_concat, adjust_vperm, handle_special_swappables,
8691 replace_swap_with_copy, dump_swap_insn_table,
8692 alignment_with_canonical_addr, alignment_mask, find_alignment_op,
8693 recombine_lvx_pattern, recombine_stvx_pattern,
8694 recombine_lvx_stvx_patterns, rs6000_analyze_swaps,
8695 make_pass_analyze_swaps): Move all code related to p8 swap optimizations
8696 to file rs6000-p8swap.c.
8697 * config/rs6000/rs6000-p8swap.c: New file.
8698 * config/rs6000/t-rs6000: Add rule to build rs6000-p8swap.o.
8699 * config.gcc: Add rs6000-p8swap.o to extra_objs for powerpc*-*-*
8700 and rs6000*-*-* targets.
8702 2017-07-06 David Malcolm <dmalcolm@redhat.com>
8704 * Makefile.in (selftest): Remove dependency on s-selftest-c++.
8706 2017-07-06 Jan Hubicka <hubicka@ucw.cz>
8708 * lto-wrapper.c (merge_and_complain): Do not merge
8709 fexceptions, fnon_call_exceptions, ftrapv, ffp_contract_, fmath_errno,
8710 fsigned_zeros, ftrapping_math, fwrapv.
8711 (append_compiler_options): Do not track these options.
8712 (append_linker_options): Likewie
8714 2017-07-06 Jan Hubicka <hubicka@ucw.cz>
8716 * cgraphunit.c (cgraph_node::finalize_function): When
8717 !flag_toplevel_reorde set no_reorder flag.
8718 (varpool_node::finalize_decl): Likewise.
8719 (symbol_table::compile): Drop no toplevel reorder path.
8721 2017-07-06 Jan Hubicka <hubicka@ucw.cz>
8723 * bb-reorder.c (better_edge_p): Do not build traces across abnormal/eh
8724 edges; zero probability is not better than uninitialized.
8726 2017-07-06 Maxim Ostapenko <m.ostapenko@samsung.com>
8728 * asan.h (asan_sanitize_allocas_p): Declare.
8729 * asan.c (asan_sanitize_allocas_p): New function.
8730 (handle_builtin_stack_restore): Bail out if !asan_sanitize_allocas_p.
8731 (handle_builtin_alloca): Likewise.
8732 * cfgexpand.c (expand_used_vars): Do not add allocas unpoisoning stuff
8733 if !asan_sanitize_allocas_p.
8734 * params.def (asan-instrument-allocas): Add new option.
8735 * params.h (ASAN_PROTECT_ALLOCAS): Define.
8736 * opts.c (common_handle_option): Disable allocas sanitization for
8739 2017-07-06 Maxim Ostapenko <m.ostapenko@samsung.com>
8741 * asan.c: Include gimple-fold.h.
8742 (get_last_alloca_addr): New function.
8743 (handle_builtin_stackrestore): Likewise.
8744 (handle_builtin_alloca): Likewise.
8745 (asan_emit_allocas_unpoison): Likewise.
8746 (get_mem_refs_of_builtin_call): Add new parameter, remove const
8747 quallifier from first paramerer. Handle BUILT_IN_ALLOCA,
8748 BUILT_IN_ALLOCA_WITH_ALIGN and BUILT_IN_STACK_RESTORE builtins.
8749 (instrument_builtin_call): Pass gimple iterator to
8750 get_mem_refs_of_builtin_call.
8751 (last_alloca_addr): New global.
8752 * asan.h (asan_emit_allocas_unpoison): Declare.
8753 * builtins.c (expand_asan_emit_allocas_unpoison): New function.
8754 (expand_builtin): Handle BUILT_IN_ASAN_ALLOCAS_UNPOISON.
8755 * cfgexpand.c (expand_used_vars): Call asan_emit_allocas_unpoison
8756 if function calls alloca.
8757 * gimple-fold.c (replace_call_with_value): Remove static keyword.
8758 * gimple-fold.h (replace_call_with_value): Declare.
8759 * internal-fn.c: Include asan.h.
8760 * sanitizer.def (BUILT_IN_ASAN_ALLOCA_POISON,
8761 BUILT_IN_ASAN_ALLOCAS_UNPOISON): New builtins.
8763 2017-07-06 David Malcolm <dmalcolm@redhat.com>
8765 * Makefile.in (SELFTEST_FLAGS): Drop "-x c", moving it to...
8766 (C_SELFTEST_FLAGS): New.
8767 (CPP_SELFTEST_FLAGS): New.
8768 (SELFTEST_DEPS): New, from deps of s-selftest.
8769 (C_SELFTEST_DEPS): New, from deps of s-selftest.
8770 (CPP_SELFTEST_DEPS): New.
8771 (selftest): Add dependency on s-selftest-c++.
8772 (s-selftest): Rename to...
8773 (s-selftest-c): ...this, moving deps to SELFTEST_DEPS
8774 and C_SELFTEST_DEPS, and using C_SELFTEST_FLAGS rather
8775 than SELFTEST_FLAGS.
8776 (selftest-gdb): Rename to...
8777 (selftest-c-gdb): ...this, using C_SELFTEST_DEPS and
8779 (selftest-gdb): Reintroduce as an alias for selftest-c-gdb.
8780 (selftest-valgrind): Rename to...
8781 (selftest-c-valgrind): ...this, using C_SELFTEST_DEPS and
8783 (selftest-valgrind): Reintroduce as an alias for
8784 selftest-c-valgrind.
8785 (s-selftest-c++): New.
8786 (selftest-c++-gdb): New.
8787 (selftest-c++-valgrind): New.
8789 2017-07-06 Olivier Hainque <hainque@adacore.com>
8791 * gcc.c (process_command): When deciding if undefined variables
8792 should be ignored when processing specs, accept "gcc -v" as well.
8794 2017-07-06 Jan Hubicka <hubicka@ucw.cz>
8796 * auto-profile.c (afdo_set_bb_count, afdo_propagate_edge,
8797 afdo_annotate_cfg): Set counts/probabilities as determined by afdo.
8799 2017-07-06 Thomas Preud'homme <thomas.preudhomme@arm.com>
8801 * config/arm/arm-cpus.in (armv8-r): Add new entry.
8802 * config/arm/arm-isa.h (ISA_ARMv8r): Define macro.
8803 * config/arm/arm-tables.opt: Regenerate.
8804 * config/arm/arm.h (enum base_architecture): Add BASE_ARCH_8R
8806 * doc/invoke.texi: Mention -march=armv8-r and its extensions.
8808 2017-07-06 Carl Love <cel@us.ibm.com>
8810 * ChangeLog: Clean up from mid air collision
8812 2017-07-06 Carl Love <cel@us.ibm.com>
8814 * config/rs6000/rs6000-c.c: Add support for built-in functions
8815 vector signed int vec_subc (vector signed int, vector signed int);
8816 vector signed __int128 vec_subc (vector signed __int128,
8817 vector signed __int128);
8818 vector unsigned __int128 vec_subc (vector unsigned __int128,
8819 vector unsigned __int128);
8820 vector signed int vec_sube (vector signed int, vector signed int,
8822 vector unsigned int vec_sube (vector unsigned int,
8823 vector unsigned int,
8824 vector unsigned int);
8825 vector signed __int128 vec_sube (vector signed __int128,
8826 vector signed __int128,
8827 vector signed__int128);
8828 vector unsigned __int128 vec_sube (vector unsigned __int128,
8829 vector unsigned __int128,
8830 vector unsigned __int128);
8831 vector signed int vec_subec (vector signed int, vector signed int,
8833 vector unsigned int vec_subec (vector unsigned int,
8834 vector unsigned int,
8835 vector unsigned int);
8836 vector signed __int128 vec_subec (vector signed __int128,
8837 vector signed __int128,
8838 vector signed__int128);
8839 vector unsigned __int128 vec_subec (vector unsigned __int128,
8840 vector unsigned __int128,
8841 vector unsigned __int128);
8842 * config/rs6000/rs6000.c (ALTIVEC_BUILTIN_VEC_SUBE,
8843 ALTIVEC_BUILTIN_VEC_SUBEC): Add ef_builtins.
8844 * config/rs6000/rs6000-builtin.def (SUBE, SUBEC): Add
8845 BU_ALTIVEC_OVERLOAD_X definitions.
8846 * config/rs6000/altivec.h (vec_sube, vec_subec): Add builtin defines.
8847 * doc/extend.texi: Update the built-in documentation file for the new
8850 2017-07-06 David Malcolm <dmalcolm@redhat.com>
8853 * diagnostic-show-locus.c (layout::layout): Use start and finish
8854 spelling location for the start and finish of each range.
8855 * genmatch.c (linemap_client_expand_location_to_spelling_point):
8856 Add unused aspect param.
8857 * input.c (expand_location_1): Add "aspect" param, and use it
8858 to access the correct part of the location.
8859 (expand_location): Pass LOCATION_ASPECT_CARET to new param of
8861 (expand_location_to_spelling_point): Likewise.
8862 (linemap_client_expand_location_to_spelling_point): Add "aspect"
8863 param, and pass it to expand_location_1.
8865 2017-07-06 Sebastian Peryt <sebastian.peryt@intel.com>
8867 * config/i386/avx512fintrin.h (_mm_mask_getexp_round_ss,
8868 _mm_maskz_getexp_round_ss, _mm_mask_getexp_round_sd,
8869 _mm_maskz_getexp_round_sd, _mm_mask_getmant_round_sd,
8870 _mm_maskz_getmant_round_sd, _mm_mask_getmant_round_ss,
8871 _mm_maskz_getmant_round_ss, _mm_mask_getexp_ss, _mm_maskz_getexp_ss,
8872 _mm_mask_getexp_sd, _mm_maskz_getexp_sd, _mm_mask_getmant_sd,
8873 _mm_maskz_getmant_sd, _mm_mask_getmant_ss,
8874 _mm_maskz_getmant_ss): New intrinsics.
8875 (__builtin_ia32_getexpss128_mask): Changed to ...
8876 __builtin_ia32_getexpss128_round ... this.
8877 (__builtin_ia32_getexpsd128_mask): Changed to ...
8878 __builtin_ia32_getexpsd128_round ... this.
8879 * config/i386/i386-builtin-types.def
8880 ((V2DF, V2DF, V2DF, INT, V2DF, UQI, INT),
8881 (V4SF, V4SF, V4SF, INT, V4SF, UQI, INT)): New function type aliases.
8882 * config/i386/i386-builtin.def (__builtin_ia32_getexpsd_mask_round,
8883 __builtin_ia32_getexpss_mask_round, __builtin_ia32_getmantsd_mask_round,
8884 __builtin_ia32_getmantss_mask_round): New builtins.
8885 * config/i386/i386.c (V2DF_FTYPE_V2DF_V2DF_INT_V2DF_UQI_INT,
8886 V4SF_FTYPE_V4SF_V4SF_INT_V4SF_UQI_INT): Handle new types.
8887 (CODE_FOR_avx512f_vgetmantv2df_mask_round,
8888 CODE_FOR_avx512f_vgetmantv4sf_mask_round): New cases.
8889 * config/i386/sse.md
8890 (avx512f_sgetexp<mode><round_saeonly_name>): Changed to ...
8891 avx512f_sgetexp<mode><mask_scalar_name>
8892 <round_saeonly_scalar_name> ... this.
8893 (vgetexp<ssescalarmodesuffix>\t{<round_saeonly_op3>%2, %1, %0|
8894 %0, %1, %2<round_saeonly_op3>}): Changed to ...
8895 vgetexp<ssescalarmodesuffix>
8896 \t{<round_saeonly_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|
8897 %0<mask_scalar_operand3>, %1, %2<round_saeonly_scalar_mask_op3>} ... this.
8898 (avx512f_vgetmant<mode><round_saeonly_name>): Changed to ...
8899 avx512f_vgetmant<mode><mask_scalar_name>
8900 <round_saeonly_scalar_name> ... this.
8901 (vgetmant<ssescalarmodesuffix>\t{%3, <round_saeonly_op4>%2, %1, %0|
8902 %0, %1, %2<round_saeonly_op4>, %3}): Changed to ...
8903 vgetmant<ssescalarmodesuffix>
8904 \t{%3, <round_saeonly_scalar_mask_op4>%2, %1, %0<mask_scalar_operand4>|
8905 %0<mask_scalar_operand4>, %1, %2
8906 <round_saeonly_scalar_mask_op4>, %3} ... this.
8907 * config/i386/subst.md (mask_scalar_operand4,
8908 round_saeonly_scalar_mask_operand4, round_saeonly_scalar_mask_op4,
8909 round_saeonly_scalar_nimm_predicate): New subst attributes.
8911 2017-07-06 Julia Koval <julia.koval@intel.com>
8913 * config/i386/i386.c (ix86_erase_embedded_rounding):
8914 Remove code for old rounding pattern.
8916 2017-07-06 Richard Earnshaw <rearnsha@arm.com>
8918 * config/arm/t-arm (GTM_H): Add arm-cpu.h.
8920 2017-07-06 Christophe Lyon <christophe.lyon@linaro.org>
8922 * doc/sourcebuild.texi (Test Directives, Variants of
8923 dg-require-support): Add documentation for dg-require-stack-check.
8925 2017-07-05 Sebastian Peryt <sebastian.peryt@intel.com>
8927 * config/i386/subst.md (mask_scalar, round_scalar,
8928 round_saeonly_scalar): New meta-templates.
8929 (mask_scalar_name, mask_scalar_operand3, round_scalar_name,
8930 round_scalar_mask_operand3, round_scalar_mask_op3,
8931 round_scalar_constraint, round_scalar_prefix, round_saeonly_scalar_name,
8932 round_saeonly_scalar_mask_operand3, round_saeonly_scalar_mask_op3,
8933 round_saeonly_scalar_constraint,
8934 round_saeonly_scalar_prefix): New subst attribute.
8935 * config/i386/sse.md
8936 (<sse>_vm<plusminus_insn><mode>3<mask_name><round_name>): Renamed to ...
8937 <sse>_vm<plusminus_insn><mode>3<mask_scalar_name>
8938 <round_scalar_name> ... this.
8939 (<sse>_vm<multdiv_mnemonic><mode>3<mask_name><round_name>): Renamed to ...
8940 <sse>_vm<multdiv_mnemonic><mode>3<mask_scalar_name>
8941 <round_scalar_name> ... this.
8942 (<sse>_vm<code><mode>3<mask_name><round_saeonly_name>): Renamed to ...
8943 <sse>_vm<code><mode>3<mask_scalar_name>
8944 <round_saeonly_scalar_name> ... this.
8945 (v<plusminus_mnemonic><ssescalarmodesuffix>
8946 \t{<round_mask_op3>%2, %1, %0<mask_operand3>|
8947 %0<mask_operand3>, %1, %<iptr>2<round_mask_op3>}): Changed to ...
8948 v<plusminus_mnemonic><ssescalarmodesuffix>
8949 \t{<round_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|
8950 %0<mask_scalar_operand3>, %1, %<iptr>2<round_scalar_mask_op3>} ... this.
8951 (v<multdiv_mnemonic><ssescalarmodesuffix>
8952 \t{<round_mask_op3>%2, %1, %0<mask_operand3>|
8953 %0<mask_operand3>, %1, %<iptr>2<round_mask_op3>}): Changed to ...
8954 v<multdiv_mnemonic><ssescalarmodesuffix>
8955 \t{<round_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|
8956 %0<mask_scalar_operand3>, %1, %<iptr>2<round_scalar_mask_op3>} ... this.
8957 (v<maxmin_float><ssescalarmodesuffix>
8958 \t{<round_saeonly_mask_op3>%2, %1, %0<mask_operand3>|
8959 %0<mask_operand3>, %1, %<iptr>2<round_saeonly_mask_op3>}): Changed to ...
8960 v<maxmin_float><ssescalarmodesuffix>
8961 \t{<round_saeonly_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|
8962 %0<mask_scalar_operand3>, %1, %<iptr>2
8963 <round_saeonly_scalar_mask_op3>} ... this.
8965 2017-07-05 Richard Earnshaw <rearnsha@arm.com>
8967 * config/arm/arm.c (arm_fixed_condition_code_regs): New function.
8968 (TARGET_FIXED_CONDITION_CODE_REGS): Redefine.
8970 2017-07-05 Richard Sandiford <richard.sandiford@linaro.org>
8971 Alan Hayward <alan.hayward@arm.com>
8972 David Sherwood <david.sherwood@arm.com>
8974 * combine.c (simplify_if_then_else): Remove "enum" before
8976 * compare-elim.c (can_eliminate_compare): Likewise.
8977 * config/aarch64/aarch64-builtins.c (aarch64_simd_builtin_std_type):
8979 (aarch64_lookup_simd_builtin_type): Likewise.
8980 (aarch64_simd_builtin_type): Likewise.
8981 (aarch64_init_simd_builtin_types): Likewise.
8982 (aarch64_simd_expand_args): Likewise.
8983 * config/aarch64/aarch64-protos.h (aarch64_simd_attr_length_rglist):
8985 (aarch64_reverse_mask): Likewise.
8986 (aarch64_simd_emit_reg_reg_move): Likewise.
8987 (aarch64_gen_adjusted_ldpstp): Likewise.
8988 (aarch64_ccmp_mode_to_code): Likewise.
8989 (aarch64_operands_ok_for_ldpstp): Likewise.
8990 (aarch64_operands_adjust_ok_for_ldpstp): Likewise.
8991 * config/aarch64/aarch64.c (aarch64_ira_change_pseudo_allocno_class):
8993 (aarch64_min_divisions_for_recip_mul): Likewise.
8994 (aarch64_reassociation_width): Likewise.
8995 (aarch64_get_condition_code_1): Likewise.
8996 (aarch64_simd_emit_reg_reg_move): Likewise.
8997 (aarch64_simd_attr_length_rglist): Likewise.
8998 (aarch64_reverse_mask): Likewise.
8999 (aarch64_operands_ok_for_ldpstp): Likewise.
9000 (aarch64_operands_adjust_ok_for_ldpstp): Likewise.
9001 (aarch64_gen_adjusted_ldpstp): Likewise.
9002 * config/aarch64/cortex-a57-fma-steering.c (fma_node::rename):
9004 * config/arc/arc.c (legitimate_offset_address_p): Likewise.
9005 * config/arm/arm-builtins.c (arm_simd_builtin_std_type): Likewise.
9006 (arm_lookup_simd_builtin_type): Likewise.
9007 (arm_simd_builtin_type): Likewise.
9008 (arm_init_simd_builtin_types): Likewise.
9009 (arm_expand_builtin_args): Likewise.
9010 * config/arm/arm-protos.h (arm_expand_builtin): Likewise.
9011 * config/ft32/ft32.c (ft32_libcall_value): Likewise.
9012 (ft32_setup_incoming_varargs): Likewise.
9013 (ft32_function_arg): Likewise.
9014 (ft32_function_arg_advance): Likewise.
9015 (ft32_pass_by_reference): Likewise.
9016 (ft32_arg_partial_bytes): Likewise.
9017 (ft32_valid_pointer_mode): Likewise.
9018 (ft32_addr_space_pointer_mode): Likewise.
9019 (ft32_addr_space_legitimate_address_p): Likewise.
9020 * config/i386/i386-protos.h (ix86_operands_ok_for_move_multiple):
9022 * config/i386/i386.c (ix86_setup_incoming_vararg_bounds): Likewise.
9023 (ix86_emit_outlined_ms2sysv_restore): Likewise.
9024 (iamcu_alignment): Likewise.
9025 (canonicalize_vector_int_perm): Likewise.
9026 (ix86_noce_conversion_profitable_p): Likewise.
9027 (ix86_mpx_bound_mode): Likewise.
9028 (ix86_operands_ok_for_move_multiple): Likewise.
9029 * config/microblaze/microblaze-protos.h
9030 (microblaze_expand_conditional_branch_reg): Likewise.
9031 * config/microblaze/microblaze.c
9032 (microblaze_expand_conditional_branch_reg): Likewise.
9033 * config/powerpcspe/powerpcspe.c (rs6000_init_hard_regno_mode_ok):
9035 (rs6000_reassociation_width): Likewise.
9036 (rs6000_invalid_binary_op): Likewise.
9037 (fusion_p9_p): Likewise.
9038 (emit_fusion_p9_load): Likewise.
9039 (emit_fusion_p9_store): Likewise.
9040 * config/riscv/riscv-protos.h (riscv_regno_mode_ok_for_base_p):
9042 (riscv_hard_regno_mode_ok_p): Likewise.
9043 (riscv_address_insns): Likewise.
9044 (riscv_split_symbol): Likewise.
9045 (riscv_legitimize_move): Likewise.
9046 (riscv_function_value): Likewise.
9047 (riscv_hard_regno_nregs): Likewise.
9048 (riscv_expand_builtin): Likewise.
9049 * config/riscv/riscv.c (riscv_build_integer_1): Likewise.
9050 (riscv_build_integer): Likewise.
9051 (riscv_split_integer): Likewise.
9052 (riscv_legitimate_constant_p): Likewise.
9053 (riscv_cannot_force_const_mem): Likewise.
9054 (riscv_regno_mode_ok_for_base_p): Likewise.
9055 (riscv_valid_base_register_p): Likewise.
9056 (riscv_valid_offset_p): Likewise.
9057 (riscv_valid_lo_sum_p): Likewise.
9058 (riscv_classify_address): Likewise.
9059 (riscv_legitimate_address_p): Likewise.
9060 (riscv_address_insns): Likewise.
9061 (riscv_load_store_insns): Likewise.
9062 (riscv_force_binary): Likewise.
9063 (riscv_split_symbol): Likewise.
9064 (riscv_force_address): Likewise.
9065 (riscv_legitimize_address): Likewise.
9066 (riscv_move_integer): Likewise.
9067 (riscv_legitimize_const_move): Likewise.
9068 (riscv_legitimize_move): Likewise.
9069 (riscv_address_cost): Likewise.
9070 (riscv_subword): Likewise.
9071 (riscv_output_move): Likewise.
9072 (riscv_canonicalize_int_order_test): Likewise.
9073 (riscv_emit_int_order_test): Likewise.
9074 (riscv_function_arg_boundary): Likewise.
9075 (riscv_pass_mode_in_fpr_p): Likewise.
9076 (riscv_pass_fpr_single): Likewise.
9077 (riscv_pass_fpr_pair): Likewise.
9078 (riscv_get_arg_info): Likewise.
9079 (riscv_function_arg): Likewise.
9080 (riscv_function_arg_advance): Likewise.
9081 (riscv_arg_partial_bytes): Likewise.
9082 (riscv_function_value): Likewise.
9083 (riscv_pass_by_reference): Likewise.
9084 (riscv_setup_incoming_varargs): Likewise.
9085 (riscv_print_operand): Likewise.
9086 (riscv_elf_select_rtx_section): Likewise.
9087 (riscv_save_restore_reg): Likewise.
9088 (riscv_for_each_saved_reg): Likewise.
9089 (riscv_register_move_cost): Likewise.
9090 (riscv_hard_regno_mode_ok_p): Likewise.
9091 (riscv_hard_regno_nregs): Likewise.
9092 (riscv_class_max_nregs): Likewise.
9093 (riscv_memory_move_cost): Likewise.
9094 * config/rl78/rl78-protos.h (rl78_split_movsi): Likewise.
9095 * config/rl78/rl78.c (rl78_split_movsi): Likewise.
9096 (rl78_addr_space_address_mode): Likewise.
9097 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9099 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Likewise.
9100 (rs6000_reassociation_width): Likewise.
9101 (rs6000_invalid_binary_op): Likewise.
9102 (fusion_p9_p): Likewise.
9103 (emit_fusion_p9_load): Likewise.
9104 (emit_fusion_p9_store): Likewise.
9105 * config/visium/visium-protos.h (prepare_move_operands): Likewise.
9106 (ok_for_simple_move_operands): Likewise.
9107 (ok_for_simple_move_strict_operands): Likewise.
9108 (ok_for_simple_arith_logic_operands): Likewise.
9109 (visium_legitimize_reload_address): Likewise.
9110 (visium_select_cc_mode): Likewise.
9111 (output_cbranch): Likewise.
9112 (visium_split_double_move): Likewise.
9113 (visium_expand_copysign): Likewise.
9114 (visium_expand_int_cstore): Likewise.
9115 (visium_expand_fp_cstore): Likewise.
9116 * config/visium/visium.c (visium_pass_by_reference): Likewise.
9117 (visium_function_arg): Likewise.
9118 (visium_function_arg_advance): Likewise.
9119 (visium_libcall_value): Likewise.
9120 (visium_setup_incoming_varargs): Likewise.
9121 (visium_legitimate_constant_p): Likewise.
9122 (visium_legitimate_address_p): Likewise.
9123 (visium_legitimize_address): Likewise.
9124 (visium_secondary_reload): Likewise.
9125 (visium_register_move_cost): Likewise.
9126 (visium_memory_move_cost): Likewise.
9127 (prepare_move_operands): Likewise.
9128 (ok_for_simple_move_operands): Likewise.
9129 (ok_for_simple_move_strict_operands): Likewise.
9130 (ok_for_simple_arith_logic_operands): Likewise.
9131 (visium_function_value_1): Likewise.
9132 (rtx_ok_for_offset_p): Likewise.
9133 (visium_legitimize_reload_address): Likewise.
9134 (visium_split_double_move): Likewise.
9135 (visium_expand_copysign): Likewise.
9136 (visium_expand_int_cstore): Likewise.
9137 (visium_expand_fp_cstore): Likewise.
9138 (visium_split_cstore): Likewise.
9139 (visium_select_cc_mode): Likewise.
9140 (visium_split_cbranch): Likewise.
9141 (output_cbranch): Likewise.
9142 (visium_print_operand_address): Likewise.
9143 * expmed.c (flip_storage_order): Likewise.
9144 * expmed.h (emit_cstore): Likewise.
9145 (flip_storage_order): Likewise.
9146 * genrecog.c (validate_pattern): Likewise.
9147 * hsa-gen.c (gen_hsa_addr): Likewise.
9148 * internal-fn.c (expand_arith_overflow): Likewise.
9149 * ira-color.c (allocno_copy_cost_saving): Likewise.
9150 * lra-assigns.c (find_hard_regno_for_1): Likewise.
9151 * lra-constraints.c (prohibited_class_reg_set_mode_p): Likewise.
9152 (process_invariant_for_inheritance): Likewise.
9153 * lra-eliminations.c (move_plus_up): Likewise.
9154 * omp-low.c (lower_oacc_reductions): Likewise.
9155 * simplify-rtx.c (simplify_subreg): Likewise.
9156 * target.def (TARGET_SETUP_INCOMING_VARARG_BOUNDS): Likewise.
9157 (TARGET_CHKP_BOUND_MODE): Likewise..
9158 * targhooks.c (default_chkp_bound_mode): Likewise.
9159 (default_setup_incoming_vararg_bounds): Likewise.
9160 * targhooks.h (default_chkp_bound_mode): Likewise.
9161 (default_setup_incoming_vararg_bounds): Likewise.
9162 * tree-ssa-math-opts.c (divmod_candidate_p): Likewise.
9163 * tree-vect-loop.c (calc_vec_perm_mask_for_shift): Likewise.
9164 (have_whole_vector_shift): Likewise.
9165 * tree-vect-stmts.c (vectorizable_load): Likewise.
9166 * doc/tm.texi: Regenerate.
9168 2017-07-05 Georg-Johann Lay <avr@gjlay.de>
9170 Graceful degrade if Binutils PR21472 is not available.
9173 * configure.ac [target=avr]: WARN instead of ERROR if avrxmega3
9174 .rodata in flash test fails.
9175 (HAVE_LD_AVR_AVRXMEGA3_RODATA_IN_FLASH): Define it if test passes.
9176 * confgure: Regenerate.
9177 * config.in: Regenerate.
9178 * config/avr/avr.c (avr_asm_named_section)
9179 [HAVE_LD_AVR_AVRXMEGA3_RODATA_IN_FLASH]: Only trigger
9180 __do_copy_data for stuff in .rodata if flash_pm_offset = 0.
9181 (avr_asm_init_sections): Same.
9183 2017-07-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9185 * config/arm/neon.md (fma<VCVTF:mode>4): Remove %?.
9186 (fma<VH:mode>4_intrinsic): Likewise.
9187 (*fmsub<VCVTF:mode>4): Likewise.
9188 (*fmsub<VH:mode>4_intrinsic): Likewise.
9190 2017-07-05 Georg-Johann Lay <avr@gjlay.de>
9193 * config/avr/avr.c (avr_out_movhi_mr_r_xmega) [CONSTANT_ADDRESS_P]:
9194 Don't depend on "optimize > 0".
9195 (out_movhi_r_mr, out_movqi_mr_r): Same.
9196 (out_movhi_mr_r, out_movqi_r_mr): Same.
9197 (avr_address_cost) [CONSTANT_ADDRESS_P]: Don't depend cost for
9198 io_address_operand on "optimize > 0".
9200 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9202 * tree-loop-distribution.c: Add general explanantion on the pass.
9203 (generate_loops_for_partition): Mark distributed loop.
9204 (pg_add_dependence_edges): New parameter. Handle alias data
9205 dependence specially and record it in the parameter if asked.
9206 (struct pg_vdata, pg_edata, pg_edge_callback_data): New structs.
9207 (init_partition_graph_vertices, add_partition_graph_edge): New.
9208 (pg_skip_alias_edge, free_partition_graph_edata_cb): New.
9209 (free_partition_graph_vdata, build_partition_graph): New.
9210 (sort_partitions_by_post_order, merge_dep_scc_partitions): New.
9211 (pg_collect_alias_ddrs, break_alias_scc_partitions): New.
9212 (data_ref_segment_size, latch_dominated_by_data_ref): New.
9213 (compute_alias_check_pairs, version_loop_by_alias_check): New.
9214 (version_for_distribution_p, finalize_partitions): New.
9215 (distribute_loop): Handle alias data dependence specially. Factor
9216 out loop fusion code as functions and call these functions.
9218 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9220 * tree-loop-distribution.c (classify_partition): New parameter and
9221 better handle reduction statement.
9222 (rdg_build_partitions): Revise comment.
9223 (distribute_loop): Compute statements in all partitions and pass it
9224 to classify_partition.
9226 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9228 * tree-loop-distribution.c (enum partition_type): New.
9229 (struct partition): New field type.
9230 (partition_merge_into): Add parameter. Update partition type.
9231 (data_dep_in_cycle_p, update_type_for_merge): New functions.
9232 (build_rdg_partition_for_vertex): Compute partition type.
9233 (rdg_build_partitions): Dump partition type.
9234 (distribute_loop): Update calls to partition_merge_into.
9236 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9238 * tree-loop-distribution.c (struct ddr_hasher): New.
9239 (ddr_hasher::hash, ::equal, get_data_dependence): New function.
9241 (classify_partition): Call get_data_dependence.
9242 (pg_add_dependence_edges): Ditto.
9243 (distribute_loop): Release data dependence hash table.
9245 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9247 * tree-loop-distribution.c (ref_base_address): Delete.
9248 (similar_memory_accesses): Rename ...
9249 (share_memory_accesses): ... to this. Check if partitions access
9250 the same memory reference.
9251 (distribute_loop): Call share_memory_accesses.
9253 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9255 * tree-loop-distribution.c (struct partition): New field recording
9257 (partition_alloc, partition_free): Init and release data refs.
9258 (partition_merge_into): Merge data refs.
9259 (build_rdg_partition_for_vertex): Collect data refs for partition.
9260 (pg_add_dependence_edges): Change parameters from vector to bitmap.
9262 (distribute_loop): Remve data refs from vertice data of partition
9265 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9267 * tree-loop-distribution.c (params.h): Include header file.
9268 (MAX_DATAREFS_NUM, DR_INDEX): New macro.
9269 (datarefs_vec): New global var.
9270 (create_rdg_vertices): Use datarefs_vec directly.
9271 (free_rdg): Don't free data references.
9272 (build_rdg): Update use. Don't free data references.
9273 (distribute_loop): Compute global variable for data references.
9274 Bail out if there are too many data references.
9276 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9278 * tree-loop-distribution.c (loop_nest): New global var.
9279 (build_rdg): Use loop directly, rather than loop nest.
9280 (pg_add_dependence_edges): Remove loop nest parameter. Use global
9282 (distribute_loop): Compute global variable loop nest. Update use.
9284 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9286 * tree-loop-distribution.c (enum fuse_type, fuse_message): New.
9287 (partition_merge_into): New parameter. Dump reason for fusion.
9288 (distribute_loop): Update use of partition_merge_into.
9290 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9292 * tree-loop-distribution.c (bb_top_order_index): New.
9293 (bb_top_order_index_size, bb_top_order_cmp): New.
9294 (stmts_from_loop): Use topological order.
9295 (pass_loop_distribution::execute): Compute and release topological
9296 order for basic blocks.
9298 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9300 * tree-loop-distribution.c (pass_loop_distribution::execute): Skip
9303 2017-07-05 Bin Cheng <bin.cheng@arm.com>
9305 * cfgloop.h (struct loop): Add comment. New field orig_loop_num.
9306 * cfgloopmanip.c (lv_adjust_loop_entry_edge): Comment change.
9307 * internal-fn.c (expand_LOOP_DIST_ALIAS): New function.
9308 * internal-fn.def (LOOP_DIST_ALIAS): New.
9309 * tree-vectorizer.c (fold_loop_vectorized_call): Rename to ...
9310 (fold_loop_internal_call): ... this.
9311 (vect_loop_dist_alias_call): New function.
9312 (set_uid_loop_bbs): Call fold_loop_internal_call.
9313 (vectorize_loops): Fold IFN_LOOP_VECTORIZED and IFN_LOOP_DIST_ALIAS
9316 2017-07-04 Uros Bizjak <ubizjak@gmail.com>
9319 * config/i386/i386.md (setcc + movzbl/and to xor + setcc peepholes):
9320 Require dead FLAGS_REG at the beginning of a peephole.
9322 2017-07-04 Uros Bizjak <ubizjak@gmail.com>
9325 * config/i386/adxintrin.h (_subborrow_u32): Swap _X and _Y
9326 arguments in the call to __builtin_ia32_sbb_u32.
9327 (_subborrow_u64): Swap _X and _Y arguments in the call to
9328 __builtin_ia32_sbb_u64.
9330 2017-07-04 Jakub Jelinek <jakub@redhat.com>
9333 * tree-vrp.c (compare_assert_loc): Turn into a function template
9334 with stable template parameter. Only test if a->e is NULL,
9335 !a->e == !b->e has been verified already. Use e == NULL or
9336 e != NULL instead of e or ! e tests. If stable is true, don't use
9337 iterative_hash_expr, on the other side allow a or b or both NULL
9338 and sort the NULLs last.
9339 (process_assert_insertions): Sort using compare_assert_loc<false>
9340 instead of compare_assert_loc, later sort using
9341 compare_assert_loc<true> before calling process_assert_insertions_for
9342 in a loop. Use break instead of continue once seen NULL pointer.
9344 2017-07-04 Thomas Preud'homme <thomas.preudhomme@arm.com>
9346 * config/arm/driver-arm.c (arm_cpu_table): Add entry for ARM
9347 Cortex-R7 and Cortex-R8 processors.
9349 2017-07-04 Jan Hubicka <hubicka@ucw.cz>
9351 * ipa-utils.c (ipa_merge_profiles): Fix merging when dst is
9352 uninitialized while src is not.
9354 2017-07-04 Richard Earnshaw <rearnsha@arm.com>
9356 * common/config/arm/arm-common.c: Adjust include path for
9358 * config/arm/t-arm (TM_H): Adjust path for arm-cpu.h.
9359 (arm-cpu.h): Create in build directory. Adjust dependency rules.
9360 (arm-cpu-data.h): Likewise.
9361 (arm-cpu-cdata.h): Likewise.
9362 * config/arm/arm-cpu.h: Delete.
9363 * config/arm/arm-cpu-cdata.h: Delete.
9364 * config/arm/arm-cpu-data.h: Delete.
9366 2017-07-04 James Greenhalgh <james.greenhalgh@arm.com>
9368 * config/arm/arm-cpus.in (cortex-a55): New.
9369 (cortex-a75): Likewise.
9370 (cortex-a75.cortex-a55): Likewise.
9371 * config/arm/driver-arm.c (arm_cpu_table): Add cortex-a55 and
9373 * doc/invoke.texi (-mcpu): Document cortex-a55 and cortex-a75.
9374 * config/arm/arm-cpu-cdata.h: Regenerate.
9375 * config/arm/arm-cpu-data.h: Regenerate.
9376 * config/arm/arm-cpu.h: Regenerate.
9377 * config/arm/arm-tables.opt: Regenerate.
9378 * config/arm/arm-tune.md: Regenerate.
9380 2017-07-04 Jan Hubicka <hubicka@ucw.cz>
9382 * haifa-sched.c (sched_create_recovery_edges): Update profile.
9384 2017-07-04 Jan Hubicka <hubicka@ucw.cz>
9386 * bb-reorder.c (better_edge_p): Fix handling of uninitialized
9389 2017-07-04 Richard Sandiford <richard.sandiford@linaro.org>
9391 PR tree-optimization/81292
9392 * tree-ssa-strlen.c (handle_builtin_strlen): When setting
9393 full_string_p, also call adjust_related_strinfos if the adjustment
9394 is simple, otherwise invalidate related strinfos.
9396 2017-07-04 Martin Liska <mliska@suse.cz>
9399 * sanopt.c (sanitize_rewrite_addressable_params): Mark the
9400 newly created variable as DECL_IGNORED_P.
9402 2017-07-04 Martin Liska <mliska@suse.cz>
9405 * ipa-inline.c (inline_small_functions):
9406 Use xstrdup_for_dump.
9408 2017-07-04 Tom de Vries <tom@codesourcery.com>
9410 * graph.c (draw_cfg_edges): Save and restore EDGE_DFS_BACK.
9412 2017-07-03 Dominique d'Humieres <dominiq@lps.ens.fr>
9415 * config/darwin.c (darwin_function_switched_text_sections):
9418 2017-07-03 Jan Hubicka <hubicka@ucw.cz>
9420 * tree-vect-loop-manip.c (vect_do_peeling): Fix scaling up.
9422 2017-07-03 Richard Earnshaw <rearnsha@arm.com>
9424 * doc/invoke.texi (ARM Options): Add -mbe8 and -mbe32 to option summary.
9426 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9428 * tree-vect-loop.c (vect_analyze_loop_2): Treat min_scalar_loop_bound,
9429 min_profitable_iters, and th as inclusive lower bounds.
9430 Fix LOOP_VINFO_PEELING_FOR_GAPS condition.
9431 (vect_estimate_min_profitable_iters): Return inclusive lower bounds
9432 for min_profitable_iters and min_profitable_estimate.
9433 (vect_transform_loop): Treat th as an inclusive lower bound.
9434 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
9436 2017-07-03 Dominique d'Humieres <dominiq@lps.ens.fr>
9439 * config/darwin.c (darwin_function_switched_text_sections):
9440 Replace DECL_NAME with DECL_ASSEMBLER_NAME, split assemble_name_raw
9441 in two pieces, and suppress the use of buf.
9443 2017-07-03 Nathan Sidwell <nathan@acm.org>
9445 * hash-table.h (hash_table_mod1): Fix indentation.
9447 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9450 * predict.c (force_edge_cold): Be more careful about propagation
9452 * profile-count.h (profile_probability::guessed,
9453 profile_probability::fdo, profile_count::guessed, profile_count::fdo):
9455 * tree-ssa-threadupdate.c (recompute_probabilities): Result is guessed.
9457 2017-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9459 * doc/invoke.texi (rcpc architecture extension): Document it.
9461 2017-07-03 Richard Biener <rguenther@suse.de>
9463 PR tree-optimization/60510
9464 * tree-vect-loop.c (vect_create_epilog_for_reduction): Pass in
9465 the scalar reduction PHI and use it.
9466 (vectorizable_reduction): Properly guard the single_defuse_cycle
9467 path for non-SLP reduction chains where we cannot use it.
9468 Rework reduc_def/index and vector type deduction. Rework
9469 vector operand gathering during reduction op code-gen.
9470 * tree-vect-slp.c (vect_analyze_slp): For failed SLP reduction
9471 chains dissolve the chain and leave it to non-SLP reduction
9474 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9476 * tree-data-ref.h (dr_alignment): Declare.
9477 * tree-data-ref.c (dr_alignment): New function.
9478 * tree-vectorizer.h (dataref_aux): Remove base_element_aligned.
9479 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Don't
9481 * tree-vect-stmts.c (vectorizable_store): Use dr_alignment.
9483 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9485 * tree-data-ref.h (innermost_loop_behavior): Add base_alignment
9486 and base_misalignment fields.
9487 (DR_BASE_ALIGNMENT, DR_BASE_MISALIGNMENT): New macros.
9488 * tree-data-ref.c: Include builtins.h.
9489 (dr_analyze_innermost): Set up the new innmost_loop_behavior fields.
9490 * tree-vectorizer.h (STMT_VINFO_DR_BASE_ALIGNMENT): New macro.
9491 (STMT_VINFO_DR_BASE_MISALIGNMENT): Likewise.
9492 * tree-vect-data-refs.c: Include tree-cfg.h.
9493 (vect_compute_data_ref_alignment): Use the new innermost_loop_behavior
9494 fields instead of calculating an alignment here.
9495 (vect_analyze_data_refs): Use dr_analyze_innermost. Dump the new
9496 innermost_loop_behavior fields.
9498 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9500 * tree-data-ref.h (innermost_loop_behavior): Add a step_alignment
9502 (DR_STEP_ALIGNMENT): New macro.
9503 * tree-vectorizer.h (STMT_VINFO_DR_STEP_ALIGNMENT): Likewise.
9504 * tree-data-ref.c (dr_analyze_innermost): Initalize step_alignment.
9505 (create_data_ref): Print it.
9506 * tree-vect-stmts.c (vectorizable_load): Use the step alignment
9507 to tell whether the step preserves vector (mis)alignment.
9508 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Likewise.
9509 Move the check for an integer step and generalise to all INTEGER_CST.
9510 (vect_analyze_data_refs): Set DR_STEP_ALIGNMENT when setting DR_STEP.
9511 Print the outer step alignment.
9513 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9515 * tree-data-ref.h (innermost_loop_behavior): Replace aligned_to
9516 with offset_alignment.
9517 (DR_ALIGNED_TO): Delete.
9518 (DR_OFFSET_ALIGNMENT): New macro.
9519 * tree-vectorizer.h (STMT_VINFO_DR_ALIGNED_TO): Delete.
9520 (STMT_VINFO_DR_OFFSET_ALIGNMENT): New macro.
9521 * tree-data-ref.c (dr_analyze_innermost): Update after above changes.
9522 (create_data_ref): Likewise.
9523 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Likewise.
9524 (vect_analyze_data_refs): Likewise.
9525 * tree-if-conv.c (if_convertible_loop_p_1): Use memset before
9526 creating dummy innermost behavior.
9528 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9530 * tree-data-ref.h (dr_analyze_innermost): Replace the dr argument
9531 with a "innermost_loop_behavior *" and refeence tree.
9532 * tree-data-ref.c (dr_analyze_innermost): Likewise.
9533 (create_data_ref): Update call accordingly.
9534 * tree-predcom.c (find_looparound_phi): Likewise.
9536 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9538 * tree-vectorizer.h (_stmt_vec_info): Replace individual dr_*
9539 fields with dr_wrt_vec_loop.
9540 (STMT_VINFO_DR_BASE_ADDRESS, STMT_VINFO_DR_INIT, STMT_VINFO_DR_OFFSET)
9541 (STMT_VINFO_DR_STEP, STMT_VINFO_DR_ALIGNED_TO): Update accordingly.
9542 (STMT_VINFO_DR_WRT_VEC_LOOP): New macro.
9543 (vect_dr_behavior): New function.
9544 (vect_create_addr_base_for_vector_ref): Remove loop parameter.
9545 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Use
9546 vect_dr_behavior. Use a step_preserves_misalignment_p boolean to
9547 track whether the step preserves the misalignment.
9548 (vect_create_addr_base_for_vector_ref): Remove loop parameter.
9549 Use vect_dr_behavior.
9550 (vect_setup_realignment): Update call accordingly.
9551 (vect_create_data_ref_ptr): Likewise. Use vect_dr_behavior.
9552 * tree-vect-loop-manip.c (vect_gen_prolog_loop_niters): Update
9553 call to vect_create_addr_base_for_vector_ref.
9554 (vect_create_cond_for_align_checks): Likewise.
9555 * tree-vect-patterns.c (vect_recog_bool_pattern): Copy
9556 STMT_VINFO_DR_WRT_VEC_LOOP as a block.
9557 (vect_recog_mask_conversion_pattern): Likewise.
9558 * tree-vect-stmts.c (compare_step_with_zero): Use vect_dr_behavior.
9559 (new_stmt_vec_info): Remove redundant zeroing.
9561 2017-07-03 Richard Earnshaw <rearnsha@arm.com>
9563 * common/config/arm/arm-common.c (arm_be8_option): New function.
9564 * config/arm/arm-isa.h (isa_feature): Add new feature bit isa_bit_be8.
9565 (ISA_ARMv6): Add isa_bit_be8.
9566 * config/arm/arm.h (arm_be8_option): Add prototype.
9567 (BE8_SPEC_FUNCTION): New define.
9568 (EXTRA_SPEC_FUNCTIONS): Add BE8_SPEC_FUNCTION.
9569 * config/arm/arm.opt (mbig-endian): Mark as Negative of mlittle-endian.
9570 (mlittle-endian): Similarly.
9571 (mbe8, mbe32): New options.
9572 * config/arm/bpabi.h (BE8_LINK_SPEC): Call arm_be8_option.
9573 * doc/invoke.texi (ARM Options): Document -mbe8 and -mbe32.
9575 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9577 * tree-cfg.c (gimple_find_sub_bbs): Fix profile updating.
9579 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9581 * tree-cfgcleanup.c (want_merge_blocks_p): New function.
9582 (cleanup_tree_cfg_bb): Use it.
9583 * profile-count.h (profile_count::of_for_merging, profile_count::merge):
9585 * tree-cfg.c (gimple_merge_blocks): Use profile_count::merge.
9587 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9590 * loop-doloop.c (add_test): Update profile.
9592 2017-07-03 Martin Liska <mliska@suse.cz>
9595 * sanopt.c (rewrite_usage_of_param): New function.
9596 (sanitize_rewrite_addressable_params): Likewise.
9597 (pass_sanopt::execute): Call rewrite_usage_of_param.
9599 2017-07-03 Richard Biener <rguenther@suse.de>
9601 * tree-vect-loop.c (vect_create_epilog_for_reduction): Revert
9602 back to using VIEW_CONVERT_EXPR.
9604 2017-07-03 Martin Liska <mliska@suse.cz>
9607 * doc/extend.texi: Document when a resolver function is
9608 generated for target_clones.
9610 2017-07-03 Martin Liska <mliska@suse.cz>
9612 * asan.c (asan_emit_stack_protection): Unpoison just red zones
9613 and shadow memory of auto variables which are subject of
9614 use-after-scope sanitization.
9615 (asan_expand_mark_ifn): Add do set only when is_poison.
9617 2016-07-03 Richard Biener <rguenther@suse.de>
9619 * tree-vect-loop.c (vect_analyze_loop_operations): Also analyze
9621 (vect_force_simple_reduction): Record reduction def -> phi mapping.
9622 (vectorizable_reduction): Perform reduction PHI creation when
9623 visiting a reduction PHI and adjust and simplify code generation
9624 phase of the reduction op. Cache dts, use fold_binary, not fold_build2.
9625 (vect_transform_loop): Visit reduction PHIs.
9626 * tree-vect-slp.c (vect_get_and_check_slp_defs): Record reduction
9627 defs into the SLP tree.
9628 (vect_build_slp_tree): Reduction defs terminate the recursion.
9629 * tree-vect-stmts.c (vect_get_vec_def_for_operand_1): Allow lookup
9631 (vect_get_vec_defs_for_stmt_copy): Export.
9632 (vect_get_vec_defs): Likewise.
9633 * tree-vectorizer.h (struct _stmt_vec_info): Amend reduc_def
9635 (vect_get_vec_defs_for_stmt_copy): Declare.
9636 (vect_get_vec_defs): Likewise.
9638 2017-07-03 Richard Sandiford <richard.sandiford@linaro.org>
9640 * tree-data-ref.c (dr_analyze_innermost): Replace the "nest"
9641 parameter with a "loop" parameter and use it instead of the
9642 loop containing DR_STMT. Don't check simple_iv when doing
9643 BB analysis. Describe the two analysis modes in the comment.
9645 2017-07-03 Tom de Vries <tom@codesourcery.com>
9647 PR tree-optimization/69468
9648 * tree-ssa-tail-merge.c (ignore_edge_flags): New constant.
9649 (find_same_succ_bb): Handle ignore_edge_flags.
9651 2017-07-03 Tom de Vries <tom@codesourcery.com>
9653 PR tree-optimization/81192
9654 * tree-ssa-tail-merge.c (same_succ_hash): Use bb->loop_father->num in
9656 (same_succ::equal): Don't find bbs to be equal if bb->loop_father
9658 (find_same_succ_bb): Remove obsolete test on bb->loop_father->latch.
9660 2017-07-03 Tom de Vries <tom@codesourcery.com>
9662 PR tree-optimization/81192
9663 * tree-ssa-tail-merge.c (same_succ_flush_bb): Handle
9664 BB_SAME_SUCC (bb) == NULL.
9666 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9668 * cfgrtl.c (rtl_verify_edges): Enable checking of profile_probability
9671 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9673 * dumpfile.c: Include profile-count.h
9674 * tree-cfg.c (gimple_duplicate_sese_tail): Drop UNUSED attributes;
9676 (insert_cond_bb): Update profile.
9677 * tree-cfg.h (insert_cond_bb): Update prototype.
9678 * tree-chkp-opt.c (chkp_optimize_string_function_calls): Update.
9679 * tree-dump.c: Do not include tree-cfg.
9681 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9683 * bb-reorder.c (fix_up_crossing_landing_pad): Update profile.
9685 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9687 * expect.c (dw2_build_landing_pads): Update profile of the landing pad
9690 2017-07-02 Jan Hubicka <hubicka@ucw.cz>
9692 * tree-complex.c (expand_complex_div_wide): update profile.
9694 2017-07-02 Richard Sandiford <richard.sandiford@linaro.org>
9695 Alan Hayward <alan.hayward@arm.com>
9696 David Sherwood <david.sherwood@arm.com>
9698 * Makefile.in (MACHMODE_H): Remove insn-modes.h
9699 (CORETYPES_H): New define.
9700 (MOSTLYCLEANFILES): Add insn-modes-inline.h.
9701 (insn-modes-inline.h, s-modes-inline-h): New rules.
9702 (generated_files): Add insn-modes-inline.h.
9703 (RTL_BASE_H, TREE_CORE_H): Use CORETYPES_H instead of coretypes.h.
9704 (build/gensupport.o, build/ggc-none.o, build/print-rtl.o): Likewise.
9705 (build/read-md.o, build/read-rtl.o, build/rtl.o): Likewise.
9706 (build/vec.o, build/hash-table.o, build/inchash.o): Likewise.
9707 (build/gencondmd.o, build/genattr.o, build/genattr-common.o): Likewise.
9708 (build/genattrtab.o, build/genautomata.o, build/gencheck.o): Likewise.
9709 (build/gencodes.o, build/genconditions.o): Likewise.
9710 (build/genconfig.o, build/genconstants.o, build/genemit.o): Likewise.
9711 (build/genenums.o, build/genextract.o, build/genflags.o): Likewise.
9712 (build/gentarget-def.o, build/genmddeps.o, build/genopinit.o)
9713 (build/genoutput.o, build/genpeep.o, build/genpreds.o): Likewise.
9714 (build/genrecog.o, build/genmddump.o, build/genmatch.o): Likewise.
9715 (build/gencfn-macros.o, build/gcov-iov.o): Likewise.
9716 * coretypes.h: Include everything up to real.h for generators.
9717 Include insn-modes.h first. Include wide-int-print.h after
9718 wide-int.h. Include insn-modes-inline.h and then machmode.h.
9719 * machmode.h: Don't include insn-modes.h here.
9720 * function-tests.c: Remove includes of signop.h, machmode.h,
9721 double-int.h and wide-int.h.
9723 * gcc-rich-location.c: Remove includes of machmode.h, double-int.h
9725 * optc-save-gen.awk: Likewise.
9726 * gencheck.c (BITS_PER_UNIT): Delete dummy definition.
9727 * godump.c: Remove include of wide-int-print.h.
9728 * pretty-print.h: Likewise.
9729 * wide-int-print.cc: Likewise.
9730 * wide-int.cc: Likewise.
9731 * hash-map-tests.c: Remove include of signop.h.
9732 * hash-set-tests.c: Likewise.
9733 * rtl-tests.c: Likewise.
9734 * mkconfig.sh: Remove include of machmode.h.
9735 * genmodes.c (emit_insn_modes_h): Split emission of inline functions
9737 (emit_insn_modes_inline_h): ...this new function. Emit the code
9738 into an insn-modes-inline.h header file, adding appropriate
9739 include guards and end comments.
9740 (emit_insn_modes_c_header): Remove include of machmode.h.
9741 (emit_min_insn_modes_c_header): Include coretypes.h rather than
9743 (main): Handle -i flag and call emit_insn_modes_inline_h when
9746 2017-07-02 Richard Sandiford <richard.sandiford@linaro.org>
9748 * tree-ssa-strlen.c (strinfo): Rename the length field to
9749 nonzero_chars. Add a full_string_p field.
9750 (compare_nonzero_chars, zero_length_string_p): New functions.
9751 (get_addr_stridx): Add an offset_out parameter.
9752 Use compare_nonzero_chars.
9753 (get_stridx): Update accordingly. Use compare_nonzero_chars.
9754 (new_strinfo): Update after above changes to strinfo.
9755 (set_endptr_and_length): Set full_string_p.
9756 (get_string_length): Update after above changes to strinfo.
9757 (unshare_strinfo): Update call to new_strinfo.
9758 (maybe_invalidate): Likewise.
9759 (get_stridx_plus_constant): Change off to unsigned HOST_WIDE_INT.
9760 Use compare_nonzero_chars and zero_string_p. Treat nonzero_chars
9761 as a uhwi instead of an shwi. Update after above changes to
9762 strinfo and new_strinfo.
9763 (zero_length_string): Assert that chainsi contains full strings.
9764 Use zero_length_string_p. Update call to new_strinfo.
9765 (adjust_related_strinfos): Update after above changes to strinfo.
9766 Copy full_string_p from origsi.
9767 (adjust_last_stmt): Use zero_length_string_p.
9768 (handle_builtin_strlen): Update after above changes to strinfo and
9769 new_strinfo. Install the lhs as the string length if the previous
9770 entry didn't describe a full string.
9771 (handle_builtin_strchr): Update after above changes to strinfo
9773 (handle_builtin_strcpy): Likewise.
9774 (handle_builtin_strcat): Likewise.
9775 (handle_builtin_malloc): Likewise.
9776 (handle_pointer_plus): Likewise.
9777 (handle_builtin_memcpy): Likewise. Track nonzero characters
9778 that aren't necessarily followed by a nul terminator.
9779 (handle_char_store): Likewise.
9781 2017-07-02 Richard Sandiford <richard.sandiford@linaro.org>
9783 PR tree-optimization/80769
9784 * tree-ssa-strlen.c (strinfo): Document that "stmt" is also used
9785 for malloc and calloc. Document the new invariant that all related
9786 strinfos have delayed lengths or none do.
9787 (verify_related_strinfos): Move earlier in file.
9788 (set_endptr_and_length): New function, split out from...
9789 (get_string_length): ...here. Also set the lengths of related
9791 (zero_length_string): Assert that chainsi has known (rather than
9793 (adjust_related_strinfos): Likewise.
9795 2017-07-02 Richard Sandiford <richard.sandiford@linaro.org>
9797 PR tree-optimization/81136
9798 * tree-vect-data-refs.c (vect_update_misalignment_for_peel): Only
9799 assert that two references with the same misalignment have the same
9800 compile-time misalignment if those compile-time misalignments
9803 2017-07-01 Andi Kleen <ak@linux.intel.com>
9805 * print-tree.c (print_node): Print all attributes.
9807 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9809 * cfg.c (scale_bbs_frequencies): New function.
9810 * cfg.h (scale_bbs_frequencies): Declare it.
9811 * cfgloopanal.c (single_likely_exit): Cleanup.
9812 * cfgloopmanip.c (scale_loop_frequencies): Take profile_probability
9814 (scale_loop_profile): Likewise.
9815 (loop_version): Likewise.
9816 (create_empty_loop_on_edge): Update.
9817 * cfgloopmanip.h (scale_loop_frequencies, scale_loop_profile,
9818 scale_loop_frequencies, scale_loop_profile, loopify,
9819 loop_version): Update prototypes.
9820 * modulo-sched.c (sms_schedule): Update.
9821 * predict.c (unlikely_executed_edge_p): Also check probability.
9822 (probably_never_executed_edge_p): Fix typo.
9823 * tree-if-conv.c (version_loop_for_if_conversion): Update.
9824 * tree-parloops.c (gen_parallel_loop): Update.
9825 * tree-ssa-loop-ivcanon.c (try_peel_loop): Update.
9826 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Update.
9827 * tree-ssa-loop-split.c (split_loop): Update.
9828 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Update.
9829 * tree-vect-loop-manip.c (vect_do_peeling): Update.
9830 (vect_loop_versioning): Update.
9831 * tree-vect-loop.c (scale_profile_for_vect_loop): Update.
9833 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9835 * trans-mem.c (split_bb_make_tm_edge): Update profile.
9837 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9839 * tree-if-conv.c (combine_blocks): Use make_single_succ_edge
9840 to keep profile consistent.
9842 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9844 * cfgrtl.c (rtl_flow_call_edges_add): Update profile.
9845 * tree-cfg.c (gimple_flow_call_edges_add): Likewise.
9846 * profile-count.h (max_safe_multiplier): Make unsigned.
9847 (profile_count::guessed_zero): New.
9849 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9851 * bb-reorder.c (fix_up_crossing_landing_pad,
9852 fix_crossing_conditional_branches): Use make_single_succ_edge
9853 to keep profile consistent.
9855 2017-07-01 Jan Hubicka <hubicka@ucw.cz>
9857 * tree-vect-loop.c (optimize_mask_stores): Use make_single_succ_edge
9860 2017-07-01 Jakub Jelinek <jakub@redhat.com>
9863 * bb-reorder.c (fix_up_fall_thru_edges): Move variable declarations to
9864 the right scopes, make sure cond_jump isn't preserved between multiple
9865 iterations. Search for fallthru edge whenever there are 3+ edges and
9866 use find_fallthru_edge for it.
9868 2017-06-29 Jan Hubicka <hubicka@ucw.cz>
9870 Patch by Alexander Monakov <amonakov@ispras.ru>
9871 * sel-sched-ir.c (compute_succs_info): Handle uninitialized
9872 probabilities consistently.
9874 2017-06-29 Jan Hubicka <hubicka@ucw.cz>
9876 * pa.c (pa_expand_compare_and_swap_loop): Update call of
9877 emit_cmp_and_jump_insns.
9879 2017-06-29 Jan Hubicka <hubicka@ucw.cz>
9882 * tree-inline.c (expand_call_inline): Combine profile statuses.
9884 2017-06-30 Andrew Pinski <apinski@cavium.com>
9886 * tree-if-conv.c (predicate_scalar_phi): Update new_stmt if
9887 fold_stmt returned true.
9889 2017-06-30 Nathan Sidwell <nathan@acm.org>
9891 * ggc.h (empty_string): Delete.
9892 * cfgexpand.c (expand_asm_stmt): Use plain "".
9893 * optabs.c (expand_asm_memory_barrier): Likewise.
9894 * stringpool.c (empty_string): Delete.
9895 (digit_vector, digit_string): Delete.
9896 (ggc_alloc_string): Use plain "", don't optimize single digit
9897 strings. Use ggc_alloc_atomic.
9899 2017-06-30 Richard Earnshaw <rearnsha@arm.com>
9901 * rtlanal.c (insn_rtx_cost): If a parallel contains exactly one
9902 comparison set and one other set, use the cost of the non-comparison
9905 2017-06-30 Nathan Sidwell <nathan@acm.org>
9907 * ggc.h: Replace all 'static inline' with plain 'inline'. Fix
9910 2017-06-30 Peter Bergner <bergner@vnet.ibm.com>
9912 * tree-cfg.c (group_case_labels_stmt): Merge scanning and compressing
9913 loops. Remove now unneeded calls to gimple_switch_set_label() that
9914 just set removed labels to NULL_TREE.
9916 2017-06-30 Aldy Hernandez <aldyh@redhat.com>
9918 * tree-ssanames.c (set_range_info_raw): Abstract from ...
9919 (set_range_info): ...here. Only call set_range_info_raw if domain
9921 (set_nonzero_bits): Call set_range_info_raw.
9922 * tree-ssanames.h (set_range_info_raw): New.
9924 2017-06-30 Jakub Jelinek <jakub@redhat.com>
9927 * config/i386/sse.md (vec_extract_lo_<mode><mask_name>): For V8FI,
9928 V16FI and VI8F_256 iterators, use <store_mask_predicate> instead
9929 of nonimmediate_operand and <store_mask_constraint> instead of m
9930 for the input operand. For V8FI iterator, always split if input
9931 is a MEM. For V16FI and V8SF_256 iterators, don't test if both
9932 operands are MEM if <mask_applied>. For VI4F_256 iterator, use
9933 <store_mask_predicate> instead of register_operand and
9934 <store_mask_constraint> instead of v for the input operand. Make
9935 sure both operands aren't MEMs for if not <mask_applied>.
9937 2017-06-30 Sylvestre Ledru <sylvestre@debian.org>
9939 * lto-wrapper.c (copy_file) Close both file descriptors before
9942 2017-06-30 Martin Liska <mliska@suse.cz>
9945 * multiple_target.c (create_dispatcher_calls): Make ifunc
9946 also for function that don't have calls or are not referenced.
9948 2017-06-30 Richard Biener <rguenther@suse.de>
9950 * tree-vect-slp.c (vect_slp_analyze_node_operations): Only
9951 analyze the first scalar stmt. Move vector type computation
9952 for the BB case here from ...
9953 * tree-vect-stmts.c (vect_analyze_stmt): ... here. Guard
9954 live operation processing in the SLP case properly.
9956 2017-06-30 Richard Biener <rguenther@suse.de>
9958 * graph.c (draw_cfg_node_succ_edges): Fix broken dot syntax.
9960 2017-06-30 Martin Liska <mliska@suse.cz>
9963 * tree-eh.c (lower_resx): Call BUILT_IN_ASAN_HANDLE_NO_RETURN
9964 before BUILT_IN_UNWIND_RESUME when ASAN is used.
9966 2017-06-30 Yvan Roux <yvan.roux@linaro.org>
9968 * doc/invoke.texi (AArch64): Add missing options and remove redundant
9971 2017-06-30 Richard Biener <rguenther@suse.de>
9973 PR tree-optimization/81249
9974 * tree-vect-loop.c (vect_create_epilog_for_reduction): Convert
9975 condition reduction result to original scalar type.
9977 2017-06-30 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
9979 * profile-count.h (enum profile_quality): Fix typos and whitespace
9982 2017-06-30 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
9984 * config/s390/s390.c (s390_expand_setmem): Adjust to the new data
9985 type for branch probabilities.
9987 2017-06-29 Julian Brown <julian@codesourcery.com>
9988 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
9990 * config/aarch64/aarch64-fusion-pairs.def: Add ALU_BRANCH entry.
9991 * config/aarch64/aarch64.c (AARCH64_FUSE_ALU_BRANCH): New fusion type.
9992 (thunderx2t99_tunings): Set AARCH64_FUSE_ALU_BRANCH flag.
9993 (aarch_macro_fusion_pair_p): Add support for AARCH64_FUSE_ALU_BRANCH.
9995 2017-06-29 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
9997 * config/aarch64/aarch64.c (aarch_macro_fusion_pair_p): Push the
9998 check for CC usage into AARCH64_FUSE_CMP_BRANCH.
9999 * config/i386/i386.c (ix86_macro_fusion_pair_p): Push the check for
10000 CC usage from generic code to here.
10001 * sched-deps.c (sched_macro_fuse_insns): Move the condition for
10002 CC usage into the target macros.
10004 2017-06-29 Maya Rashish <coypu@sdf.org>
10006 * config/netbsd.h (NETBSD_LIB_SPEC): Add -lc when creating shared
10009 2017-06-29 Jan Hubicka <hubicka@ucw.cz>
10011 * arm/arm-builtins.c: Include profile-count.h
10012 * except.c (sjlj_emit_function_enter): Use
10013 profile_probability::unlikely.
10015 2017-06-29 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
10017 * config/rs6000/rs6000.c (toc_relative_expr_p): Make tocrel_base
10018 and tocrel_offset be pointer args rather than implicitly using
10020 (legitimate_constant_pool_address_p, rs6000_emit_move,
10021 const_load_sequence_p, adjust_vperm): Add local tocrel_base and
10022 tocrel_offset and use in toc_relative_expr_p call.
10023 (print_operand, print_operand_address): Use static tocrel_base_oac
10024 and tocrel_offset_oac.
10025 (rs6000_output_addr_const_extra): Use static tocrel_base_oac and
10028 2017-06-29 Maya Rashish <coypu@sdf.org>
10030 * config/vax/builtins.md (ffssi2_internal): Correct constraint.
10032 2017-06-29 Eric Botcazou <ebotcazou@adacore.com>
10034 * expr.c (expand_expr) <normal_inner_ref>: When testing for unaligned
10035 objects, take into account only the alignment of 'op0' and 'mode1' if
10038 2017-06-29 Steve Ellcey <sellcey@cavium.com>
10040 * ccmp.c (ccmp_tree_comparison_p): New function.
10041 (ccmp_candidate_p): Update to use above function.
10042 (get_compare_parts): New function.
10043 (expand_ccmp_next): Update to use new functions.
10044 (expand_ccmp_expr_1): Take tree arg instead of gimple, update to use
10046 (expand_ccmp_expr): Pass tree instead of gimple to expand_ccmp_expr_1,
10047 take mode as argument.
10048 * ccmp.h (expand_ccmp_expr): Add mode as argument.
10049 * expr.c (expand_expr_real_1): Pass mode as argument.
10051 2017-06-29 Segher Boessenkool <segher@kernel.crashing.org>
10053 * combine.c (combine_instructions): Print insns to dump_file, together
10056 2017-06-29 Jan Hubicka <hubicka@ucw.cz>
10058 * asan.c (asan_emit_stack_protection): Update.
10059 (create_cond_insert_point): Update.
10060 * auto-profile.c (afdo_propagate_circuit): Update.
10061 * basic-block.h (struct edge_def): Turn probability to
10062 profile_probability.
10063 (EDGE_FREQUENCY): Update.
10064 * bb-reorder.c (find_traces_1_round): Update.
10065 (better_edge_p): Update.
10066 (sanitize_hot_paths): Update.
10067 * cfg.c (unchecked_make_edge): Initialize probability to uninitialized.
10068 (make_single_succ_edge): Update.
10069 (check_bb_profile): Update.
10070 (dump_edge_info): Update.
10071 (update_bb_profile_for_threading): Update.
10072 * cfganal.c (connect_infinite_loops_to_exit): Initialize new edge
10073 probabilitycount to 0.
10074 * cfgbuild.c (compute_outgoing_frequencies): Update.
10075 * cfgcleanup.c (try_forward_edges): Update.
10076 (outgoing_edges_match): Update.
10077 (try_crossjump_to_edge): Update.
10078 * cfgexpand.c (expand_gimple_cond): Update make_single_succ_edge.
10079 (expand_gimple_tailcall): Update.
10080 (construct_init_block): Use make_single_succ_edge.
10081 (construct_exit_block): Use make_single_succ_edge.
10082 * cfghooks.c (verify_flow_info): Update.
10083 (redirect_edge_succ_nodup): Update.
10084 (split_edge): Update.
10085 (account_profile_record): Update.
10086 * cfgloopanal.c (single_likely_exit): Update.
10087 * cfgloopmanip.c (scale_loop_profile): Update.
10088 (set_zero_probability): Remove.
10089 (duplicate_loop_to_header_edge): Update.
10090 * cfgloopmanip.h (loop_version): Update prototype.
10091 * cfgrtl.c (try_redirect_by_replacing_jump): Update.
10092 (force_nonfallthru_and_redirect): Update.
10093 (update_br_prob_note): Update.
10094 (rtl_verify_edges): Update.
10095 (purge_dead_edges): Update.
10096 (rtl_lv_add_condition_to_bb): Update.
10097 * cgraph.c: (cgraph_edge::redirect_call_stmt_to_calle): Update.
10098 * cgraphunit.c (init_lowered_empty_function): Update.
10099 (cgraph_node::expand_thunk): Update.
10100 * cilk-common.c: Include profile-count.h
10101 * dojump.c (inv): Remove.
10102 (jumpifnot): Update.
10103 (jumpifnot_1): Update.
10104 (do_jump_1): Update.
10106 (do_jump_by_parts_greater_rtx): Update.
10107 (do_compare_rtx_and_jump): Update.
10108 * dojump.h (jumpifnot, jumpifnot_1, jumpif_1, jumpif, do_jump,
10109 do_jump_1. do_compare_rtx_and_jump): Update prototype.
10110 * dwarf2cfi.c: Include profile-count.h
10111 * except.c (dw2_build_landing_pads): Use make_single_succ_edge.
10112 (sjlj_emit_dispatch_table): Likewise.
10113 * explow.c: Include profile-count.h
10114 * expmed.c (emit_store_flag_force): Update.
10115 (do_cmp_and_jump): Update.
10116 * expr.c (compare_by_pieces_d::generate): Update.
10117 (compare_by_pieces_d::finish_mode): Update.
10118 (emit_block_move_via_loop): Update.
10119 (store_expr_with_bounds): Update.
10120 (store_constructor): Update.
10121 (expand_expr_real_2): Update.
10122 (expand_expr_real_1): Update.
10123 * expr.h (try_casesi, try_tablejump): Update prototypes.
10124 * gimple-pretty-print.c (dump_probability): Update.
10125 (dump_profile): New.
10126 (dump_gimple_label): Update.
10127 (dump_gimple_bb_header): Update.
10128 * graph.c (draw_cfg_node_succ_edges): Update.
10129 * hsa-gen.c (convert_switch_statements): Update.
10130 * ifcvt.c (cheap_bb_rtx_cost_p): Update.
10131 (find_if_case_1): Update.
10132 (find_if_case_2): Update.
10133 * internal-fn.c (expand_arith_overflow_result_store): Update.
10134 (expand_addsub_overflow): Update.
10135 (expand_neg_overflow): Update.
10136 (expand_mul_overflow): Update.
10137 (expand_vector_ubsan_overflow): Update.
10138 * ipa-cp.c (good_cloning_opportunity_p): Update.
10139 * ipa-split.c (split_function): Use make_single_succ_edge.
10140 * ipa-utils.c (ipa_merge_profiles): Update.
10141 * loop-doloop.c (add_test): Update.
10142 (doloop_modify): Update.
10143 * loop-unroll.c (compare_and_jump_seq): Update.
10144 (unroll_loop_runtime_iterations): Update.
10145 * lra-constraints.c (lra_inheritance): Update.
10146 * lto-streamer-in.c (input_cfg): Update.
10147 * lto-streamer-out.c (output_cfg): Update.
10148 * mcf.c (adjust_cfg_counts): Update.
10149 * modulo-sched.c (sms_schedule): Update.
10150 * omp-expand.c (expand_omp_for_init_counts): Update.
10151 (extract_omp_for_update_vars): Update.
10152 (expand_omp_ordered_sink): Update.
10153 (expand_omp_for_ordered_loops): Update.
10154 (expand_omp_for_generic): Update.
10155 (expand_omp_for_static_nochunk): Update.
10156 (expand_omp_for_static_chunk): Update.
10157 (expand_cilk_for): Update.
10158 (expand_omp_simd): Update.
10159 (expand_omp_taskloop_for_outer): Update.
10160 (expand_omp_taskloop_for_inner): Update.
10161 * omp-simd-clone.c (simd_clone_adjust): Update.
10162 * optabs.c (expand_doubleword_shift): Update.
10163 (expand_abs): Update.
10164 (emit_cmp_and_jump_insn_1): Update.
10165 (expand_compare_and_swap_loop): Update.
10166 * optabs.h (emit_cmp_and_jump_insns): Update prototype.
10167 * predict.c (predictable_edge_p): Update.
10168 (edge_probability_reliable_p): Update.
10169 (set_even_probabilities): Update.
10170 (combine_predictions_for_insn): Update.
10171 (combine_predictions_for_bb): Update.
10172 (propagate_freq): Update.
10173 (estimate_bb_frequencies): Update.
10174 (force_edge_cold): Update.
10175 * profile-count.c (profile_count::dump): Add missing space into dump.
10176 (profile_count::debug): Add newline.
10177 (profile_count::differs_from_p): Explicitly convert to unsigned.
10178 (profile_count::stream_in): Update.
10179 (profile_probability::dump): New member function.
10180 (profile_probability::debug): New member function.
10181 (profile_probability::differs_from_p): New member function.
10182 (profile_probability::differs_lot_from_p): New member function.
10183 (profile_probability::stream_in): New member function.
10184 (profile_probability::stream_out): New member function.
10185 * profile-count.h (profile_count_quality): Rename to ...
10186 (profile_quality): ... this one.
10187 (profile_probability): New.
10188 (profile_count): Update.
10189 * profile.c (compute_branch_probabilities): Update.
10190 * recog.c (peep2_attempt): Update.
10191 * sched-ebb.c (schedule_ebbs): Update.
10192 * sched-rgn.c (find_single_block_region): Update.
10193 (compute_dom_prob_ps): Update.
10194 (schedule_region): Update.
10195 * sel-sched-ir.c (compute_succs_info): Update.
10196 * stmt.c (struct case_node): Update.
10197 (do_jump_if_equal): Update.
10198 (get_outgoing_edge_probs): Update.
10199 (conditional_probability): Update.
10200 (emit_case_dispatch_table): Update.
10201 (expand_case): Update.
10202 (expand_sjlj_dispatch_table): Update.
10203 (emit_case_nodes): Update.
10204 * targhooks.c: Update.
10205 * tracer.c (better_p): Update.
10206 (find_best_successor): Update.
10207 * trans-mem.c (expand_transaction): Update.
10208 * tree-call-cdce.c: Update.
10209 * tree-cfg.c (gimple_split_edge): Upate.
10210 (move_sese_region_to_fn): Upate.
10211 * tree-cfgcleanup.c (cleanup_control_expr_graph): Upate.
10212 * tree-eh.c (lower_resx): Upate.
10213 (cleanup_empty_eh_move_lp): Upate.
10214 * tree-if-conv.c (version_loop_for_if_conversion): Update.
10215 * tree-inline.c (copy_edges_for_bb): Update.
10216 (copy_cfg_body): Update.
10217 * tree-parloops.c (gen_parallel_loop): Update.
10218 * tree-profile.c (gimple_gen_ic_func_profiler): Update.
10219 (gimple_gen_time_profiler): Update.
10220 * tree-ssa-dce.c (remove_dead_stmt): Update.
10221 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): Update.
10222 * tree-ssa-loop-im.c (execute_sm_if_changed): Update.
10223 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Update.
10224 (unloop_loops): Update.
10225 (try_peel_loop): Update.
10226 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Update.
10227 * tree-ssa-loop-split.c (connect_loops): Update.
10228 (split_loop): Update.
10229 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Update.
10230 (hoist_guard): Update.
10231 * tree-ssa-phionlycprop.c (propagate_rhs_into_lhs): Update.
10232 * tree-ssa-phiopt.c (replace_phi_edge_with_variable): Update.
10233 (value_replacement): Update.
10234 * tree-ssa-reassoc.c (branch_fixup): Update.
10235 * tree-ssa-tail-merge.c (replace_block_by): Update.
10236 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges): Update.
10237 (create_edge_and_update_destination_phis): Update.
10238 (compute_path_counts): Update.
10239 (recompute_probabilities): Update.
10240 (update_joiner_offpath_counts): Update.
10241 (freqs_to_counts_path): Update.
10242 (duplicate_thread_path): Update.
10243 * tree-switch-conversion.c (hoist_edge_and_branch_if_true): Update.
10244 (struct switch_conv_info): Update.
10245 (gen_inbound_check): Update.
10246 * tree-vect-loop-manip.c (slpeel_add_loop_guard): Update.
10247 (vect_do_peeling): Update.
10248 (vect_loop_versioning): Update.
10249 * tree-vect-loop.c (scale_profile_for_vect_loop): Update.
10250 (optimize_mask_stores): Update.
10251 * ubsan.c (ubsan_expand_null_ifn): Update.
10252 * value-prof.c (gimple_divmod_fixed_value): Update.
10253 (gimple_divmod_fixed_value_transform): Update.
10254 (gimple_mod_pow2): Update.
10255 (gimple_mod_pow2_value_transform): Update.
10256 (gimple_mod_subtract): Update.
10257 (gimple_mod_subtract_transform): Update.
10258 (gimple_ic): Update.
10259 (gimple_stringop_fixed_value): Update.
10260 (gimple_stringops_transform): Update.
10261 * value-prof.h: Update.
10263 2017-06-29 Carl Love <cel@us.ibm.com>
10265 * config/rs6000/rs6000-c.c: Add support for built-in functions
10266 vector signed int vec_signed (vector float);
10267 vector signed long long vec_signed (vector double);
10268 vector signed int vec_signed2 (vector double, vector double);
10269 vector signed int vec_signede (vector double);
10270 vector signed int vec_signedo (vector double);
10271 * config/rs6000/rs6000.c (rs6000_generate_vsigned2_code): Add
10272 instruction generator.
10273 * config/rs6000/vsx.md (UNSPEC_VSX_XVCVSPSXWS, UNSPEC_VSX_XVCVSPSXDS,
10274 UNSPEC_VSX_VSIGNED2): Add UNSPECS.
10275 (vsx_xvcvspsxws, vsx_xvcvdpuxds_scale, vsx_xvcvspuxws, vsigned2_v2df):
10277 (vsignedo_v2df, vsignede_v2df, vunsigned2_v2df, vunsignedo_v2df,
10278 vunsignede_v2df): Add define_expands.
10279 * config/rs6000/rs6000-builtin.def (VEC_SIGNED, VEC_UNSIGNED,
10280 VEC_SIGNED2, VEC_UNSIGNED2, VEC_SIGNEDE, VEC_UNSIGNEDE, VEC_SIGNEDO,
10281 VEC_UNSIGNEDO): Add definitions.
10282 * config/vsx.md (UNSPEC_VSX_XVCVSPSXWS, UNSPEC_VSX_XVCVSPSXDS,
10283 UNSPEC_VSX_VSIGNED2): Add UNSPECs.
10284 (vsx_xvcvspsxws, vsx_xvcvspuxws): Add define_insn.
10285 (vsigned2_v2df, vsigendo_v2df, vsignede_v2df,
10286 vunsigned2_v2df, vunsignedo_v2df, vunsignede_v2df): Add define_expands.
10287 * config/rs6000/altivec.h (vec_signed, vec_signed2,
10288 vec_signede and vec_signedo, vec_unsigned, vec_unsigned2,
10289 vec_unsignede, vec_unsignedo): Add builtin defines.
10290 * config/rs6000-protos.h (rs6000_generate_vsigned2_code): Add extern
10292 * doc/extend.texi: Update the built-in documentation file for the
10293 new built-in functions.
10295 2017-06-29 Richard Biener <rguenther@suse.de>
10297 * tree-vect-loop.c (vect_analyze_scalar_cycles_1): Do not add
10298 reduction chains to LOOP_VINFO_REDUCTIONS.
10299 * tree-vect-slp.c (vect_analyze_slp): Continue looking for
10300 SLP reductions after processing reduction chains.
10302 2017-06-29 Nathan Sidwell <nathan@acm.org>
10304 * builtins.c (fold_builtin_FUNCTION): Use
10305 lang_hooks.decl_printable_name.
10307 2017-06-29 Peter Bergner <bergner@vnet.ibm.com>
10309 PR middle-end/81194
10310 * cfgexpand.c (expand_gimple_stmt_1): Handle switch statements
10311 with only one label.
10312 * stmt.c (expand_case): Assert NCASES is greater than one.
10314 2017-06-29 Richard Biener <rguenther@suse.de>
10316 * tree-cfg.c (group_case_labels_stmt): Return whether we changed
10318 (group_case_labels): Likewise.
10319 (find_taken_edge): Push sanity checking on val to workers...
10320 (find_taken_edge_cond_expr): ... here
10321 (find_taken_edge_switch_expr): ... and here, handle cases
10322 with just a default label.
10323 * tree-cfg.h (group_case_labels_stmt): Adjust prototype.
10324 (group_case_labels): Likewise.
10325 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing): When
10326 group_case_labels does anything cleanup the CFG again.
10328 2017-06-29 Bin Cheng <bin.cheng@arm.com>
10330 PR tree-optimization/81196
10331 * tree-ssa-loop-niter.c (number_of_iterations_cond): Handle loop
10332 exit condition comparing two IVs.
10334 2017-06-29 Richard Earnshaw <rearnsha@arm.com>
10336 * config/arm/parsecpu.awk (gen_comm_data): Add initializer for
10337 profile to the dummy entry at the end of the list of architectures.
10338 * config/arm/arm-cpu-cdata.h: Regenerated.
10340 2017-06-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10341 Michael Collison <michael.collison@arm.com>
10344 * config/aarch64/aarch64.md (*aarch64_<optab>_reg_<mode>3_mask1):
10346 (*aarch64_reg_<mode>3_neg_mask2): New pattern.
10347 (*aarch64_reg_<mode>3_minus_mask): New pattern.
10348 (*aarch64_<optab>_reg_di3_mask2): New pattern.
10349 * config/aarch64/aarch64.c (aarch64_rtx_costs): Account for cost
10350 of shift when the shift amount is masked with constant equal to
10351 the size of the mode.
10352 * config/aarch64/predicates.md (subreg_lowpart_operator): New
10355 2017-06-29 Martin Liska <mliska@suse.cz>
10357 * config/i386/i386.opt: Change range from [1,5] to [0,5].
10359 2017-06-29 Yury Gribov <tetra2005@gmail.com>
10362 * ipa-cp.c (allocate_and_init_ipcp_value): Add initialization
10365 (edge_growth_cache_entry::edge_growth_cache_entry): New
10367 (reset_edge_growth_cache): Update to use constructor.
10369 2017-06-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10371 * config/aarch64/aarch64.h (AARCH64_EXPAND_ALIGNMENT): New.
10372 (DATA_ALIGNMENT): Update to use AARCH64_EXPAND_ALIGNMENT.
10373 (LOCAL_ALIGNMENT): Update to use AARCH64_EXPAND_ALIGNMENT.
10375 2017-06-28 Sebastian Peryt <sebastian.peryt@intel.com>
10377 * config/i386/avx512vlintrin.h (_mm256_permutexvar_epi64)
10378 (_mm256_permutexvar_epi32, _mm256_permutex_epi64): New intrinsics.
10380 2017-06-28 Szabolcs Nagy <szabolcs.nagy@arm.com>
10382 * config.gcc (*-linux-musl*): Add t-musl tmake_file.
10383 (*-linux-uclibc*): Add t-uclibc tmake_file.
10384 * config/t-musl: New.
10385 * config/t-uclibc: New.
10387 2017-06-28 Richard Earnshaw <rearnsha@arm.com>
10389 * config/arm/parsecpu.awk (profile): Parse new keyword in an arch
10391 (gen_comm_data): Emit architectural setting of arch_prof.
10392 * config/arm/arm-cpus.in (armv6-m, armv6s-m, armv7-a, armv7ve): Set the
10394 (armv7-r, armv7-m, armv7e-m, armv8-a, armv8.1-a, armv8.2-a): Likewise.
10395 (armv8-m.base, armv8-m.main): Likewise.
10396 * arm-protos.h (arm_build_target): Add profile field.
10397 (arch_option): Likewise.
10398 * config/arm/arm.c (arm_configure_build_target): Copy the profile to
10400 * config/arm/arm.h (TARGET_ARM_ARCH_PROFILE): Use
10401 arm_active_target.profile.
10403 2017-06-28 Richard Biener <rguenther@suse.de>
10405 PR middle-end/81227
10406 * fold-const.c (negate_expr_p): Use TYPE_UNSIGNED, not
10407 TYPE_OVERFLOW_WRAPS.
10408 * match.pd (negate_expr_p): Likewise.
10409 * tree-ssa-reassoc.c (optimize_range_tests_diff): Use
10410 fold_build2, not fold_binary.
10412 2017-06-28 Wilco Dijkstra <wdijkstr@arm.com>
10414 * config/aarch64/aarch64 (aarch64_expand_mov_immediate):
10415 Convert memory address to Pmode.
10416 (aarch64_print_operand): Assert MEM operands are always Pmode.
10418 2017-06-28 Wilco Dijkstra <wdijkstr@arm.com>
10421 * config/arm/aarch-common.c (arm_no_early_alu_shift_dep):
10422 Remove redundant if.
10423 (aarch_forward_to_shift_is_not_shifted_reg): Remove.
10424 * config/arm/aarch-common-protos.h
10425 (aarch_forward_to_shift_is_not_shifted_re): Remove.
10426 * config/arm/cortex-a53.md: Use arm_no_early_alu_shift_dep in bypass.
10428 2017-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10431 * multiple_target.c (create_dispatcher_calls): Set the default
10432 clone to be static, not public.
10434 2017-06-28 Richard Biener <rguenther@suse.de>
10436 * tree-vect-loop.c (vectorizable_reduction): Move special
10437 cond reduction IV var creation ...
10438 (vect_create_epilog_for_reduction): ... here. Remove induction_index
10439 parameter. Use STMT_VINFO_VECTYPE.
10440 * tree-vect-slp.c (vect_get_constant_vectors): Properly reset
10443 2017-06-28 Martin Liska <mliska@suse.cz>
10446 * ipa-visibility.c (non_local_p): Handle visibility.
10448 2017-06-28 Martin Liska <mliska@suse.cz>
10451 * common.opt: Add IntegerRange to various options.
10452 * opt-functions.awk (integer_range_info): New function.
10453 * optc-gen.awk: Add integer_range_info to cl_options struct.
10454 * opts-common.c (decode_cmdline_option): Handle
10455 CL_ERR_INT_RANGE_ARG.
10456 (cmdline_handle_error): Likewise.
10457 * opts.c (print_filtered_help): Show valid interval in
10458 when --help is provided.
10459 * opts.h (struct cl_option): Add range_min and range_max fields.
10460 * config/i386/i386.opt: Add IntegerRange for -mbranch-cost.
10462 2017-06-28 Marc Glisse <marc.glisse@inria.fr>
10464 * match.pd ((X & ~Y) | (~X & Y)): Generalize to + and ^.
10465 (x * C EQ/NE y * C): New transformation.
10467 2017-06-28 Christophe Lyon <christophe.lyon@linaro.org>
10469 * genmultilib (combination_space): Accept '+' in option names.
10471 2017-06-28 Martin Liska <mliska@suse.cz>
10474 * asan.c (instrument_derefs): Bail out inner references
10475 that are hard register variables.
10477 2017-06-28 Jakub Jelinek <jakub@redhat.com>
10480 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Use def_builtin
10481 rather than def_builtin_pure for __builtin_ia32_gatherpf*.
10483 2017-06-28 Richard Biener <rguenther@suse.de>
10485 * tree-vectorizer.h (vect_get_vec_defs): Remove.
10486 (vect_get_slp_defs): Adjust.
10487 * tree-vect-loop.c (get_initial_defs_for_reduction): Split
10489 * tree-vect-slp.c (vect_get_constant_vectors): ... here and
10491 * tree-vect-loop.c (vect_create_epilog_for_reduction): Use
10492 get_initial_defs_for_reduction instead of vect_get_vec_defs.
10493 (vectorizable_reduction): Adjust.
10494 * tree-vect-slp.c (vect_get_constant_vectors): Remove reduction
10496 (vect_get_slp_defs): Likewise.
10497 * tree-vect-stmts.c (vect_get_vec_defs): Make static and adjust.
10498 (vectorizable_bswap): Adjust.
10499 (vectorizable_call): Likewise.
10500 (vectorizable_conversion): Likewise.
10501 (vectorizable_assignment): Likewise.
10502 (vectorizable_shift): Likewise.
10503 (vectorizable_operation): Likewise.
10504 (vectorizable_store): Likewise.
10505 (vectorizable_condition): Likewise.
10506 (vectorizable_comparison): Likewise.
10508 2017-06-28 Michael Collison <michael.collison@arm.com>
10511 * config/arm/arm.c (gen_ldm_seq): Remove last unnecessary
10513 (arm_gen_movmemqi): Removed unused variable 'i'.
10514 Convert 'for' loop into 'while' loop.
10515 (arm_expand_prologue): Remove last unnecessary set of insn.
10516 (thumb_pop): Remove unused variable 'pushed_words'.
10517 (thumb_exit): Remove last unnecessary set of regs_to_pop.
10519 2017-06-28 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
10521 * config/s390/predicates.md: Use s390_rel_address_ok_p.
10522 * config/s390/s390-protos.h: Add prototype of
10523 s390_rel_address_ok_p.
10524 * config/s390/s390.c (s390_got_symbol): New function.
10525 (s390_rel_address_ok_p): New function.
10526 (legitimize_pic_address): Use s390_rel_address_ok_p.
10527 (s390_load_got): Use s390_got_symbol.
10528 (s390_option_override): Issue error if
10529 -mno-pic-data-is-text-relative is used without -fpic/-fPIC.
10530 * config/s390/s390.h (TARGET_DEFAULT_PIC_DATA_IS_TEXT_RELATIVE):
10532 * config/s390/s390.opt: New option mpic-data-is-text-relative.
10534 2017-06-27 Andrew Pinski <apinski@cavium.com>
10536 * match.pd (X >/>=/</<= 0 ? 1.0 : -1.0): New patterns.
10537 (X * copysign (1.0, X)): New pattern.
10538 (X * copysign (1.0, -X)): New pattern.
10539 (copysign (-1.0, CST)): New pattern.
10541 2017-06-27 Joseph Myers <joseph@codesourcery.com>
10543 * genmultilib (combination_space): Remove variable.
10544 Validate reuse rules against regular expression for any sequence
10545 of multilib options in any order.
10547 2017-06-27 Michael Collison <michael.collison@arm.com>
10549 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Directly
10550 call aarch64_split_simd_combine.
10551 * (aarch64_combine_internal<mode>): Delete pattern.
10552 * config/aarch64/aarch64.c (aarch64_split_simd_combine):
10553 Allow register and subreg operands.
10555 2017-06-27 Jerome Lambourg <lambourg@adacore.com>
10557 * config/i386/vxworks.h (ASM_SPEC): Remove definition. No target
10558 specific need, just fallback on defaults.
10559 (ASM_OUTPUT_ALIGNED_BSS): Add #undef before #define.
10561 2017-06-27 Jerome Lambourg <lambourg@adacore.com>
10562 Olivier Hainque <hainque@adacore.com>
10564 * config/i386/vxworks.h (DBX_REGISTER_NUMBER): Pick distinct
10566 (TARGET_OS_CPP_BUILTINS): builtin_define CPU to X86_64 for 64bit
10567 targets. Pick a default if no particular attempt applied.
10568 (STACK_CHECK_PROTECT): Double for 64bit targets, which have
10571 2017-06-27 Jerome Lambourg <lambourg@adacore.com>
10573 * config.gcc (i*86-wrs-vxworks7): Handle new acceptable triplet.
10574 (x86_64-wrs-vxworks7): Likewise.
10576 2017-06-27 Marek Polacek <polacek@redhat.com>
10579 * ubsan.c (instrument_null): Check get_base_address's result for null.
10581 2017-06-27 Marc Glisse <marc.glisse@inria.fr>
10583 * match.pd ((A+-B)+(C-A), (A+B)-(A-C)): New transformations.
10585 2017-06-27 Marc Glisse <marc.glisse@inria.fr>
10587 * builtin-types.def (BT_FENV_T_PTR, BT_CONST_FENV_T_PTR,
10588 BT_FEXCEPT_T_PTR, BT_CONST_FEXCEPT_T_PTR): New primitive types.
10589 (BT_FN_INT_FENV_T_PTR, BT_FN_INT_CONST_FENV_T_PTR,
10590 BT_FN_INT_FEXCEPT_T_PTR_INT, BT_FN_INT_CONST_FEXCEPT_T_PTR_INT):
10591 New function types.
10592 * builtins.def (BUILT_IN_FECLEAREXCEPT, BUILT_IN_FEGETENV,
10593 BUILT_IN_FEGETEXCEPTFLAG, BUILT_IN_FEGETROUND,
10594 BUILT_IN_FEHOLDEXCEPT, BUILT_IN_FERAISEEXCEPT,
10595 BUILT_IN_FESETENV, BUILT_IN_FESETEXCEPTFLAG,
10596 BUILT_IN_FESETROUND, BUILT_IN_FETESTEXCEPT,
10597 BUILT_IN_FEUPDATEENV): New builtins.
10598 * tree-core.h (TI_FENV_T_PTR_TYPE, TI_CONST_FENV_T_PTR_TYPE,
10599 TI_FEXCEPT_T_PTR_TYPE, TI_CONST_FEXCEPT_T_PTR_TYPE): New entries.
10600 * tree.h (fenv_t_ptr_type_node, const_fenv_t_ptr_type_node,
10601 fexcept_t_ptr_type_node, const_fexcept_t_ptr_type_node): New
10603 (builtin_structptr_types): Adjust size.
10604 * tree.c (builtin_structptr_types): Add four entries.
10606 2017-06-27 Jerome Lambourg <lambourg@adacore.com>
10607 Olivier Hainque <hainque@adacore.com>
10609 * config/vxworks.h (VXWORKS_LIB_SPEC): Incorporate ...
10610 (TLS_SYM): New local macro, forcing reference to __tls__ on
10611 link command lines for VxWorks 7 RTPs, triggering initialization
10613 (VXWORKS_HAVE_TLS): New macro. State whether the target VxWorks
10614 OS features TLS support, true for RTPs on VxWorks 7.
10615 * config/vxworks.c (vxworks_override_options): Setup emutls
10618 2017-06-27 Jakub Jelinek <jakub@redhat.com>
10620 * predict.c (test_prediction_value_range): Use -1U instead of -1
10621 to avoid narrowing conversion warning.
10622 * dumpfile.c (dump_options): Wrap all value into dump_flags_t cast
10623 to avoid narrowing conversion warning.
10624 * opt-functions.awk (var_ref): Return (unsigned short) -1 instead of
10626 * optc-gen.awk (END): Expect (unsigned short) -1 instead of -1.
10628 2017-06-27 Jerome Lambourg <lambourg@adacore.com>
10630 * config/vxworks.h (VXWORKS_LIBS_RTP): Alternative definition for
10631 64bit configurations.
10632 (PTR_DIFF_TYPE): Alternative definition for TARGET_LP64.
10633 (SIZE_TYPE): Likewise.
10634 * config/vxworks.c (vxworks_emutls_var_fields): Use
10635 long_unsigned_type_node instead of unsigned_type_node as the offset
10636 field type, which is "pointer" mode in emutls.c.
10638 2017-06-27 Jakub Jelinek <jakub@redhat.com>
10641 * ubsan.c (ubsan_encode_value): Initialize DECL_CONTEXT on var.
10643 PR middle-end/81207
10644 * gimple-fold.c (replace_call_with_call_and_fold): Handle
10645 gimple_vuse copying separately from gimple_vdef copying.
10647 2017-06-27 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
10649 * value-prof.c (free_hist): Remove call to memset and the enclosing if
10652 2017-06-26 Jerome Lambourg <lambourg@adacore.com>
10653 Olivier Hainque <hainque@adacore.com>
10655 * config.gcc (*-*-vxworks*): Add TARGET_VXWORKS7=1 to tm_defines
10656 for all vxworks7 targets.
10657 * config/vxworks.h (TARGET_VXWORKS7): If not defined, define to 0.
10658 (VXWORKS_ADDITIONAL_CPP_SPEC): Alternative definition for VXWORKS7.
10659 (VXWORKS_LIBS_RTP, VXWORKS_LIBS_RTP_DIR): New macros, allowing
10660 variations for VX6/VX7 and 32/64bits later on in ...
10661 (VXWORKS_LIB_SPEC): Leverage new macros.
10662 (VXWORKS_OS_CPP_BUILTINS): Define _VSB_CONFIG_FILE for VXWORKS7,
10663 as well as _ALLOW_KEYWORD_MACROS when "inline" is not a keyword.
10665 2017-06-26 Jerome Lambourg <lambourg@adacore.com>
10667 * config/vxworks.h (VXWORKS_OS_CPP_BUILTINS): builtin_define
10668 _VX_TOOL_FAMILY and _VX_TOOL to gnu.
10670 2017-06-26 Carl Love <cel@us.ibm.com>
10672 * config/rs6000/rs6000-c.c: Add support for built-in functions
10673 vector bool char vec_reve (vector bool char);
10674 vector signed char vec_reve (vector signed char);
10675 vector unsigned char vec_reve (vector unsigned char);
10676 vector bool int vec_reve (vector bool int);
10677 vector signed int vec_reve (vector signed int);
10678 vector unsigned int vec_reve (vector unsigned int);
10679 vector bool long long vec_reve (vector bool long long);
10680 vector signed long long vec_reve (vector signed long long);
10681 vector unsigned long long vec_reve (vector unsigned long long);
10682 vector bool short vec_reve (vector bool short);
10683 vector signed short vec_reve (vector signed short);
10684 vector double vec_reve (vector double);
10685 vector float vec_reve (vector float);
10686 * config/rs6000/rs6000-builtin.def (VREVE_V2DI, VREVE_V4SI,
10687 VREVE_V8HI, VREVE_V16QI, VREVE_V2DF, VREVE_V4SF, VREVE): New builtin.
10688 * config/rs6000/altivec.md (UNSPEC_VREVEV): New UNSPEC.
10689 (altivec_vreve): New pattern.
10690 * config/rs6000/altivec.h (vec_reve): New define.
10691 * doc/extend.texi (vec_rev): Update the built-in documentation file
10692 for the new built-in functions.
10694 2016-06-26 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10696 PR tree-optimization/71815
10697 * gimple-ssa-strength-reduction.c (uses_consumed_by_stmt): New
10699 (find_basis_for_candidate): Call uses_consumed_by_stmt rather than
10701 (slsr_process_phi): Likewise.
10702 (replace_uncond_cands_and_profitable_phis): Don't replace a
10703 multiply candidate with a stride of 1 (copy or cast).
10704 (phi_incr_cost): Call uses_consumed_by_stmt rather than
10706 (lowest_cost_path): Likewise.
10707 (total_savings): Likewise.
10709 2017-06-26 Richard Biener <rguenther@suse.de>
10712 * config/i386/i386.c (ix86_init_mmx_sse_builtins):
10713 Use def_builtin_pure for all gather builtins.
10715 2017-06-26 Richard Biener <rguenther@suse.de>
10717 PR tree-optimization/81203
10718 * tree-tailcall.c (find_tail_calls): Do not move stmts into
10719 non-dominating BBs.
10721 2017-06-26 Marek Polacek <polacek@redhat.com>
10724 * doc/invoke.texi: Document -Wmultistatement-macros.
10726 2017-06-26 Christophe Lyon <christophe.lyon@linaro.org>
10728 * doc/sourcebuild.texi (ARM-specific attributes): Document new
10729 arm_neon_ok_no_float_abi effective target.
10731 2017-06-26 Richard Biener <rguenther@suse.de>
10733 PR tree-optimization/80928
10734 * cfghooks.c (duplicate_block): Do not copy BB_DUPLICATED flag.
10735 (copy_bbs): Set BB_DUPLICATED flag early.
10736 (execute_on_growing_pred): Do not execute for BB_DUPLICATED
10738 (execute_on_shrinking_pred): Likewise.
10739 * tree-ssa.c (ssa_redirect_edge): Do not look for PHI args in
10740 BB_DUPLICATED blocks.
10741 * tree-ssa-phionlycoprop.c (eliminate_degenerate_phis_1): Properly
10742 iterate over all PHIs considering removal of *gsi.
10744 2017-06-23 Jim Wilson <jim.wilson@linaro.org>
10746 * doc/invoke.texi (AArch64 Options, -mtune): Re-add falkor and
10749 2017-06-23 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
10751 * config/rs6000/rs6000-string.c: (expand_block_clear,
10752 do_load_for_compare, select_block_compare_mode,
10753 compute_current_alignment, expand_block_compare,
10754 expand_strncmp_align_check, expand_strn_compare,
10755 expand_block_move, rs6000_output_load_multiple)
10756 Move functions related to string/block move/compare
10757 to a separate file.
10758 * config/rs6000/rs6000.c: Move above functions to rs6000-string.c.
10759 * config/rs6000/rs6000-protos.h (rs6000_emit_dot_insn): Add prototype
10760 for this function which is now used in two files.
10761 * config/rs6000/t-rs6000: Add rule to compile rs6000-string.o.
10762 * config.gcc: Add rs6000-string.o to extra_objs for
10763 targets powerpc*-*-* and rs6000*-*-*.
10765 2017-06-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10768 * config/rs6000/rs6000.md (ALTIVEC_DFORM): Do not allow DImode in
10769 32-bit, since indexed is not valid for DImode.
10770 (mov<mode>_hardfloat32): Reorder ISA 2.07 load/stores before ISA
10771 3.0 d-form load/stores to be the same as mov<mode>_hardfloat64.
10772 (define_peephole2 for Altivec d-form load): Add 32-bit support.
10773 (define_peephole2 for Altivec d-form store): Likewise.
10776 * multiple_target.c (create_dispatcher_calls): Only create the
10777 dispatcher call if the function is the default clone of a
10778 versioned function.
10780 2017-06-23 Segher Boessenkool <segher@kernel.crashing.org>
10782 PR middle-end/80902
10783 * builtins.c (expand_builtin_atomic_fetch_op): If emitting code after
10784 a call, force the call to not be a tail call.
10786 2017-06-23 Jeff Law <law@redhat.com>
10788 * doc/contrib.texi: Add entry for Steven Pemberton's work on
10791 2017-06-23 Will Schmidt <will_schmidt@vnet.ibm.com>
10793 * config/rs6000/rs6000.c: Add include of ssa-propagate.h for
10794 update_call_from_tree(). (rs6000_gimple_fold_builtin): Add
10795 handling for early expansion of vector shifts (sl,sr,sra,rl).
10796 (builtin_function_type): Add vector shift right instructions
10797 to the unsigned argument list.
10799 2017-06-23 Bernd Edlinger <bernd.edlinger@hotmail.de>
10801 rtl-optimizatoin/79286
10802 * ira.c (update_equiv_regs): Revert to using may_trap_or_fault_p again.
10803 * rtlanal.c (rtx_addr_can_trap_p_1): SYMBOL_REF_FUNCTION_P can never
10804 trap. PIC register plus a const unspec without offset can never trap.
10806 2017-06-23 Marc Glisse <marc.glisse@inria.fr>
10808 * tree.h (builtin_structptr_type): New type.
10809 (builtin_structptr_types): Declare new array.
10810 * tree.c (builtin_structptr_types): New array.
10811 (free_lang_data, build_common_tree_nodes): Use it.
10813 2017-06-23 Jonathan Wakely <jwakely@redhat.com>
10816 * doc/invoke.texi (-Wnoexcept-type): Fix name of option, from
10819 2017-06-22 Matt Turner <mattst88@gmail.com>
10821 * config/i386/driver-i386.c (host_detect_local_cpu): Add Kaby
10822 Lake models to skylake case. Assume skylake for unknown
10823 models with clflushopt.
10825 2017-06-22 Jeff Law <law@redhat.com>
10827 * config/aarch64/aarch64.c (aarch64_emit_probe_stack_range): Handle
10828 frame sizes that do not satisfy aarch64_uimm12_shift.
10830 2017-06-22 Jan Hubicka <hubicka@ucw.cz>
10832 * profile-count.h (apply_probability,
10833 apply_scale, probability_in): Fix checks for zero.
10835 2017-06-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10837 * incpath.c (add_sysroot_to_chain): Allow for $SYSROOT prefix.
10838 * doc/cppdiropts.texi (-I @var{dir}): Document it.
10840 2016-06-22 Richard Biener <rguenther@suse.de>
10842 * tree-vect-loop.c (vect_model_reduction_cost): Handle
10843 COND_REDUCTION and INTEGER_INDUC_COND_REDUCTION without
10844 REDUC_MAX_EXPR support.
10845 (vectorizable_reduction): Likewise.
10846 (vect_create_epilog_for_reduction): Likewise.
10848 2017-06-22 James Greenhalgh <james.greenhalgh@arm.com>
10850 * match.pd (A / (1 << B) -> A >> B): New.
10851 * generic-match-head.c: Include optabs-tree.h.
10852 * gimple-match-head.c: Likewise.
10853 * optabs-tree.h (target_supports_op_p): New.
10854 * optabs-tree.c (target_supports_op_p): New.
10856 2017-06-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10858 * configure.ac (gcc_cv_ld_static_dynamic): Also check stderr for
10859 $gcc_cv_ld --help output.
10860 (gcc_cv_ld_demangle): Likewise.
10861 (gcc_cv_ld_eh_frame_hdr): Likewise.
10862 (gcc_cv_ld_pie): Likewise.
10863 (gcc_cv_ld_as_needed): Likewise. Prefer native forms unless $gnu_ld.
10864 (gcc_cv_ld_buildid): Likewise.
10865 (gcc_cv_ld_sysroot): Likewise.
10866 (ld_bndplt_support): Likewise.
10867 (ld_pushpopstate_support): Likewise.
10868 * configure: Regenerate.
10869 * config/sol2.h [!USE_GLD] (SYSROOT_SPEC): Define.
10871 2017-06-21 Jakub Jelinek <jakub@redhat.com>
10874 * config/i386/sse.md (round<mode>2): Renumber match_dup and
10875 operands indexes to avoid gap between operands and match_dups.
10877 2017-06-21 Andrew Pinski <apinski@cavium.com>
10879 * config/aarch64/aarch64-cost-tables.h (thunderx_extra_costs):
10880 Increment Arith_shift and Arith_shift_reg by 1.
10881 * config/aarch64/aarch64-tuning-flags.def (cheap_shift_extend):
10883 * config/aarch64/aarch64.c (thunderx_tunings): Enable
10884 AARCH64_EXTRA_TUNE_CHEAP_SHIFT_EXTEND.
10885 (aarch64_strip_extend): Add new argument and test for it.
10886 (aarch64_cheap_mult_shift_p): New function.
10887 (aarch64_rtx_mult_cost): Call aarch64_cheap_mult_shift_p and don't
10888 add a cost if it is true.
10889 Update calls to aarch64_strip_extend.
10890 (aarch64_rtx_costs): Update calls to aarch64_strip_extend.
10892 2017-06-21 Andrew Pinski <apinski@cavium.com>
10894 * config/aarch64/aarch64-cores.def (thunderxt88p1): Use thunderxt88
10896 (thunderxt88): Likewise.
10897 * config/aarch64/aarch64.c (thunderxt88_prefetch_tune): New variable.
10898 (thunderx_prefetch_tune): New variable.
10899 (thunderx2t99_prefetch_tune): Update for the correct values.
10900 (thunderxt88_tunings): New variable.
10901 (thunderx_tunings): Use thunderx_prefetch_tune instead of
10902 generic_prefetch_tune.
10903 (thunderx2t99_tunings): Use AUTOPREFETCHER_WEAK.
10905 2017-06-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10907 * config/aarch64/atomics.md (aarch64_compare_and_swap<mode>_lse,
10908 SHORT): Relax operand 3 to aarch64_reg_or_zero and constraint to Z.
10909 (aarch64_compare_and_swap<mode>_lse, GPI): Likewise.
10910 (aarch64_atomic_cas<mode>, SHORT): Likewise for operand 2.
10911 (aarch64_atomic_cas<mode>, GPI): Likewise.
10913 2017-06-21 Martin Liska <mliska@suse.cz>
10915 * gimplify.c (gimplify_label_expr): Insert GIMPLE_PREDICT
10916 statements on cold and hot labels.
10917 * predict.c (tree_estimate_probability_bb): Remove the
10918 prediction from this place.
10920 2017-06-21 Martin Liska <mliska@suse.cz>
10922 PR tree-optimization/79489
10923 * gimplify.c (maybe_add_early_return_predict_stmt): New
10925 (gimplify_return_expr): Call the function.
10926 * predict.c (tree_estimate_probability_bb): Remove handling
10928 * predict.def: Update comment about early return predictor.
10929 * gimple-predict.h (is_gimple_predict): New function.
10930 * predict.def: Change default value of early return to 66.
10931 * tree-tailcall.c (find_tail_calls): Skip GIMPLE_PREDICT
10933 * passes.def: Put pass_strip_predict_hints to the beginning of
10936 2017-06-21 Pierre-Marie de Rodat <derodat@adacore.com>
10938 * dwarf2out.c (gen_decl_die): Remove the guard to skip file-scope
10939 FUNCTION_DECL declarations.
10940 (dwarf2out_early_global_decl): Remove the guard to skip FUNCTION_DECL
10942 (dwaf2out_decl): Likewise.
10943 * godump.c (go_early_global_decl): Skip call to the real debug hook
10944 for FUNCTION_DECL declarations.
10945 * passes.c (rest_of_decl_compilation): Skip call to the
10946 early_global_decl debug hook for FUNCTION_DECL declarations, unless
10947 -fdump-go-spec is passed.
10949 2017-06-21 Marc Glisse <marc.glisse@inria.fr>
10951 * config/i386/i386.c (struct builtin_isa): New field pure_p.
10952 Reorder for compactness.
10953 (def_builtin, def_builtin2, ix86_add_new_builtins): Handle pure_p.
10954 (def_builtin_pure, def_builtin_pure2): New functions.
10955 (ix86_init_mmx_sse_builtins) [__builtin_ia32_stmxcsr]: Mark as pure.
10957 2017-06-21 Marc Glisse <marc.glisse@inria.fr>
10959 * match.pd (nop_convert): New predicate.
10960 ((A +- CST1) +- CST2): Allow some NOP conversions.
10962 2017-06-21 Jakub Jelinek <jakub@redhat.com>
10965 * gimplify.c (omp_add_variable): Don't force GOVD_SEEN for types
10966 with ctors/dtors if GOVD_SHARED is set.
10968 2017-06-21 Wilco Dijkstra <wdijkstr@arm.com>
10970 * config/aarch64/aarch64.md (movti_aarch64):
10971 Emit mov rather than orr.
10972 (movtf_aarch64): Likewise.
10973 * config/aarch64/aarch64-simd.md (aarch64_simd_mov):
10974 Emit mov rather than orr.
10976 2017-06-21 Wilco Dijkstra <wdijkstr@arm.com>
10978 * config/aarch64/aarch64-simd.md (aarch64_simd_dup):
10979 Swap alternatives, make integer dup more expensive.
10981 2017-06-21 Wilco Dijkstra <wdijkstr@arm.com>
10983 * config/aarch64/aarch64.c (aarch64_legitimate_constant_p):
10984 Return true for non-tls symbols.
10986 2017-06-21 James Greenhalgh <james.greenhalgh@arm.com>
10988 * config/aarch64/aarch64-cores.def (cortex-a55): New.
10989 (cortex-a75): Likewise.
10990 (cortex-a75.cortex-a55): Likewise.
10991 * config/aarch64/aarch64-tune.md: Regenerate.
10992 * doc/invoke.texi (-mtune): Document new values for -mtune.
10994 2017-06-21 Tom de Vries <tom@codesourcery.com>
10996 * doc/sourcebuild.texi (Add Options, Features for dg-add-options): Add
10997 stack_size feature.
10998 (Effective-Target Keywords, Other attributes): Suggest using
10999 dg-add-options stack_size feature to get stack limit in stack_size
11000 effective target documentation.
11002 2017-06-21 Julian Brown <julian@codesourcery.com>
11003 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
11005 * config/aarch64/aarch64-simd.md (aarch64_crypto_pmulldi)
11006 (aarch64_crypto_pmullv2di): Change type attribute to crypto_pmull.
11007 * config/aarch64/thunderx2t99.md (thunderx2t99_pmull): New
11009 * config/arm/cortex-a53.md (cortex_a53_advsimd_type): Add crypto_pmull to
11010 attribute type list for neon_multiply.
11011 * config/arm/cortex-a57.md (cortex_a57_neon_type): Add crypto_pmull to
11012 attribute type list for neon_multiply.
11013 * config/arm/crypto.md (crypto_vmullp64): Change type to crypto_pmull.
11014 * config/arm/exynos-m1.md (exynos_m1_neon_type): Add crypto_pmull to
11015 attribute type list for neon_multiply.
11016 * config/arm/types.md (crypto_pmull): Add.
11017 * config/arm/xgene1.md (xgene1_neon_pmull): Add crypto_pmull to
11018 attribute type list.
11020 2017-06-20 Andreas Tobler <andreast@gcc.gnu.org>
11022 * config.gcc (armv6*-*-freebsd*): Change the target_cpu_cname to
11025 2017-06-20 Jakub Jelinek <jakub@redhat.com>
11027 * ira-costs.c (find_costs_and_classes): Initialize cost_classes later
11028 to make sure not to dereference a NULL cost_classes_ptr pointer.
11030 2017-06-20 Carl Love <cel@us.ibm.com>
11032 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11033 ALTIVEC_BUILTIN_VMULESW, ALTIVEC_BUILTIN_VMULEUW,
11034 ALTIVEC_BUILTIN_VMULOSW, ALTIVEC_BUILTIN_VMULOUW entries.
11035 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin,
11036 builtin_function_type): Add ALTIVEC_BUILTIN_* case statements.
11037 * config/rs6000/altivec.md (MVULEUW, VMULESW, VMULOUW,
11038 VMULOSW): New enum "unspec" values.
11039 (vec_widen_umult_even_v4si, vec_widen_smult_even_v4si,
11040 vec_widen_umult_odd_v4si, vec_widen_smult_odd_v4si,
11041 altivec_vmuleuw, altivec_vmulesw, altivec_vmulouw,
11042 altivec_vmulosw): New patterns.
11043 * config/rs6000/rs6000-builtin.def (VMLEUW, VMULESW, VMULOUW,
11044 VMULOSW): Add definitions.
11046 2017-06-20 Julia Koval <julia.koval@intel.com>
11048 * config/i386/i386.c: Fix rounding expand for new pattern.
11049 * config/i386/subst.md: Fix pattern (parallel -> unspec).
11051 2017-06-20 James Greenhalgh <james.greenhalgh@arm.com>
11053 * config/aarch64/aarch64-option-extensions.def (rcpc): New.
11054 * config/aarch64/aarch64.h (AARCH64_FL_RCPC): New.
11056 2017-06-20 James Greenhalgh <james.greenhalgh@arm.com>
11058 * config/aarch64/aarch64-option-extensions.def (fp16): Fix expected
11061 2017-06-20 James Greenhalgh <james.greenhalgh@arm.com>
11063 * config/aarch64/aarch64-cores.def: Rearrange to sort by
11064 architecture, then by implementer ID.
11065 * config/aarch64/aarch64-tune.md: Regenerate.
11067 2017-06-20 Richard Biener <rguenther@suse.de>
11069 PR middle-end/81097
11070 * fold-const.c (split_tree): Fold to type before negating.
11072 2017-06-20 David Malcolm <dmalcolm@redhat.com>
11074 * diagnostic-show-locus.c
11075 (selftest::test_fixit_deletion_affecting_newline): New function.
11076 (selftest::diagnostic_show_locus_c_tests): Call it.
11078 2017-06-20 Andreas Schwab <schwab@suse.de>
11081 * config/m68k/m68k.md (bsetdreg, bchgdreg, bclrdreg): Use "=d"
11084 2017-06-20 Prakhar Bahuguna <prakhar.bahuguna@arm.com>
11086 * config/arm/arm-c.c (arm_cpu_builtins): New block to define
11087 __ARM_FEATURE_COPROC according to support.
11089 2017-06-20 Jakub Jelinek <jakub@redhat.com>
11091 * tree-chkp.c (chkp_get_hard_register_var_fake_base_address):
11092 Rewritten to avoid overflow for > 32-bit pointers.
11095 * ubsan.h (ubsan_encode_value): Workaround buggy clang++ parser
11096 by removing enum keyword.
11097 (ubsan_type_descriptor): Likewise. Formatting fix.
11100 * config/i386/i386.md (TARGET_USE_VECTOR_CONVERTS float si->{sf,df}
11101 splitter): Require TARGET_SSE2 in the condition.
11103 2017-06-20 Michael Meissner <meissner@linux.vnet.ibm.com>
11106 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Add support
11107 for doing vector set of SFmode on ISA 3.0.
11108 * config/rs6000/vsx.md (vsx_set_v4sf_p9): Likewise.
11109 (vsx_set_v4sf_p9_zero): Special case setting 0.0f to a V4SF
11111 (vsx_insert_extract_v4sf_p9): Add an optimization for inserting a
11112 SFmode value into a V4SF variable that was extracted from another
11113 V4SF variable without converting the element to double precision
11114 and back to single precision vector format.
11115 (vsx_insert_extract_v4sf_p9_2): Likewise.
11117 2017-06-19 Jakub Jelinek <jakub@redhat.com>
11119 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Multiply
11120 in UWHI to avoid undefined overflow.
11123 * ubsan.h (enum ubsan_encode_value_phase): New.
11124 (ubsan_encode_value): Change second argument to
11125 enum ubsan_encode_value_phase with default value of
11126 UBSAN_ENCODE_VALUE_GENERIC.
11127 * ubsan.c (ubsan_encode_value): Change second argument to
11128 enum ubsan_encode_value_phase PHASE from bool IN_EXPAND_P,
11129 adjust uses, for UBSAN_ENCODE_VALUE_GENERIC use just
11130 create_tmp_var_raw instead of create_tmp_var and use a
11132 (ubsan_expand_bounds_ifn, ubsan_build_overflow_builtin,
11133 instrument_bool_enum_load, ubsan_instrument_float_cast): Adjust
11134 ubsan_encode_value callers.
11137 * ubsan.c (ubsan_encode_value): If current_function_decl is NULL,
11138 use create_tmp_var_raw instead of create_tmp_var, mark it addressable
11139 just by setting TREE_ADDRESSABLE on the result and use a TARGET_EXPR.
11141 2017-06-19 Richard Biener <rguenther@suse.de>
11143 PR middle-end/81118
11144 * tree-cfgcleanup.c (cleanup_tree_cfg_noloop): Clear niter
11145 estimates if we changed anything.
11147 2017-06-19 Richard Biener <rguenther@suse.de>
11149 PR tree-optimization/80887
11150 * tree-ssa-sccvn.c (mprts_hook_cnt): New global.
11151 (vn_lookup_simplify_result): Allow only mprts_hook_cnt succesful
11152 simplified lookups, then reset mprts_hook.
11153 (vn_nary_build_or_lookup_1): Set mprts_hook_cnt to 9 before
11155 (try_to_simplify): Likewise.
11157 2017-06-19 Martin Liska <mliska@suse.cz>
11160 * gimplify.c (gimplify_switch_expr):
11161 Initialize live_switch_vars for SWITCH_BODY == STATEMENT_LIST.
11163 2017-06-19 Martin Liska <mliska@suse.cz>
11165 * doc/install.texi: Document that PGO runs in 4 stages.
11167 2017-06-19 Martin Liska <mliska@suse.cz>
11170 * attribs.c (make_dispatcher_decl): Do not append '.ifunc'
11171 to dispatcher function name.
11172 * multiple_target.c (replace_function_decl): New function.
11173 (create_dispatcher_calls): Redirect both edges and references.
11175 2017-06-19 Jan Hubicka <hubicka@ucw.cz>
11177 * profile-count.c (profile_count::dump): Dump quality.
11178 (profile_count::differs_from_p): Update for unsigned val.
11179 * profile-count.h (profile_count_quality): New enum.
11180 (profile_count): Turn m_val to 62bit unsigned, add quality tracking.
11182 2017-06-19 Richard Biener <rguenther@suse.de>
11184 * tree-ssa-loop-niter.h (estimate_numbers_of_iterations): Take
11185 struct function as arg.
11186 (estimate_numbers_of_iterations): Export overload with loop arg.
11187 (free_numbers_of_iterations_estimates_loop): Use an overload of
11188 free_numbers_of_iterations_estimates instead.
11189 * tree-cfg.c (remove_bb): Adjust.
11190 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Likewise.
11191 * tree-parloops.c (gen_parallel_loop): Likewise.
11192 * tree-ssa-loop-ivcanon.c (canonicalize_induction_variables):
11194 (tree_unroll_loops_completely): Likewise.
11195 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop):
11196 Use an overload instead and export.
11197 (estimated_loop_iterations): Adjust.
11198 (max_loop_iterations): Likewise.
11199 (likely_max_loop_iterations): Likewise.
11200 (estimate_numbers_of_iterations): Take struct function as arg
11202 (loop_exits_before_overflow): Adjust.
11203 (free_numbers_of_iterations_estimates_loop): Use an overload.
11204 * tree-vect-loop.c (vect_analyze_loop_form): Adjust.
11205 * tree-vectorizer.c (vect_free_loop_info_assumptions): Likewise.
11207 2017-06-19 Richard Biener <rguenther@suse.de>
11210 * ipa-prop.c (find_constructor_constant_at_offset): Handle
11211 RANGE_EXPR conservatively.
11213 2017-06-16 Carl Love <cel@us.ibm.com>
11215 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11216 definitions for vec_float, vec_float2, vec_floato,
11217 vec_floate built-ins.
11218 * config/rs6000/vsx.md (define_c_enum "unspec"): Add RTL code
11219 for instructions vsx_xvcvsxws vsx_xvcvuxwsp, float2, floato and
11221 * config/rs6000/rs6000-builtin.def (FLOAT2_V2DI, FLOATE_V2DF,
11222 FLOATE_2DI, FLOATO_V2DF, FLOATEE_V2DI, XVCVSXWSP_V4SF,
11223 UNS_FLOATO_V2DI, UNS_FLOATE_V2DI): Add definitions.
11224 * config/altivec.md (define_insn "p8_vmrgew_<mode>",
11225 define_mode_attr VF_sxddp): Add V4SF type to p8_vmrgew.
11226 * config/rs6000/altivec.h (vec_float, vec_float2, vec_floate,
11227 vec_floato): Add builtin defines.
11228 * doc/extend.texi (vec_float, vec_float2, vec_floate, vec_floato):
11229 Update the built-in documentation file for the new built-in
11232 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11234 * config/arm/arm.opt (marm): Mark as the negative of of -mthumb.
11235 (mthumb): Mark as the negative of -marm.
11237 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11239 * doc/invoke.texi (ARM Options, -mcpu): Document supported
11241 (ARM Options, -mtune): Document that this accepts the same
11242 extension options as -mcpu.
11243 (ARM Options, -mfpu): Document addition of -mfpu=auto.
11245 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11247 * doc/invoke.texi (ARM Options, -march=): Document new syntax and
11248 permitted extensions.
11250 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11252 * config/arm/arm-cpus.in (armv7): Add extension +nofp.
11253 (armv7-r): Add aliases vfpv3xd and vfpv3-d16.
11254 (armv8-m.main): Add option +nodsp.
11255 * config/arm/arm-cpu-cdata.h: Regenerated.
11257 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11259 * config/arm/t-fuchsia: New file.
11260 * config.gcc (arm*-*-fuchsia*): Use it.
11262 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11264 * config/arm/t-symbian: Rewrite for new option infrastructure.
11266 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11268 * config/arm/t-phoenix (MULTILIB_REUSE): Clear variable.
11269 (MULTILIB_REQUIRED): Likewise.
11271 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11273 * config/arm/t-linux-eabi (MULTILIB_EXCEPTIONS): Set to empty.
11274 (MULTILIB_RESUE): Likewise.
11275 (MULTILIB_MATCHES): Likewise.
11276 (MULTLIB_REQUIRED): Likewise.
11278 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11280 * config/arm/t-rtems: Rewrite for new option framework.
11282 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11284 * config/arm/t-aprofile (v7_a_nosimd_variants, v7_a_simd_variants)
11285 (v7ve_nosimd_variatns, v7ve_vfpv3_simd_variants)
11286 (v7ve_vfpv4_simd_variants, v8_a_nosimd_variants, v8_a_simd_variants)
11287 (v8_1_a_simd_variants, v8_2_a_simd_variants): Move to ...
11288 * config/arm/t-multilib: ... here.
11289 (MULTILIB_OPTIONS): Add armv7 and armv7+fp architectures.
11290 (MULTILIB_MATCHES): Use armv7 libraries for armv7-r. Also use for
11291 armv7-a and armv8*-a when A-profile libraries have not been built.
11292 * config/arm/t-rmprofile: Rewrite.
11294 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11296 * genmultilib (multilib_reuse): Allow an explicit period to be escaped
11297 with a backslash. Remove the backslash after substituting unescaped
11299 * doc/fragments.texi (MULTILIB_REUSE): Document it.
11301 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11303 * config.gcc: (arm*-*-*): When building a-profile libraries, force
11304 the driver to pass through the default setting of -mfloat-abi.
11305 * common/config/arm/arm-common.c (arm_target_thumb_only): Return -marm
11307 * config/arm/t-multilib (MULTILIB_REUSE): Initialize to empty.
11308 (all_feat_combs): New rule.
11309 (MULTILIB_OPTIONS): Use explicit ARM and Thumb directories. Rework
11311 * config/arm/t-aprofile: Rewrite.
11313 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11315 * config/arm/arm.h (FPUTYPE_AUTO): Define.
11316 * config/arm/arm.c (arm_option_override): Use FPUTYPE_AUTO if the
11317 fpu is not specified by the user/command-line.
11318 * config/arm/bpabi.h (FPUTYPE_DEFAULT): Delete.
11319 * config/arm/netbsd-elf.h (FPUTYPE_DEFAULT): Delete.
11320 * config/arm/linux-elf.h (FPUTYPE_DEFAULT): Delete.
11321 * config/arm/vxworks.h (FPUTYPE_DEFAULT): Delete.
11322 * common/config/arm/arm-common.c (arm_canon_arch_option): Use
11323 FPUTYPE_AUTO insted of FPUTYPE_DEFAULT.
11325 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11327 * config/arm/elf.h (MULTILIB_DEFAULTS): Delete.
11328 * config/arm/t-arm-elf: Rewritten.
11330 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11332 * config/arm/arm.h (TARGET_HARD_FLOAT): Also check that we
11333 have some floating-point instructions.
11334 (TARGET_SOFT_FLOAT): Define as inverse of TARGET_HARD_FLOAT.
11335 (TARGET_MAYBE_HARD_FLOAT): New macro.
11336 * config/arm/arm-builtins.c (arm_init_builtins): Use
11337 TARGET_MAYBE_HARD_FLOAT.
11338 * config/arm/arm.c (arm_option_override): Use TARGET_HARD_FLOAT_ABI.
11340 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11342 * common/config/arm/arm-common.c: Define INCLUDE_LIST.
11343 (configargs.h): Include it.
11344 (arm_print_hint_for_fpu_option): New function.
11345 (arm_parse_fpu_option): New function.
11346 (candidate_extension): New class.
11347 (arm_canon_for_multilib): New function.
11348 * config/arm/arm.h (CANON_ARCH_SPEC_FUNCTION): New macro.
11349 (EXTRA_SPEC_FUNCTIONS): Add CANON_ARCH_SPEC_FUNCTION.
11350 (ARCH_CANONICAL_SPECS): New macro.
11351 (DRIVER_SELF_SPECS): Add ARCH_CANONICAL_SPECS.
11353 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11355 * config.gcc (arm*-*-*): Ensure both target_cpu_cname and with_cpu
11356 are set after handling multilib fragments. Set target_cpu_default2
11359 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11361 * config.gcc (arm*-*-fucshia*): Set target_cpu_cname to the real
11363 (arm*-*-*): Set target_cpu_default2 to a quoted string.
11364 * config/arm/parsecpu.awk (check_cpu): Validate any extension
11366 (check_arch): Likewise.
11367 * config/arm/arm.c (arm_configure_build_target): Handle
11368 TARGET_CPU_DEFAULT being a string constant. Scan any feature
11369 options in the default.
11371 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11373 * config/arm/arm-protos.h (cpu_arch_extension): Add field to record
11374 when an option is an alias of another.
11375 * config/arm/parsecpu.awk (optalias): New parser token.
11376 (gen_comm_data): Mark non-alias options as such. Emit entries
11377 for extension aliases.
11378 * config/arm/arm-cpus.in (armv5e): Make vfpv2 an alias.
11379 (armv5te, armv5tej, armv6, armv6j, armv6k, armv6z): Likewise.
11380 (armv6kz, armv6zk, armv6t2): Likewise.
11381 (armv7): Make vfpv3-d16 an alias.
11382 (armv7-a): Make vfpv3-d16, neon and neon-vfpv3 aliases. Sort in
11384 (armv7ve): Make vfpv4-d16, neon-vfpv3 and neon-vfpv4 aliases.
11385 Sort in canonical order.
11386 (armv8-a): Sort in canonical order.
11387 (armv8.1-a, armv8.2-a): Likewise.
11388 (generic-armv7-a): Make neon and neon-vfpv3 aliases. Sort in
11390 (cortex-a9): Sort in canonical order.
11391 * config/arm/arm.c (selftests.h): Include it.
11392 (arm_test_cpu_arch_data): New function.
11393 (arm_run_self_tests): New function.
11394 (TARGET_RUN_TARGET_SELFTESTS): Redefine.
11395 (targetm): Move declaration to the end of the file.
11396 * arm-cpu-cdata.h: Regenerated.
11398 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11400 * config/arm/arm.h (TARGET_MODE_SPECS): Add additional parameter to
11401 call to target_mode_check describing the type of option passed.
11402 * common/config/arm/arm-common.c (arm_arch_core_flag): Delete.
11403 (arm_target_thumb_only): Use arm_parse_arch_option_name or
11404 arm_parse_cpu_option_name to match parameters against list of
11406 * config/arm/parsecpu.awk (gen_comm_data): Don't generate
11407 arm_arch_core_flags data structure.
11408 * config/arm/arm-cpu_cdata.h: Regenerated.
11410 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11412 * common/config/arm/arm-common.c (arm_initialize_isa): Moved here from
11414 (arm_print_hint_for_cpu_option): Likewise.
11415 (arm_print_hint_for_arch_option): Likewise.
11416 (arm_parse_cpu_option_name): Likewise.
11417 (arm_parse_arch_option_name): Likewise.
11418 * config/arm/arm.c (arm_identify_fpu_from_isa): Use the computed number
11419 of entries in the all_fpus list.
11420 * config/arm/arm-protos.h (all_architectures, all_cores): Declare.
11421 (arm_parse_cpu_option_name): Declare.
11422 (arm_parse_arch_option_name): Declare.
11423 (arm_parse_option_features): Declare.
11424 (arm_intialize_isa): Declare.
11425 * config/arm/parsecpu.awk (gen_data): Move CPU and architecture
11427 (gen_comm_data): ... here. Make definitions non-static.
11428 * config/arm/arm-cpu-data.h: Regenerated.
11429 * config/arm/arm-cpu-cdata.h: Regenerated.
11431 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11433 * config/arm/arm-protos.h (arm_build_target): Remove arch_core.
11434 (cpu_arch_extension): New structure.
11435 (cpu_arch_option, arch_option, cpu_option): New structures.
11436 * config/arm/parsecpu.awk (gen_headers): Build an enumeration of
11437 architecture types.
11438 (gen_data): Generate new format data tables.
11439 * config/arm/arm.c (cpu_tune): New structure.
11440 (cpu_option, processors): Delete.
11441 (arm_print_hint_for_core_or_arch): Delete. Replace with ...
11442 (arm_print_hint_for_cpu_option): ... this and ...
11443 (arm_print_hint_for_arch_option): ... this.
11444 (arm_parse_arch_cpu_name): Delete. Replace with ...
11445 (arm_parse_cpu_option_name): ... this and ...
11446 (arm_parse_arch_option_name): ... this.
11447 (arm_unrecognized_feature): Change type of target parameter to
11449 (arm_parse_arch_cpu_features): Delete. Replace with ...
11450 (arm_parse_option_features): ... this.
11451 (arm_configure_build_target): Rework to use new configuration data
11453 (arm_print_tune_info): Rework for new configuration data tables.
11454 * config/arm/arm-cpu-data.h: Regenerated.
11455 * config/arm/arm-cpu.h: Regenerated.
11457 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11459 * Makefile.in (OBJS): Move sbitmap.o from here ...
11460 (OBJS-libcommon): ... to here.
11462 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11464 * config/arm/arm-isa.h (ISA_ALL_FPU_INTERNAL): Renamed from ISA_ALL_FPU.
11465 (ISA_ALL_CRYPTO): New macro.
11466 (ISA_ALL_SIMD): New macro
11467 (ISA_ALL_FP): New macro.
11468 * config/arm/arm.c (fpu_bitlist): Update initializer.
11469 * config/arm/arm-cpus.in: Use new ISA_ALL macros to disable crypto,
11471 (arm9e): Add fpu. Add option for nofp
11472 (arm946e-s, arm966e-s, arm968e-s, arm10e, arm1020e, arm1022e): Likewise.
11473 (arm926ej-s, arm1026ej-s): Likewise.
11474 (generic-armv7-a): Add fpu. Add options for simd, vfpv3, vfpv3-d16,
11475 vfpv3-fp16, vfpv3-d16-fp16, vfpv4, vfpv4-d16, neon, neon-vfp3,
11476 neon-fp16, neon-vfpv4, nofp and nosimd.
11477 (cortex-a5, cortex-a7): Add fpu. Add options for nosimd and nofp.
11478 (cortex-a8): Add fpu. Add option for nofp.
11479 (cortex-a9): Add fpu. Add options for nosimd and nofp.
11480 (cortex-a12, cortex-a15, cortex-a17): Add fpu. Add option for nofp.
11481 (cortex-r4f): Add fpu.
11482 (cortex-r5): Add fpu. Add options for nofp.dp and nofp.
11483 (cortex-r7): Use idiv option from architecture. Add fpu. Add option
11485 (cortex-r8): Likewise.
11486 (cortex-m4): Add fpu. Add option for nofp.
11487 (cortex-a15.cortex-a7): Add fpu. Add option for nofp.
11488 (cortex-a17.cortex-a7): Likewise.
11489 (cortex-a32): Add fpu. Add options for crypto and nofp.
11490 (cortex-a35, cortex-a53): Likewise.
11491 (cortex-a57): Add fpu. Add option for crypto.
11492 (cortex-a72, cortex-a73): Likewise.
11493 (exynos-m1): Likewise.
11494 (cortex-a57.cortex-a53, cortex-a72.cortex-a53): Likewise.
11495 (cortex-a73.cortex-a35, cortex-a73.cortex-a53): Likewise.
11496 (cortex-m33): Add fpu. Add option for nofp.
11497 * config/arm/arm-cpu-cdata.h: Regenerated
11498 * config/arm/arm-cpu-data.h: Regenerated.
11500 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11502 * arm-cpus.in (armv5e): Add options fp, vfpv2 and nofp.
11503 (armv5te, armv5tej): Likewise.
11504 (armv6, armv6j, armv6k, armv6z, armv6kz, armv6zk, armv6t2): Likewise.
11505 (armv7): Add options fp and vfpv3-d16.
11506 (armv7-a): Add options fp, simd, vfpv3, vfpv3-d16, vfpv3-d16-fp16,
11507 vfpv3-fp16, vfpv4, vfpv4-d16, neon, neon-vfpv3, neon-fp16, neon-vfpv4,
11509 (armv7ve): Likewise.
11510 (armv7-r): Add options fp, fp.sp, idiv, nofp and noidiv.
11511 (armv7e-m): Add options fp, fpv5, fp.dp and nofp.
11512 (armv8-a): Add nocrypto option.
11513 (armv8.1-a, armv8.2-a): Likewise.
11514 (armv8-m.main): add options fp, fp.dp and nofp.
11516 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11518 * config/arm/arm-cpus.in (armv8-a): Add options crc, simd crypto and
11520 (armv8-a+crc): Delete.
11521 (armv8.1-a): Add options simd, crypto and nofp.
11522 (armv8.2-a): Add options fp16, simd, crypto and nofp.
11523 (armv8.2-a+fp16): Delete.
11524 (armv8-m.main): Add option dsp.
11525 (armv8-m.main+dsp): Delete.
11526 (cortex-a8): Add fpu. Add nofp option.
11527 (cortex-a9): Add fpu. Add nofp and nosimd options.
11528 * config/arm/parsecpu.awk (gen_data): Generate option tables and
11529 link to main cpu and architecture data structures.
11530 (gen_comm_data): Only put isa attributes from the main architecture
11532 (option): New statement for architecture and CPU entries.
11533 * arm.c (struct cpu_option): New structure.
11534 (struct processors): Add entry for options.
11535 (arm_unrecognized_feature): New function.
11536 (arm_parse_arch_cpu_name): Ignore any characters after the first
11538 (arm_parse_arch_cpu_feature): New function.
11539 (arm_configure_build_target): Separate out any CPU and architecture
11540 features and parse separately. Don't error out if -mfpu=auto is
11541 used with only an architecture string.
11542 (arm_print_asm_arch_directives): New function.
11543 (arm_file_start): Call it.
11544 * config/arm/arm-cpu-cdata.h: Regenerated.
11545 * config/arm/arm-cpu-data.h: Likewise.
11546 * config/arm/arm-tables.opt: Likewise.
11548 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11550 * config/arm/elf.h (ASM_SPEC): Only pass -mfpu through to the
11551 assembler when it is not -mfpu=auto.
11553 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11555 * config/arm/arm.h (BIG_LITTLE_SPEC): Delete macro.
11556 (ASM_REWRITE_SPEC_FUNCTIONS): New macro.
11557 (BIG_LITTLE_CPU_SPEC_FUNCTIONS): Delete macro.
11558 (ASM_CPU_SPEC): Rewrite.
11559 (MCPU_MTUNE_NATIVE_FUNCTIONS): New macro.
11560 (EXTRA_SPEC_FUNCTIONS): Move outside of ifdef. Use
11561 MCPU_MTUNE_NATIVE_FUNCTIONS and ASM_REWRITE_SPEC_FUNCTIONS. Remove
11562 reference to BIG_LITTLE_CPU_SPEC_FUNCTIONS.
11563 * common/config/arm/arm-common.c (arm_rewrite_selected_cpu): Ensure
11564 copied string is NUL-terminated. Also strip any characters prefixed
11566 (arm_rewrite_selected_arch): New function.
11567 (arm_rewrite_march): New function.
11569 2017-06-16 Richard Earnshaw <rearnsha@arm.com>
11571 * config/arm/arm.opt (x_arm_arch_string): New TargetSave option.
11572 (x_arm_cpu_string, x_arm_tune_string): Likewise.
11573 (march, mcpu, mtune): Convert to string-based options.
11574 * config/arm/arm.c (arm_print_hint_for_core_or_arch): New function.
11575 (arm_parse_arch_cpu_name): New function.
11576 (arm_configure_build_target): Use arm_parse_arch_cpu_name to
11577 identify selected architecture or CPU.
11578 (arm_option_save): New function.
11579 (TARGET_OPTION_SAVE): Redefine.
11580 (arm_option_restore): Restore string options.
11581 (arm_option_print): Print string options.
11583 2017-06-16 Martin Sebor <msebor@redhat.com>
11585 PR tree-optimization/80933
11586 PR tree-optimization/80934
11587 * builtins.c (fold_builtin_3): Do not handle bcmp here.
11588 * gimple-fold.c (gimple_fold_builtin_bcmp): New function.
11589 (gimple_fold_builtin_bcopy, gimple_fold_builtin_bzero): Likewise.
11590 (gimple_fold_builtin): Call them.
11592 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11594 * gimple-ssa-isolate-paths.c (isolate_path): Set edge leading to path
11595 as unlikely; update profile.
11597 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11599 * predict.c (force_edge_cold): Handle declaring edges impossible
11602 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11604 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Update
11606 (try_unroll_loop_completely): Fix reporting.
11608 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11610 * tree-ssa-tail-merge.c (replace_block_by): Fix profile updating.
11612 2017-06-16 James Greenhalgh <james.greenhalgh@arm.com>
11615 * config/arm/arm-builtins.c (arm_expand_builtin_args): Return TARGET
11616 if given a non-constant argument for an intrinsic which requires a
11619 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11621 * profile.c (compare_freqs): New function.
11622 (branch_prob): Sort edge list.
11623 (find_spanning_tree): Assume that the list is priority sorted.
11625 2017-06-16 Richard Biener <rguenther@suse.de>
11627 PR tree-optimization/81090
11628 * passes.def (pass_record_bounds): Remove.
11629 * tree-pass.h (make_pass_record_bounds): Likewise.
11630 * tree-ssa-loop.c (pass_data_record_bounds, pass_record_bounds,
11631 make_pass_record_bounds): Likewise.
11632 * tree-ssa-loop-ivcanon.c (canonicalize_induction_variables): Do
11633 not free niter estimates at the beginning but at the end.
11634 * tree-scalar-evolution.c (scev_finalize): Free niter estimates.
11636 2017-06-16 Richard Biener <rguenther@suse.de>
11638 * tree-switch-conversion.c (emit_case_bit_tests): Adjust
11639 initializer to workaround ICE in host GCC 4.8.
11641 2017-06-16 Jan Hubicka <hubicka@ucw.cz>
11643 * ipa-inline-transform.c (update_noncloned_frequencies): Update also
11645 (clone_inlined_nodes): Update.
11647 2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
11649 * config/aarch64/aarch64.c (qdf24xx_prefetch_tune): Update
11650 prefetch settings, and enable prefetching by default at -O3.
11652 2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
11654 * config/aarch64/aarch64.c (aarch64_override_options_internal):
11655 Set flag_prefetch_loop_arrays according to tuning data.
11657 2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
11659 * config/aarch64/aarch64-protos.h (struct cpu_prefetch_tune):
11660 New tune structure.
11661 (struct tune_params): Use cpu_prefetch_tune instead of cache_line_size.
11662 [Unrelated to main purpose of the patch] Place the pointer field last
11663 to enable type checking errors when tune structure are wrongly merged.
11664 * config/aarch64/aarch64.c (generic_prefetch_tune,)
11665 (exynosm1_prefetch_tune, qdf24xx_prefetch_tune,)
11666 (thunderx2t99_prefetch_tune): New tune constants.
11667 (tune_params *_tunings): Update all tunings (no functional change).
11668 (aarch64_override_options_internal): Set PARAM_SIMULTANEOUS_PREFETCHES,
11669 PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE, and PARAM_L2_CACHE_SIZE
11670 from tunings structures.
11672 2017-06-16 Jakub Jelinek <jakub@redhat.com>
11675 * ubsan.c (instrument_null): Add T argument, use it instead
11676 of computing it based on IS_LHS.
11677 (instrument_object_size): Likewise.
11678 (pass_ubsan::execute): Adjust instrument_null and
11679 instrument_object_size callers to pass gimple_get_lhs or
11680 gimple_assign_rhs1 result to it. Use instrument_null instead of
11681 calling get_base_address and instrument_mem_ref. Handle
11682 aggregate call arguments for object-size sanitization.
11684 2017-06-16 Yury Gribov <tetra2005@gmail.com>
11686 PR tree-optimization/81089
11687 * tree-vrp.c (is_masked_range_test): Validate operands of
11690 2017-06-15 Martin Sebor <msebor@redhat.com>
11693 * dumpfile.c (dump_register): Avoid calling memset to initialize
11694 a class with a default ctor.
11695 * gcc.c (struct compiler): Remove const qualification.
11696 * genattrtab.c (gen_insn_reserv): Replace memset with initialization.
11697 * hash-table.h: Ditto.
11698 * ipa-cp.c (allocate_and_init_ipcp_value): Replace memset with
11700 * ipa-prop.c (ipa_free_edge_args_substructures): Ditto.
11701 * omp-low.c (lower_omp_ordered_clauses): Replace memset with
11703 * params.h (struct param_info): Make struct members non-const.
11704 * tree-switch-conversion.c (emit_case_bit_tests): Replace memset
11705 with default initialization.
11706 * vec.h (vec_copy_construct, vec_default_construct): New helper
11708 (vec<T>::copy, vec<T>::splice, vec<T>::reserve): Replace memcpy
11709 with vec_copy_construct.
11710 (vect<T>::quick_grow_cleared): Replace memset with default ctor.
11711 (vect<T>::vec_safe_grow_cleared, vec_safe_grow_cleared): Same.
11712 * doc/invoke.texi (-Wclass-memaccess): Document.
11714 2017-06-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11716 * emit-rtl.h (is_leaf): Update comment about local
11717 register allocator.
11719 2017-06-15 Jozef Lawrynowicz <jozef.l@somniumtech.com>
11722 * config/msp430/msp430.c (msp430_data_attr): Check that it's possible
11723 for a variable to have a section before checking if the section has a
11725 Set section to.persistent if persistent attribute is set.
11726 Warn if .persistent attribute is used on an automatic variable.
11728 2017-06-15 Eric Botcazou <ebotcazou@adacore.com>
11730 PR rtl-optimization/80474
11731 * reorg.c (update_block): Do not ignore instructions in a delay slot.
11733 2017-06-15 Segher Boessenkool <segher@kernel.crashing.org>
11735 * config/rs6000/rs6000.md (add<mode>3): Use reg_or_subregno instead
11738 2017-06-14 Maciej W. Rozycki <macro@imgtec.com>
11740 * config/mips/mips.md (MIPS16_T_REGNUM): Remove constant.
11741 (casesi): Emit bounds checking as RTL.
11742 (casesi_internal_mips16_<mode>): Remove bounds checking.
11744 2017-06-14 Max Filippov <jcmvbkbc@gmail.com>
11746 * config/xtensa/xtensa.c (xtensa_option_override): Append
11747 MASK_CONST16 to target_flags in the absence of TARGET_L32R.
11748 (hwloop_optimize, hwloop_fail, hwloop_pattern_reg,
11749 xtensa_doloop_hooks): Define unconditionally.
11750 (xtensa_reorg_loops): Only call reorg_loops in the presence of
11752 * config/xtensa/xtensa.h (TARGET_L32R): New definition.
11753 (TARGET_DEFAULT): Remove XCHAL_HAVE_L32R condition and account
11754 for it in xtensa_option_override.
11755 (HARD_FRAME_POINTER_IS_FRAME_POINTER,
11756 HARD_FRAME_POINTER_IS_ARG_POINTER): New definitions.
11758 2017-06-14 Boris Kolpackov <boris@codesynthesis.com>
11760 * doc/cppopts.texi: Document '-' special value to -MF.
11762 2017-06-14 Wilco Dijkstra <wdijkstr@arm.com>
11764 * config/arm/cortex-a53.md (cortex_a53_fpalu) Adjust latency.
11765 (cortex_a53_fconst): Likewise.
11766 (cortex_a53_fpmul): Likewise.
11767 (cortex_a53_f_load_64): Likewise.
11768 (cortex_a53_f_load_many): Likewise.
11769 (cortex_a53_advsimd_alu): Likewise.
11770 (cortex_a53_advsimd_alu_q): Likewise.
11771 (cortex_a53_advsimd_mul): Likewise.
11772 (cortex_a53_advsimd_mul_q): Likewise.
11773 (fpmac bypass): Add new bypass for fpmac-fpmac case.
11774 Add missing fmul, r2f_cvt and fconst cases.
11776 2017-06-14 Richard Biener <rguenther@suse.de>
11778 PR middle-end/81088
11779 * fold-const.c (split_tree): Drop TREE_OVERFLOW flag from
11781 (fold_binary_loc): When associating do not treat pre-existing
11782 TREE_OVERFLOW on literal constants as a reason to allow
11783 TREE_OVERFLOW on associated literal constants.
11785 2017-06-14 Eric Botcazou <ebotcazou@adacore.com>
11787 * config/sparc/sparc.h (MASK_ISA): Add MASK_LEON and MASK_LEON3.
11788 (MASK_FEATURES): New macro.
11789 * config/sparc/sparc.c (sparc_option_override): Remove the special
11790 handling of -mfpu and generalize it to all MASK_FEATURES switches.
11792 2017-06-14 Eric Botcazou <ebotcazou@adacore.com>
11794 * simplify-rtx.c (simplify_binary_operation_1) <UDIV>: Do not simplify
11795 a division of 0 if non-call exceptions are enabled.
11797 2017-06-14 Andrew Pinski <apinski@cavium.com>
11798 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
11801 * config/aarch64/aarch64.c (aarch64_expand_vector_init):
11802 Improve vector initialization code gen for only variable case.
11804 2017-06-14 Eric Botcazou <ebotcazou@adacore.com>
11806 * config/sparc/driver-sparc.c (cpu_names): Add SPARC-T5 entry.
11808 2017-06-14 Richard Biener <rguenther@suse.de>
11810 PR tree-optimization/81083
11811 * tree-ssa-sccvn.c (vn_reference_lookup_3): Do not use abnormals
11814 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11816 * config/rs6000/rs6000.c: Update all comments that mentioned SPE.
11817 (rs6000_expand_builtin): Remove RS6000_BTC_EVSEL.
11818 * config/rs6000/rs6000.h (RS6000_BTC_EVSEL): Delete.
11819 * config/rs6000/vxworks.h (VXCPU_FOR_8548): Delete. Adjust former use.
11820 * config/rs6000/vxworksae.h (VXCPU_FOR_8548): Delete.
11821 * config/rs6000/vxworksmils.h (VXCPU_FOR_8548): Delete.
11823 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11825 * config/rs6000/rs6000-opts.h (enum rs6000_vector): Delete VECTOR_SPE.
11826 * config/rs6000/rs6000.c (rs6000_debug_vector_unit): Delete VECTOR_SPE.
11828 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11830 * config/rs6000/rs6000.h (FIXED_SCRATCH): Delete.
11832 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11834 * config/rs6000/t-rtems: Don't handle SPE.
11836 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11838 * config/rs6000/t-linux: Don't handle SPE.
11840 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11842 * config/rs6000/eabispe.h: Delete file.
11844 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11846 * config/rs6000/t-spe: Delete file.
11848 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11850 * config/rs6000/rs6000.c (SPE_CONST_OFFSET_OK): Delete.
11851 (rs6000_legitimate_offset_address_p): Return false for anything in
11852 V2SImode or V2SFmode.
11854 2017-06-13 Segher Boessenkool <segher@kernel.crashing.org>
11856 * config/rs6000/rs6000-modes.def: Remove all 8-byte vector modes
11857 except V2SF and V2SI. Rearrange the vector modes, and add comments.
11858 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Remove V8QImode
11860 (reg_offset_addressing_ok_p): Remove V4HImode and V1DImode.
11861 (rs6000_legitimate_offset_address_p): Ditto.
11862 (rs6000_emit_move): Ditto.
11863 (rs6000_init_builtins): Remove V4HI_type_node.
11865 2017-06-13 Martin Liska <mliska@suse.cz>
11868 * asan.c (asan_sanitize_stack_p): Use sanitize_flags_p.
11869 (gate_asan): Likewise.
11870 * asan.h (asan_no_sanitize_address_p): Remove the function.
11871 (sanitize_flags_p): New function.
11872 * builtins.def: Fix coding style.
11873 * common.opt: Use renamed enum value.
11874 * convert.c (convert_to_integer_1): Use sanitize_flags_p.
11875 * doc/extend.texi: Document no_sanitize attribute.
11876 * flag-types.h (enum sanitize_code): Rename SANITIZE_NONDEFAULT
11877 to SANITIZE_UNDEFINED_NONDEFAULT.
11878 * gcc.c (sanitize_spec_function): Use the renamed enum value.
11879 * gimple-fold.c (optimize_atomic_compare_exchange_p):
11880 Use sanitize_flags_p.
11881 * gimplify.c (gimplify_function_tree): Likewise.
11882 * ipa-inline.c (sanitize_attrs_match_for_inline_p): Likewise.
11883 * opts.c (parse_no_sanitize_attribute): New function.
11884 (common_handle_option): Use renamed enum value.
11885 * opts.h (parse_no_sanitize_attribute): Declare.
11886 * tree.c (sanitize_flags_p): New function.
11887 * tree.h: Declared here.
11888 * tsan.c: Use sanitize_flags_p.
11889 * ubsan.c (ubsan_expand_null_ifn): Likewise.
11890 (instrument_mem_ref): Likewise.
11891 (instrument_bool_enum_load): Likewise.
11892 (do_ubsan_in_current_function): Remove the function.
11893 (pass_ubsan::execute): Use sanitize_flags_p.
11894 * ubsan.h: Remove do_ubsan_in_current_function
11895 * tree-cfg.c (print_no_sanitize_attr_value): New function.
11896 (dump_function_to_file): Use it here.
11898 2017-06-13 Martin Jambor <mjambor@suse.cz>
11900 PR tree-optimization/80803
11901 PR tree-optimization/81063
11902 * tree-sra.c (subtree_mark_written_and_enqueue): Move up in the file.
11903 (propagate_subaccesses_across_link): Enqueue subtree whenever
11904 necessary instead of relying on the caller.
11906 2017-06-13 Martin Jambor <mjambor@suse.cz>
11908 * tree-sra.c (add_access_to_work_queue): Only enqueue accesses
11909 that have a first_link.
11910 (sort_and_splice_var_accesses): Do not check first_link before
11912 (subtree_mark_written_and_enqueue): Likewise.
11913 (propagate_all_subaccesses): Likewise and do not stop at first
11914 parent with a first_link.
11916 2017-06-13 Martin Jambor <mjambor@suse.cz>
11918 * tree-sra.c (dump_access_tree_1): Fix accidental dumping to stderr
11921 2017-06-13 Yury Gribov <tetra2005@gmail.com>
11923 * match.pd: New pattern.
11925 2017-06-13 Yury Gribov <tetra2005@gmail.com>
11927 * tree-vrp.c (is_masked_range_test): New function.
11928 (register_edge_assert_for): Determine ranges for
11931 2017-06-13 Yury Gribov <tetra2005@gmail.com>
11933 PR tree-optimization/67328
11934 * fold-const.c (maskable_range_p): New function.
11935 (build_range_check): Generate bittests if possible.
11937 2017-06-13 Martin Liska <mliska@suse.cz>
11939 * gimple-pretty-print.c (dump_probability): Add new argument.
11940 (dump_edge_probability): Dump both probability and count.
11941 (dump_gimple_label): Likewise.
11942 (dump_gimple_bb_header): Likewise.
11944 2017-06-13 Georg-Johann Lay <avr@gjlay.de>
11947 * config/avr/avr-devices.c: Fix indentation.
11948 * config/avr/gen-avr-mmcu-specs.c: Dito.
11950 2017-06-13 Richard Biener <rguenther@suse.de>
11952 * tree-vect-loop.c (vect_model_reduction_cost): Do not fail,
11953 instead get vector type from stmt_info.
11954 (vectorizable_reduction): Adjust. Remove dead code.
11956 2017-06-13 Richard Biener <rguenther@suse.de>
11958 PR middle-end/81065
11959 * fold-const.c (extract_muldiv_1): Remove bogus distribution
11960 case of C * (x * C2 + C3).
11961 (fold_addr_of_array_ref_difference): Properly fold index difference.
11963 2017-06-12 David S. Miller <davem@davemloft.net>
11966 * config/sparc/sparc.md (return expander): Emit frame blockage if
11967 function uses alloca.
11969 2017-06-12 Richard Sandiford <richard.sandiford@linaro.org>
11971 * combine.c (make_field_assignment): Check len rather than the mode
11972 precision when calling force_to_mode.
11974 2017-06-12 Georg-Johann Lay <avr@gjlay.de>
11976 Support multilibs and devices that see flash in RAM address range.
11979 * config/avr/avr-arch.h (avr_arch_id) <ARCH_AVRXMEGA3>: New enum.
11980 (avr_mcu_t) <flash_pm_offset>: New field.
11981 (avr_device_specific_features) <AVR_ISA_RCALL>: New enum.
11982 * config/avr/avr.h (AVR_SHORT_CALLS): New define.
11983 (AVR_HAVE_JMP_CALL): Don't set if AVR_SHORT_CALLS.
11984 (AVR_TINY_PM_OFFSET): Remove macro.
11985 * config/avr/avr.opt (-mshort-calls): New option.
11986 * config/avr/gen-avr-mmcu-specs.c (print_mcu)
11987 [*self_spec]: Add / remove -mshort-calls depending on AVR_ISA_RCALL.
11988 * config/avr/avr-c.c (avr_cpu_cpp_builtins)
11989 <__AVR_SHORT_CALLS__>: Built-in define if AVR_SHORT_CALLS.
11990 <__AVR_HAVE_JMP_CALL__>: Use AVR_HAVE_JMP_CALL as condition
11991 instead of avr_arch->have_jmp_call.
11992 <__AVR_PM_BASE_ADDRESS__>: Built-in define if avr_arch->flash_pm_offset.
11993 [AVR_TINY] <__AVR_TINY_PM_BASE_ADDRESS__>: Use
11994 avr_arch->flash_pm_offset to define.
11995 * config/avr/avr-devices.c (avr_arch_types): Add initializers for
11996 new field flash_pm_offset. Add entry for avrxmega3.
11997 (avr_texinfo): Add entry for avrxmega3.
11998 * config/avr/avr-mcus.def: Add entries for: avrxmega3,
11999 attiny212, attiny214,
12000 attiny412, attiny414, attiny416, attiny417,
12001 attiny814, attiny816, attiny817,
12002 attiny1614, attiny1616, attiny1617,
12003 attiny3214, attiny3216, attiny3217.
12004 * config/avr/avr.c (avr_assemble_integer)[AVR_TINY]: Use
12005 avr_arch->flash_pm_offset instead of AVR_TINY_PM_OFFSET.
12006 (avr_print_operand_address) [AVR_TINY]: Same.
12007 (avr_asm_init_sections) <readonly_data_section>: Only patch
12008 callback if avr_arch->flash_pm_offset = 0.
12009 (avr_asm_named_section) <avr_need_copy_data_p>: Skip setting it
12010 for rodata if avr_arch->flash_pm_offset != 0.
12011 (avr_encode_section_info) [AVR_TINY]: Adjust comment.
12012 * config/avr/genmultilib.awk (dir_rcall, opt_rcall): New vars.
12013 (opts) [AVR_ISA_RCALL]: Append opt_rcall.
12014 (m_options): Append opt_rcall.
12015 (m_dirnames): Append dir_rcall.
12016 * config/avr/t-multilib: Regenerate.
12018 * configure.ac [target=avr]: Check whether avrxmega3 default
12019 linker description file works as needed.
12020 * configure: Regenerate.
12021 * doc/avr-mmcu.texi: Regenerate.
12022 * doc/invoke.texi (AVR Options) <-mshort-calls>: Document it.
12023 <__AVR_ARCH__>: Document avrxmega3 and 103.
12024 <__AVR_HAVE_JMP_CALL__>: Adjust documentation.
12025 <__AVR_SHORT_CALLS__>: Document it.
12026 <__AVR_PM_BASE_ADDRESS__>: Document it.
12027 * doc/extend.texi (AVR Options) <-mshort-calls>: Document it.
12028 (AVR Variable Attributes) <progmem>: Document this is
12029 not needed for avrxmega3.
12030 (AVR Named Address Spaces) <__flash>: Dito.
12032 2017-06-12 Jan Hubicka <hubicka@ucw.cz>
12034 * cgraph.c (cgraph_node::dump): Complain about profile insanities.
12036 2017-06-12 Doug Rupp <rupp@adacore.com>
12038 * config.gcc (*-*-vxworks*): Set use_gcc_stdint to "provide".
12039 Append vxworks-stdint.h to the tm_file list.
12040 * config/vxworks-stdint.h: New file.
12042 2017-06-12 Martin Liska <mliska@suse.cz>
12044 PR tree-optimization/81041
12045 * tree-profile.c (gimple_gen_ic_func_profiler):
12046 Create an extra BB in profile-generate
12047 (gimple_gen_time_profiler): Likewise.
12049 2017-06-12 Jakub Jelinek <jakub@redhat.com>
12051 PR tree-optimization/81003
12052 * tree-ssa-reassoc.c (force_into_ssa_name): New function.
12053 (update_range_test): Use it instead of force_gimple_operand_gsi.
12055 2017-06-12 Richard Biener <rguenther@suse.de>
12057 PR tree-optimization/81053
12058 * tree-vect-loop.c (vect_is_simple_reduction): Handle PHI
12059 with backedge value not defined in loop. Simplify def stmt
12062 2017-06-11 Tom de Vries <tom@codesourcery.com>
12065 * config/nvptx/nvptx.c (nvptx_cannot_force_const_mem): New function.
12067 (TARGET_CANNOT_FORCE_CONST_MEM): Redefine to
12068 nvptx_cannot_force_const_mem.
12070 2017-06-10 Jan Hubicka <hubicka@ucw.cz>
12072 * opts.c (finish_options): Move test for flag_split_stack after
12073 it has been initialized.
12075 2017-06-11 Jason Merrill <jason@redhat.com>
12077 * tree.h (id_equal): New.
12078 * dwarf2out.c, hsa-gen.c, ipa-devirt.c, omp-expand.c,
12079 omp-simd-clone.c, read-rtl-function.c, tree-chkp.c, tree.c: Use it
12080 instead of strcmp of IDENTIFIER_POINTER.
12082 2017-06-10 Jan Hubicka <hubicka@ucw.cz>
12084 * ipa-inline-transform.c: Include function.h, cfg.h and basic-block.h
12085 (mark_all_inlined_calls_cdtor): Fix formating.
12086 (inline_transform): Rescale profile before inlining.
12088 2017-06-10 Jan Hubicka <hubicka@ucw.cz>
12090 * cgraph.h (cgraph_edge::clone): Update prototype.
12091 * cgraphclones.c (cgraph_edge::clone): Update profile scaling.
12092 (cgraph_node::create_clone): Update.
12093 (cgraph_node::create_version_clone): Update.
12094 * tree-inline.c (copy_bb): Update.
12095 (expand_call_inline): Update.
12097 2017-06-10 Segher Boessenkool <segher@kernel.crashing.org>
12099 * config/rs6000/rs6000.c (emit_vrsave_prologue): New function,
12100 factored out from ...
12101 (rs6000_emit_prologue): ... here.
12103 2017-06-10 Segher Boessenkool <segher@kernel.crashing.org>
12105 * config/rs6000/rs6000.c (emit_split_stack_prologue): New function,
12106 factored out from ...
12107 (rs6000_emit_prologue): ... here.
12109 2017-06-10 Jan Hubicka <hubicka@ucw.cz>
12111 * predict.c (drop_profile): Also drop individual bb/edge and cgraph
12113 (handle_missing_profiles): Fix computation of tp_first_run.
12114 (counts_to_freqs): Do not touch freqs when count is 0.
12116 2017-06-10 Jan Hubicka <hubicka@ucw.cz>
12118 * cgraphbuild.c (cgraph_edge::rebuild_references): Do not touch
12121 2017-06-10 Tom de Vries <tom@codesourcery.com>
12123 * doc/sourcebuild.texi (Effective-Target Keywords, Environment
12124 attributes): Document signal effective target.
12126 2017-06-10 Tom de Vries <tom@codesourcery.com>
12128 * doc/sourcebuild.texi (Effective-Target Keywords, Other attributes):
12129 Document effective target stack_size.
12131 2017-06-09 David Malcolm <dmalcolm@redhat.com>
12133 * diagnostic.c (diagnostic_report_diagnostic): Only add fixits
12134 to the edit_context if they can be auto-applied.
12136 2017-06-9 Ian Lance Taylor <iant@golang.org>
12138 * opts.c (finish_options): If -fsplit-stack, disable implicit
12139 -forder-blocks-and-partition.
12140 * doc/invoke.texi (Optimize Options): Document that when using
12141 -fsplit-stack -forder-blocks-and-partition is not implicitly
12144 2017-06-09 Jan Hubicka <hubicka@ucw.cz>
12146 * builtin-attrs.def (ATTR_NORETURN_NOTHROW_LEAF_COLD_LIST,
12147 ATTR_CONST_NORETURN_NOTHROW_LEAF_COLD_LIST,
12148 ATTR_TMPURE_NORETURN_NOTHROW_LEAF_COLD_LIST): New.
12149 * builtins.def (abort, trap, unreachable): Declare cold.
12150 * calls.c (flags_from_decl_or_type): Lookup ECF_COLD.
12151 * tree-core.h (ECF_COLD): New.
12152 * tree.c (set_call_expr_flags): Handle ECF_COLD.
12153 (build_common_builtin_nodes): Mark unreachable and abort as cold.
12155 2017-06-09 Jan Hubicka <hubicka@ucw.cz>
12157 * predict.c (unlikely_executed_stmt_p): Cleanup.
12159 2017-06-09 Richard Biener <rguenther@suse.de>
12161 * tree-ssa-loop-im.c (execute_sm): Do not force multi-threaded
12162 model if the ref is always written to.
12164 2017-06-09 Tamar Christina <tamar.christina@arm.com>
12166 * config/aarch64/aarch64.md (lrint<GPF:mode><GPI:mode>2): New.
12168 2017-06-09 Tamar Christina <tamar.christina@arm.com>
12170 * config/arm/arm.c (arm_rtx_costs_internal): Make sdiv more expensive
12173 2017-06-09 Tom de Vries <tom@codesourcery.com>
12176 * config/nvptx/nvptx.md (define_expand "mov<QHSDISDFM>"): Error out with
12177 "target cannot support label values" when encountering LABEL_REF.
12179 2017-06-09 Martin Liska <mliska@suse.cz>
12181 * tree-profile.c (gimple_gen_ic_profiler): Update comment.
12182 (gimple_gen_ic_func_profiler): Emit direct comparison
12183 of __gcov_indirect_call_callee with NULL.
12184 (gimple_gen_time_profiler): Change probability from
12185 PROB_VERY_UNLIKELY to PROB_UNLIKELY.
12187 2017-06-09 Jan Hubicka <hubicka@ucw.cz>
12189 * profile.c (edge_gcov_counts): Turn to pointer.
12190 (compute_branch_probabilities, compute_branch_probabilities): Update.
12191 (branch_prob): Do not clear edge_gcov_count.
12192 * profile.h (edge_gcov_counts): Turn to pointer.
12193 (edge_gcov_count): Update.
12195 2017-06-09 Jan Hubicka <hubicka@ucw.cz>
12197 * gimple.h (gimple_check_failed): Mark cold.
12199 2017-06-09 Richard Biener <rguenther@suse.de>
12201 PR tree-optimization/66623
12202 * tree-vect-loop.c (vect_is_simple_reduction): Cleanup,
12203 refactor check_reduction into two parts, properly computing
12204 whether we have to check reduction validity for outer loop
12207 2017-06-09 Richard Biener <rguenther@suse.de>
12209 PR tree-optimization/79483
12210 * graphite-scop-detection.c (order): New global.
12211 (get_order): Compute bb to order mapping that satisfies code
12212 generation constraints.
12213 (cmp_pbbs): New helper.
12214 (build_scops): Start domwalk at entry block, sort generated
12217 2017-06-09 Richard Biener <rguenther@suse.de>
12219 PR middle-end/81007
12220 * ipa-polymorphic-call.c
12221 (ipa_polymorphic_call_context::restrict_to_inner_class):
12222 Skip FIELD_DECLs with error_mark_node type.
12223 * passes.def (all_lowering_passes): Run pass_build_cgraph_edges
12226 2017-06-09 Martin Liska <mliska@suse.cz>
12228 * predict.c (struct branch_predictor): New struct.
12229 (test_prediction_value_range): New test.
12230 (predict_c_tests): New function.
12231 * selftest-run-tests.c (selftest::run_tests): Run the function.
12232 * selftest.h: Declare new tests.
12234 2017-06-09 Segher Boessenkool <segher@kernel.crashing.org>
12237 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Assert that
12238 gen_add3_insn did not fail.
12239 * config/rs6000/rs6000.md (add<mode>3): If asked to add a constant to
12240 r0, construct that number in a temporary reg and add that reg to r0.
12241 If asked to put the result in r0 as well, fail.
12243 2017-06-08 Will Schmidt <will_schmidt@vnet.ibm.com>
12245 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling
12246 for early expansion of vec_eqv.
12248 2017-06-08 Jakub Jelinek <jakub@redhat.com>
12250 PR middle-end/81005
12251 * ubsan.c (instrument_null): Avoid pointless code temporary.
12252 (pass_ubsan::execute): Instrument aggregate arguments of calls.
12254 2017-06-08 Uros Bizjak <ubizjak@gmail.com>
12258 2016-12-14 Uros Bizjak <ubizjak@gmail.com>
12261 * config/i386/i386.md (*ctzhi2): New insn_and_split pattern.
12264 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12266 * predict.c (unlikely_executed_edge_p): Move ahead.
12267 (probably_never_executed_edge_p): Use it.
12269 2017-06-08 Alexander Ivchenko <alexander.ivchenko@intel.com>
12271 PR middle-end/79988
12272 * tree-chkp.c (chkp_gimple_call_builtin_p): Remove
12273 gimple_call_builtin_p call.
12275 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12277 * system.h (fancy_abort): Annotate by ATTRIBUTE_COLD.
12278 * rtl.h (rtl_check_failed_bounds, rtl_check_failed_type1,
12279 rtl_check_failed_type2, rtl_check_failed_code1,
12280 rtl_check_failed_code2, rtl_check_failed_code_mode,
12281 rtl_check_failed_block_symbol, cwi_check_failed_bounds,
12282 rtvec_check_failed_bounds, rtl_check_failed_flag,
12283 _fatal_insn_not_found, _fatal_insn): Likewise.
12284 * tree.h (tree_contains_struct_check_failed,
12285 tree_check_failed, tree_not_check_failed,
12286 tree_class_check_failed, tree_range_check_failed,
12287 tree_not_class_check_failed, tree_int_cst_elt_check_failed,
12288 tree_vec_elt_check_failed, phi_node_elt_check_failed,
12289 tree_operand_check_failed, omp_clause_check_failed,
12290 omp_clause_operand_check_failed, omp_clause_range_check_failed):
12293 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12295 * cgraph.c (cgraph_edge::maybe_hot_p): Do not check
12296 flag_branch_probabilities.
12297 * ipa-inline.c (edge_badness): Likewise.
12298 * ipa-profile.c (ipa_propagate_frequency_1): Likewise.
12299 * postreload-gcse.c (eliminate_partially_redundant_load): Likewise.
12300 * predict.c (maybe_hot_frequency_p): Likewise.
12301 (probably_never_executed): Likewise.
12302 * sched-ebb.c (schedule_ebbs): Likewise.
12303 * sched-rgn.c (find_single_block_region): Likewise.
12304 * tracer.c (tail_duplicate): Likewise.
12306 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12308 * opts.c (finish_options): x_flag_reorder_blocks_and_partition no
12309 longer requires x_flag_profile_use.
12311 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12313 * cfgrtl.c (cfg_layout_initialize): Check crtl->has_bb_partition
12314 instead of flag_reorder_blocks_and_partition.
12315 * dbxout.c (dbxout_function_end): Likewise.
12316 * dwarf2out.c (gen_subprogram_die): Likewise.
12317 * haifa-sched.c (sched_create_recovery_edges): Likewise.
12318 * hw-doloop.c (reorg_loops): Likewise.
12319 * varasm.c (assemble_start_function,
12320 assemble_end_function): Likewise.
12321 (decide_function_section): Do not check for
12322 flag_reorder_blocks_and_partition.
12324 2017-06-08 Alexander Ivchenko <alexander.ivchenko@intel.com>
12326 * tree-chkp.c (chkp_get_hard_register_var_fake_base_address):
12328 (chkp_get_hard_register_fake_addr_expr): Ditto.
12329 (chkp_build_addr_expr): Add check for hard reg case.
12330 (chkp_parse_array_and_component_ref): Ditto.
12331 (chkp_find_bounds_1): Ditto.
12332 (chkp_process_stmt): Don't generate bounds store for
12335 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12337 * predict.c (maybe_hot_bb_p): Do not check profile status.
12338 (maybe_hot_edge_p): Likewise.
12339 (probably_never_executed): Check for zero counts even if profile
12341 (unlikely_executed_edge_p): New function.
12342 (unlikely_executed_stmt_p): New function.
12343 (unlikely_executed_bb_p): New function.
12344 (set_even_probabilities): Use unlikely predicates.
12345 (combine_predictions_for_bb): Likewise.
12346 (predict_paths_for_bb): Likewise.
12347 (predict_paths_leading_to_edge): Likewise.
12348 (determine_unlikely_bbs): New function.
12349 (estimate_bb_frequencies): Use it.
12350 (compute_function_frequency): Use zero counts even if profile is
12352 * profile-count.h: Fix typo.
12354 2017-08-08 Julia Koval <julia.koval@intel.com>
12356 * config/i386/avx512bwintrin.h (_mm512_mask_cvtepi16_storeu_epi8,
12357 _mm512_mask_cvtsepi16_storeu_epi8,
12358 _mm512_mask_cvtusepi16_storeu_epi8): New intrinsics.
12359 * config/i386/avx512vlbwintrin.h (_mm256_mask_cvtepi16_storeu_epi8,
12360 _mm_mask_cvtsepi16_storeu_epi8, _mm256_mask_cvtsepi16_storeu_epi8,
12361 _mm_mask_cvtusepi16_storeu_epi8, _mm256_mask_cvtusepi16_storeu_epi8,
12362 _mm_mask_cvtepi16_storeu_epi8): New intrinsics.
12363 * config/i386/i386-builtin-types.def (PV8Q, V8QI): New pointer type.
12364 (VOID_FTYPE_PV32QI_V32HI_USI, VOID_FTYPE_PV8QI_V8HI_UQI,
12365 VOID_FTYPE_PV16QI_V16HI_UHI): New function types.
12366 * config/i386/i386-builtin.def (__builtin_ia32_pmovwb128mem_mask,
12367 __builtin_ia32_pmovwb256mem_mask, __builtin_ia32_pmovswb128mem_mask,
12368 __builtin_ia32_pmovswb256mem_mask, __builtin_ia32_pmovuswb128mem_mask,
12369 __builtin_ia32_pmovuswb256mem_mask,
12370 __builtin_ia32_pmovuswb512mem_mask, __builtin_ia32_pmovswb512mem_mask)
12371 __builtin_ia32_pmovwb512mem_mask): New builtins.
12373 2017-08-08 Julia Koval <julia.koval@intel.com>
12375 PR target/73350,80862
12376 * config/i386/subst.md (round): Fix round pattern.
12377 * config/i386/i386.c (ix86_erase_embedded_rounding):
12378 Fix erasing rounding for the fixed pattern.
12380 2017-06-08 Jan Hubicka <hubicka@ucw.cz>
12382 * cfgbuild.c (find_many_sub_basic_blocks): Fix thinko.
12384 2017-06-08 Martin Liska <mliska@suse.cz>
12386 PR gcov-profile/80911
12387 * gcov.c (block_info::block_info): New constructor.
12389 2017-06-07 Carl Love <cel@us.ibm.com>
12391 * config/rs6000/rs6000-c: The return type of the following
12392 built-in functions was implemented as int not long long. Fix sign
12393 of return value for the unsigned version of vec_mulo and vec_mule.
12394 vector unsigned long long vec_bperm (vector unsigned long long,
12395 vector unsigned char)
12396 vector signed long long vec_mule (vector signed int,
12398 vector unsigned long long vec_mule (vector unsigned int,
12399 vector unsigned int)
12400 vector signed long long vec_mulo (vector signed int,
12402 vector unsigned long long vec_mulo (vector unsigned int,
12403 vector unsigned int)
12404 * doc/extend.texi: Fix the documentation for the built-in
12407 2017-06-07 Carl Love <cel@us.ibm.com>
12410 * config/rs6000/altivec.md (double<mode>2): Fix the implementation of
12413 2017-06-07 Carl Love <cel@us.ibm.com>
12415 * config/rs6000/altivec.md: Fix argument swizzle in vec_doublel
12416 support, Generate doublehv for signed int/float for BE case only.
12418 2017-06-07 Alexander Monakov <amonakov@ispras.ru>
12420 * doc/invoke.texi (mcx16): Rewrite.
12422 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12424 * config/rs6000/predicates.md (rs6000_nonimmediate_operand): Delete.
12425 * config/rs6000/rs6000.md (*movsi_internal1, movsi_from_sf,
12426 *mov<mode>_softfloat, and an anonymous splitter): Use
12427 nonimmediate_operand instead of rs6000_nonimmediate_operand.
12429 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12431 * config/rs6000/darwin.h (REGISTER_NAMES): Delete the SPE_ACC and
12433 * config/rs6000/rs6000.c (rs6000_reg_names, alt_reg_names): Ditto.
12434 (enum rs6000_reg_type): Delete SPE_ACC_TYPE and SPEFSCR_REG_TYPE.
12435 (rs6000_debug_reg_global): Adjust.
12436 (rs6000_init_hard_regno_mode_ok): Adjust.
12437 (rs6000_dbx_register_number): Adjust.
12438 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTER): Change to 115.
12439 (FIXED_REGISTERS, CALL_USED_REGISTERS, CALL_REALLY_USED_REGISTERS):
12440 Remove SPE_ACC and SPEFSCR.
12441 (REG_ALLOC_ORDER): Ditto.
12442 (FRAME_POINTER_REGNUM): Change to 111.
12443 (enum reg_class): Remove the SPE_ACC and SPEFSCR registers.
12444 (REG_CLASS_NAMES): Ditto.
12445 (REG_CLASS_CONTENTS): Delete the SPE_ACC and SPEFSCR registers.
12446 (REGISTER_NAMES): Ditto.
12447 (ADDITIONAL_REG_NAMES): Ditto.
12448 (rs6000_reg_names): Ditto.
12449 * config/rs6000/rs6000.md: Renumber some register number
12452 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12454 * config/rs6000/darwin.h (REGISTER_NAMES): Delete the SPE high
12456 * config/rs6000/rs6000.c (rs6000_reg_names, alt_reg_names): Ditto.
12457 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTER): Change from 149
12459 (DWARF_REG_TO_UNWIND_COLUMN): Do not define.
12460 (FIXED_REGISTERS, CALL_USED_REGISTERS, CALL_REALLY_USED_REGISTERS):
12461 Delete the SPE high registers.
12462 (REG_ALLOC_ORDER): Ditto.
12463 (enum reg_class): Remove SPE_HIGH_REGS.
12464 (REG_CLASS_NAMES): Ditto.
12465 (REG_CLASS_CONTENTS): Delete the SPE high registers.
12466 (REGISTER_NAMES): Ditto.
12467 (rs6000_reg_names): Ditto.
12468 * doc/tm.texi.in: Remove SPE as example.
12469 * doc/tm.texi: Regenerate.
12471 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12473 * config/rs6000/8540.md (ppc8540_brinc): Delete.
12474 * config/rs6000/e500mc.md (e500mc_brinc): Delete.
12475 * config/rs6000/e500mc64.md (e500mc64_brinc): Delete.
12476 * config/rs6000/rs6000.md (type): Remove "brinc".
12478 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12480 * config.gcc (powerpc*-*-*): Don't add spe.h to extra_headers.
12481 (powerpc*-linux*spe*): Use ${cpu_type} instead of rs6000.
12482 * config/rs6000/linuxspe.h: Delete file.
12483 * config/rs6000/rs6000.md: Don't include spe.md.
12484 * config/rs6000/spe.h: Delete file.
12485 * config/rs6000/spe.md: Delete file.
12486 * config/rs6000/t-rs6000: Remove spe.md.
12488 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12490 * config/rs6000/predicates.md (reg_or_mem_operand): Reformat.
12491 (reg_or_none500mem_operand): Delete.
12492 * config/rs6000/rs6000.md (extendsfdf2): Use reg_or_mem_operand
12493 instead of reg_or_none500mem_operand.
12495 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12497 * config/rs6000/rs6000.c (rs6000_option_override_internal): Delete
12498 handling of SPE flags.
12499 * config/rs6000/rs6000.opt (-mspe, -mspe=no, -mspe=yes): Delete.
12501 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12503 * config/rs6000/rs6000-common.c (rs6000_handle_option): Remove
12505 * config/rs6000/paired.md (paired_negv2sf2): Rename to negv2sf2.
12506 (paired_absv2sf2, paired_addv2sf3, paired_subv2sf3, paired_mulv2sf3,
12507 paired_divv2sf3): Similar.
12508 * config/rs6000/predicates.md: Replace TARGET_SPE, TARGET_SPE_ABI,
12509 SPE_VECTOR_MODE and SPE_HIGH_REGNO_P by 0; simplify.
12510 * config/rs6000/rs6000-builtin.def: Delete RS6000_BUILTIN_E and
12512 Delete BU_SPE_1, BU_SPE_2, BU_SPE_3, BU_SPE_E, BU_SPE_P, and BU_SPE_X.
12513 Rename the paired_* instruction patterns.
12514 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Do not
12516 * config/rs6000/rs6000-protos.h (invalid_e500_subreg): Delete.
12517 * config/rs6000/rs6000.c: Delete RS6000_BUILTIN_E and RS6000_BUILTIN_S.
12518 (struct rs6000_stack): Delete fields spe_gp_save_offset, spe_gp_size,
12519 spe_padding_size, and spe_64bit_regs_used. Replace TARGET_SPE and
12520 TARGET_SPE_ABI with 0, simplify. Replace SPE_VECTOR_MODE with
12521 PAIRED_VECTOR_MODE.
12522 (struct machine_function): Delete field spe_insn_chain_scanned_p.
12523 (spe_func_has_64bit_regs_p): Delete.
12524 (spe_expand_predicate_builtin): Delete.
12525 (spe_expand_evsel_builtin): Delete.
12526 (TARGET_DWARF_REGISTER_SPAN): Do not define.
12527 (TARGET_MEMBER_TYPE_FORCES_BLK): Do not define.
12528 (invalid_e500_subreg): Delete.
12529 (rs6000_legitimize_address): Always force_reg op2 as well, for
12530 paired single memory accesses.
12531 (rs6000_member_type_forces_blk): Delete.
12532 (rs6000_spe_function_arg): Delete.
12533 (rs6000_expand_unop_builtin): Delete SPE handling.
12534 (rs6000_expand_binop_builtin): Ditto.
12535 (spe_expand_stv_builtin): Delete.
12536 (bdesc_2arg_spe): Delete.
12537 (spe_expand_builtin): Delete.
12538 (spe_expand_predicate_builtin): Delete.
12539 (spe_expand_evsel_builtin): Delete.
12540 (rs6000_invalid_builtin): Remove RS6000_BTM_SPE handling.
12541 (spe_init_builtins): Delete.
12542 (spe_func_has_64bit_regs_p): Delete.
12543 (savres_routine_name): Delete "info" parameter. Adjust callers.
12544 (rs6000_emit_stack_reset): Ditto.
12545 (rs6000_dwarf_register_span): Delete.
12546 * config/rs6000/rs6000.h (TARGET_SPE_ABI, TARGET_SPE,
12547 UNITS_PER_SPE_WORD, SPE_HIGH_REGNO_P, SPE_SIMD_REGNO_P,
12548 SPE_VECTOR_MODE, RS6000_BTM_SPE, RS6000_BUILTIN_E, RS6000_BUILTIN_S):
12550 * config/rs6000/rs6000.md (FIRST_SPE_HIGH_REGNO, LAST_SPE_HIGH_REGNO):
12552 * config/rs6000/rs6000.opt (-mabi=spe, -mabi=no-spe): Delete.
12553 * config/rs6000/spe.md: Delete every pattern that uses TARGET_SPE.
12554 * config/rs6000/vector.md (absv2sf2, negv2sf2, addv2sf3, subv2sf3,
12555 mulv2sf3, divv2sf3): Delete expanders.
12557 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12559 config/rs6000/rs6000.md (UNSPEC_MV_CR_GT): Delete.
12561 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12563 * config/rs6000/rs6000-protos.h (output_e500_flip_gt_bit): Delete.
12564 * config/rs6000/rs6000.c: Ditto.
12566 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12568 * config/rs6000/predicated.md (rs6000_cbranch_operator): Delete.
12569 * config/rs6000/rs6000.md: Replace rs6000_cbranch_operator by
12570 comparison_operator.
12572 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12574 * config/rs6000/rs6000.c: Remove everything related to -mfloat-gprs.
12575 * config/rs6000/rs6000.opt: Ditto.
12576 * config/rs6000/t-rtems: Ditto.
12578 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12580 * config/rs6000/predicates.md: Replace TARGET_E500_DOUBLE and
12581 TARGET_E500_SINGLE by 0, simplify.
12582 * config/rs6000/rs6000.c: Ditto.
12583 (rs6000_option_override_internal): Delete CHECK_E500_OPTIONS.
12584 (spe_build_register_parallel): Delete.
12585 * config/rs6000/rs6000.h: Delete TARGET_E500_SINGLE,
12586 TARGET_E500_DOUBLE, and CHECK_E500_OPTIONS.
12587 * config/rs6000/rs6000.md: Replace TARGET_E500_DOUBLE,
12588 TARGET_E500_SINGLE, and <E500_CONVERT> by 0, simplify.
12589 (E500_CONVERT): Delete.
12590 * config/rs6000/spe.md: Remove many patterns and all define_constants.
12592 2017-06-07 Segher Boessenkool <segher@kernel.crashing.org>
12594 * config/rs6000/darwin.md: Replace TARGET_FPRS by 1 and simplify.
12595 * config/rs6000/dfp.md: Ditto.
12596 (negdd2, *negdd2_fpr): Merge.
12597 (absdd2, *absdd2_fpr): Merge.
12598 (negtd2, *negtd2_fpr): Merge.
12599 (abstd2, *abstd2_fpr): Merge.
12600 * config/rs6000/e500.h: Delete file.
12601 * config/rs6000/predicates.md (rs6000_cbranch_operator): Replace
12602 TARGET_FPRS by 1 and simplify.
12603 * config/rs6000/rs6000-c.c: Ditto.
12604 * config/rs6000/rs6000.c: Ditto. Also replace TARGET_SF_SPE and
12605 TARGET_DF_SPE by 0.
12606 * config/rs6000/rs6000.h: Ditto. Delete TARGET_SF_SPE and
12608 * config/rs6000/rs6000.md: Ditto.
12609 (floatdidf2, *floatdidf2_fpr): Merge.
12610 (move_from_CR_gt_bit): Delete.
12611 * config/rs6000/spe.md: Replace TARGET_FPRS by 1 and simplify.
12612 (E500_CR_IOR_COMPARE): Delete.
12613 (All patterns that require !TARGET_FPRS): Delete.
12614 * config/rs6000/vsx.md: Replace TARGET_FPRS by 1 and simplify.
12616 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12618 * passes.def (pass_iv_canon): Move before pass_loop_distribution.
12620 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12622 * graphds.c (add_edge): Intitialize edge's attached data.
12623 (foll_in_subgraph, dfs_fst_edge, dfs_next_edge): New function
12624 pointer parameter. Call pointed function on each edge during
12625 graph traversing. Skip traversing the edge when the function
12627 (graphds_dfs, graphds_scc): Ditto.
12628 (for_each_edge): New parameter. Pass the new parameter to callback
12630 * graphds.h (skip_edge_callback): New function pointer type.
12631 (graphds_dfs, graphds_scc): New function pointer parameter.
12632 (graphds_edge_callback, for_each_edge): New parameter.
12634 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12636 * tree-vect-data-refs.c (vect_mark_for_runtime_alias_test): Factor
12637 out code checking if runtime alias check is possible to below ...
12638 Call the new function.
12639 * tree-data-ref.c (runtime_alias_check_p): ... to new function.
12640 * tree-data-ref.h (runtime_alias_check_p): New decalaration.
12642 2017-06-07 Marek Polacek <polacek@redhat.com>
12645 * fold-const.c (extract_muldiv_1) <case MINUS_EXPR>: Add
12646 TYPE_OVERFLOW_WRAPS check.
12648 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12650 * tree-vect-loop-manip.c (vect_do_peeling): Don't skip vector loop
12651 if versioning is required.
12652 * tree-vect-loop.c (vect_analyze_loop_2): Merge niter check for loop
12653 peeling with the check for versioning.
12655 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12657 * tree-vectorizer.h (vect_build_loop_niters): New parameter.
12658 * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter.
12659 Set true to new parameter if new ssa variable is defined.
12660 (vect_gen_vector_loop_niters): Refactor. Set range information
12661 for the new vector loop bound variable.
12662 (vect_do_peeling): Ditto.
12664 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12666 * tree-affine.c (ssa.h): Include header file.
12667 (tree_to_aff_combination): Handle (T1)(X - CST) when inner type
12668 has wrapping overflow behavior.
12670 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12672 * tree-affine.c (tree_to_aff_combination): Handle (T1)(X + X).
12674 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12676 (aff_combination_expand): Move (T1)(X *+- CST) simplification to ...
12677 (tree_to_aff_combination): ... here.
12679 2017-06-07 Bin Cheng <bin.cheng@arm.com>
12681 * tree-ssa-loop-ivopts.c (ivopts_estimate_reg_pressure): New
12682 reg_pressure model function.
12683 (ivopts_global_cost_for_size): Delete.
12684 (determine_set_costs, iv_ca_recount_cost): Call new model function
12685 ivopts_estimate_reg_pressure.
12687 2017-06-07 Tamar Christina <tamar.christina@arm.com>
12689 * config/aarch64/aarch64.c (aarch64_rtx_costs): Make sdiv more
12690 expensive than udiv. Remove floating point cases from mod.
12692 2017-06-07 Tamar Christina <tamar.christina@arm.com>
12694 * config/arm/aarch-cost-tables.h (cortexa53_extra_cost):
12695 Increase idiv cost.
12697 2017-06-07 Tamar Christina <tamar.christina@arm.com>
12699 * config/aarch64/aarch64.md
12700 (copysignsf3): Fix mask generation.
12702 2017-06-07 Jakub Jelinek <jakub@redhat.com>
12704 * dumpfile.h (enum tree_dump_index): Rename TDI_generic to
12706 (class dump_manager): Add register_dumps method.
12707 * dumpfile.c: Include langhooks.h.
12708 (dump_files): Use 0 instead of 3/4/5 for TDI_{original,gimple,nested}.
12709 (FIRST_AUTO_NUMBERED_DUMP): Decrease to 1.
12710 (FIRST_ME_AUTO_NUMBERED_DUMP): Define.
12711 (dump_manager::dump_register): Start with 512 entries instead of 32.
12712 (dump_manager::register_dumps): New method.
12713 * toplev.c (general_init): Instead of invoking register_dumps
12714 langhook, invoke register_dumps method on the dump manager.
12715 * gimplify.c (gimplify_function_tree): Use TDI_gimple instead of
12718 2017-06-07 Richard Sandiford <richard.sandiford@linaro.org>
12720 * doc/md.texi: Clarify the restrictions on a define_insn condition.
12721 Say that # requires an associated define_split to exist, and that
12722 the define_split must be suitable for use after register allocation.
12724 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12726 * cfgbuild.c (find_bb_boundaries): Initialize profile of split blocks.
12727 (compute_outgoing_frequencies): Also initialize zero counts.
12728 (find_many_sub_basic_blocks): Do not produce uninitialized profile
12729 around loops; preserve more of profile when nothing changes.
12731 2017-06-06 Jim Wilson <jim.wilson@linaro.org>
12733 * config/aarch64/aarch64-cost-tables.h (qdf24xx_extra_costs): Move to
12735 * config/arm/aarch-cost-tables.h (qdf24xx_extra_costs): From here.
12736 * config/arm/arm-cpu-cdata.h: Regenerate.
12737 * config/arm/arm-cpu-data.h, config/arm/arm-cpu.h: Likewise.
12738 * config/arm/arm-tables.opt, config/arm/arm-tune.md: Likewise.
12739 * config/arm/arm-cpus.in: Delete falkor and qdf24xx entries.
12740 * config/arm/arm.c (arm_qdf24xx_tune): Delete.
12741 * config/arm/bpabi.h (BE8_LINK_SPEC): Delete falkor and qdf24xx
12743 * config/arm/t-aprofile (MULTILIB_MATCHES): Delete falkor and qdf24xx
12745 * config/arm/t-rmprofile: Likewise.
12746 * doc/invoke.texi (ARM Options): Drop falkor and qdf24xx support.
12748 2017-06-06 David S. Miller <davem@davemloft.net>
12751 * config/sparc/sparc.c (sparc_expand_prologue): Emit frame
12752 blockage if function uses alloca.
12754 2017-06-06 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
12756 * tree-ssa-loop-prefetch.c (struct mem_ref_group, struct mem_ref):
12757 New "uid" fields to hold pretty-print IDs of group and ref.
12758 Memory references are now identified as <group_id>:<ref_id>
12759 instead of using [random] addresses.
12760 (dump_mem_details): Simplify, no functional change.
12761 (dump_mem_ref): Simplify and make output more concise.
12762 Replace couple of fprintf's throughout code with calls to dump_mem_ref.
12763 (find_or_create_group): Initialize group uid.
12764 (record_ref): Initialize ref uid. Improve debug output.
12765 (prune_group_by_reuse, should_issue_prefetch_p,)
12766 (should_issue_prefetch_p, schedule_prefetches, issue_prefetch_ref,)
12767 (mark_nontemporal_store, determine_loop_nest_reuse):
12768 Improve debug output.
12770 2017-06-06 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
12772 * dbgcnt.def (prefetch): New debug counter.
12773 * tree-ssa-loop-prefetch.c (dbgcnt.h): New include.
12774 (schedule_prefetches): Stop issueing prefetches if debug counter
12777 2017-06-06 Tom de Vries <tom@codesourcery.com>
12779 * doc/sourcebuild.texi (Testsuites, C Language Testsuites,
12780 gcc.c-torture/compile): Remove mention of NO_LABEL_VALUES in fixme.
12782 2017-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12784 * config/aarch64/atomics.md (atomic_compare_and_swap<mode> expander):
12785 Use aarch64_reg_or_zero predicate for operand 4.
12786 (aarch64_compare_and_swap<mode> define_insn_and_split):
12787 Use aarch64_reg_or_zero predicate for operand 3. Add 'Z' constraint.
12788 (aarch64_store_exclusive<mode>): Likewise for operand 2.
12790 2017-06-06 Thomas Preud'homme <thomas.preudhomme@arm.com>
12792 * config/arm/arm.c (arm_compute_save_reg_mask): Rename into ...
12793 (arm_compute_save_core_reg_mask): This.
12794 (thumb1_compute_save_reg_mask): Rename into ...
12795 (thumb1_compute_save_core_reg_mask): This.
12796 (arm_compute_save_reg0_reg12_mask): Adapt comment.
12797 (arm_compute_frame_layout): Likewise.
12799 2017-06-06 Richard Biener <rguenther@suse.de>
12801 PR tree-optimization/80974
12802 * tree-ssa-sccvn.c (set_ssa_val_to): Do not change but only
12803 keep or clear leaders SSA info.
12805 2017-06-06 Tom de Vries <tom@codesourcery.com>
12807 * config/nvptx/nvptx.c (split_mode_p): New function.
12808 (nvptx_declare_function_name, nvptx_print_operand): Use split_mode_p.
12810 2017-06-06 Tom de Vries <tom@codesourcery.com>
12812 * config/nvptx/nvptx.c (nvptx_print_operand): Use maybe_split_mode.
12814 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12817 * tree-cfg.c (execute_fixup_cfg): Fix condition on when to rescale
12820 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12822 * shrink-wrap.c (handle_simple_exit): Update profile.
12823 (try_shrink_wrapping): Upate profile.
12825 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12827 * predict.c (tree_estimate_probability_bb): Add LOCAL_ONLY.
12828 (tree_guess_outgoing_edge_probabilities): New.
12829 * predict.h (tree_guess_outgoing_edge_probabilities): Declare.
12830 * tree-cfg.c (gimple_find_sub_bbs): Propagate profile.
12832 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12834 * ipa-split.c (split_function): Initialize return bb profile.
12836 2017-06-06 Jan Hubicka <hubicka@ucw.cz>
12838 * profile.c (compute_branch_probabilities): Also initialize
12839 EXIT_BLOCK profile.
12841 2017-06-06 Richard Biener <rguenther@suse.de>
12843 PR tree-optimization/80928
12844 * tree-vect-loop.c (vect_update_vf_for_slp): Amend dumps.
12845 (vect_analyze_loop_operations): Properly guard analysis for
12847 (vect_transform_loop): Likewise.
12848 (vect_analyze_loop_2): Also reset SLP type on PHIs.
12849 (vect_model_induction_cost): Do not cost for pure SLP.
12850 (vectorizable_induction): Pass in SLP node, implement SLP vectorization
12851 of induction in inner loop vectorization.
12852 * tree-vect-slp.c (vect_create_new_slp_node): Handle PHIs.
12853 (vect_get_and_check_slp_defs): Handle vect_induction_def.
12854 (vect_build_slp_tree): Likewise. Handle PHIs as terminating the
12856 (vect_analyze_slp_cost_1): Cost induction.
12857 (vect_detect_hybrid_slp_stmts): Handle PHIs.
12858 (vect_get_slp_vect_defs): Likewise.
12859 * tree-vect-stmts.c (vect_analyze_stmt): Handle induction.
12860 (vect_transform_stmt): Handle SLP reductions.
12861 * tree-vectorizer.h (vectorizable_induction): Adjust.
12863 2017-06-05 Michael Meissner <meissner@linux.vnet.ibm.com>
12865 * config/rs6000/rs6000.c (make_resolver_func): Update
12866 init_lowered_empty_function call.
12868 2017-06-05 Bernd Edlinger <bernd.edlinger@hotmail.de>
12870 * doc/invoke.texi: Document the -fprofile-abs-path option.
12871 * common.opt (fprofile-abs-path): New option.
12872 * gcov-io.h (gcov_write_filename): Declare.
12873 * gcov-io.c (gcov_write_filename): New function.
12874 * coverage.c (coverage_begin_function): Use gcov_write_filename.
12875 * profile.c (output_location): Likewise.
12877 2017-06-05 Jan Hubicka <hubicka@ucw.cz>
12879 * shring-wrap.c: Revert accidental commit.
12881 2017-06-05 Volker Reichelt <v.reichelt@netcologne.de>
12883 * doc/invoke.texi (-Wduplicated-branches): Add to warning list.
12885 2017-06-05 Jan Hubicka <hubicka@ucw.cz>
12887 * cfgexpand.c (expand_gimple_tailcall): Initialize profile of
12889 * ipa-inline.c (want_inline_self_recursive_call_p): Watch for missing
12890 profile in callgraph edge.
12891 * profile-count.h (apply_probability): If THIS is 0, then result is 0
12892 (apply_scale): Likewise.
12893 * tree-inline.c (copy_bb, copy_edges_for_bb, copy_cfg_body):
12894 Also scale profile when inlining function with zero profile.
12895 (initialize_cfun): Update exit block profile even when it is zero.
12896 * tree-ssa-threadupdate.c (clear_counts_path): Handle correctly case
12897 when profile is read.
12899 2017-06-05 Michael Meissner <meissner@linux.vnet.ibm.com>
12901 * config/rs6000/rs6000.c (toplevel): Include attribs.h.
12902 (CLONE_*): New constants to define the processors we can generate
12903 code for with the target_clone attribute.
12904 (rs6000_clone_map): New array to identify which clone processors
12905 the current program is running on.
12906 (TARGET_COMPARE_VERSION_PRIORITY): Define to enable the
12907 target_clone attribute.
12908 (TARGET_GENERATE_VERSION_DISPATCHER_BODY): Likewise.
12909 (TARGET_GET_FUNCTION_VERSIONS_DISPATCHER): Likewise.
12910 (TARGET_OPTION_FUNCTION_VERSIONS): Likewise.
12911 (cpu_expand_builtin): Add support for target_clone attribute.
12912 (rs6000_valid_attribute_p): Allow "default" attribute.
12913 (get_decl_name): New debug function to simplify printing the
12914 current function name in debugging statements.
12915 (rs6000_clone_priority): New functions to support the target_clone
12916 attribute, and be able to generate code to switch between ISA 2.05
12917 through ISA 3.0 (power6 through power9).
12918 (rs6000_compare_version_priority): Likewise.
12919 (rs6000_get_function_versions_dispatcher): Likewise.
12920 (make_resolver_func): Likewise.
12921 (add_condition_to_bb): Likewise.
12922 (dispatch_function_versions): Likewise.
12923 (rs6000_generate_version_dispatcher_body): Likewise.
12924 (rs6000_can_inline_p): Call get_decl_name for debugging usage.
12925 (fusion_gpr_load_p): Fix a spacing issue.
12926 * doc/extend.texi (Common Function Attributes): Document that the
12927 PowerPC supports the target_clone attribute.
12929 2017-06-05 Thomas Preud'homme <thomas.preudhomme@arm.com>
12931 * config/arm/arm.h: explain F symbol found in description of ARM
12932 register allocation in its legend.
12934 2017-06-05 Jan Hubicka <hubicka@ucw.cz>
12936 * config/mips/frame-header-opt.c: Include profile-count.h.
12937 * config/riscv/riscv.c: Include profile-count.h
12939 2017-06-05 Jan Hubicka <hubicka@ucw.cz>
12941 * tree-ssa-loop-im.c (execute_sm_if_changed): Add FLAG_BBS parameter;
12943 (sm_set_flag_if_changed): Add bbs field.
12944 (execute_sm_if_changed_flag_set): Pass BBS.
12945 (execute_sm): Update.
12947 2017-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12949 * config/aarch64/aarch64-simd.md (aarch64_store_lane0<mode>):
12952 2017-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12954 * config/aarch64/aarch64.md (sub<mode>3_compare1_imm): New define_insn.
12955 (peephole2): New peephole2 to emit the above.
12956 * config/aarch64/predicates.md (aarch64_sub_immediate): New predicate.
12958 2017-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12960 * config/aarch64/aarch64.c (define_peephole2 above
12961 *sub_<shift>_<mode>): New peephole.
12963 2017-05-23 Jan Hubicka <hubicka@ucw.cz>
12965 * config/i386/i386.c (make_resolver_func): Update.
12966 * Makefile.in: Add profile-count.h and profile-count.o
12967 * auto-profile.c (afdo_indirect_call): Update to new API.
12968 (afdo_set_bb_count): Update.
12969 (afdo_propagate_edge): Update.
12970 (afdo_propagate_circuit): Update.
12971 (afdo_calculate_branch_prob): Update.
12972 (afdo_annotate_cfg): Update.
12973 * basic-block.h: Include profile-count.h
12974 (struct edge_def): Turn count to profile_count.
12975 (struct basic_block_def): Likewie.
12976 (REG_BR_PROB_BASE): Move to profile-count.h
12977 (RDIV): Move to profile-count.h
12978 * bb-reorder.c (max_entry_count): Turn to profile_count.
12979 (find_traces): Update.
12980 (rotate_loop):Update.
12981 (connect_traces):Update.
12982 (sanitize_hot_paths):Update.
12983 * bt-load.c (migrate_btr_defs): Update.
12984 * cfg.c (RDIV): Remove.
12985 (init_flow): Use alloc_block.
12986 (alloc_block): Uninitialize count.
12987 (unchecked_make_edge): Uninitialize count.
12988 (check_bb_profile): Update.
12989 (dump_edge_info): Update.
12990 (dump_bb_info): Update.
12991 (update_bb_profile_for_threading): Update.
12992 (scale_bbs_frequencies_int): Update.
12993 (scale_bbs_frequencies_gcov_type): Update.
12994 (scale_bbs_frequencies_profile_count): New.
12995 * cfg.h (update_bb_profile_for_threading): Update.
12996 (scale_bbs_frequencies_profile_count): Declare.
12997 * cfgbuild.c (compute_outgoing_frequencies): Update.
12998 (find_many_sub_basic_blocks): Update.
12999 * cfgcleanup.c (try_forward_edges): Update.
13000 (try_crossjump_to_edge): Update.
13001 * cfgexpand.c (expand_gimple_tailcall): Update.
13002 (construct_exit_block): Update.
13003 * cfghooks.c (verify_flow_info): Update.
13004 (dump_bb_for_graph): Update.
13005 (split_edge): Update.
13006 (make_forwarder_block): Update.
13007 (duplicate_block): Update.
13008 (account_profile_record): Update.
13009 * cfgloop.c (find_subloop_latch_edge_by_profile): Update.
13010 (get_estimated_loop_iterations): Update.
13011 * cfgloopanal.c (expected_loop_iterations_unbounded): Update.
13012 (single_likely_exit): Update.
13013 * cfgloopmanip.c (scale_loop_profile): Update.
13015 (set_zero_probability): Update.
13016 (lv_adjust_loop_entry_edge): Update.
13017 * cfgrtl.c (force_nonfallthru_and_redirect): Update.
13018 (purge_dead_edges): Update.
13019 (rtl_account_profile_record): Update.
13020 * cgraph.c (cgraph_node::create): Uninitialize count.
13021 (symbol_table::create_edge): Uninitialize count.
13022 (cgraph_update_edges_for_call_stmt_node): Update.
13023 (cgraph_edge::dump_edge_flags): Update.
13024 (cgraph_node::dump): Update.
13025 (cgraph_edge::maybe_hot_p): Update.
13026 * cgraph.h: Include profile-count.h
13027 (create_clone), create_edge, create_indirect_edge): Update.
13028 (cgraph_node): Turn count to profile_count.
13029 (cgraph_edge0: Likewise.
13030 (make_speculative, clone): Update.
13031 (create_edge): Update.
13032 (init_lowered_empty_function): Update.
13033 * cgraphclones.c (cgraph_edge::clone): Update.
13034 (duplicate_thunk_for_node): Update.
13035 (cgraph_node::create_clone): Update.
13036 * cgraphunit.c (cgraph_node::analyze): Update.
13037 (cgraph_node::expand_thunk): Update.
13038 * final.c (dump_basic_block_info): Update.
13039 * gimple-streamer-in.c (input_bb): Update.
13040 * gimple-streamer-out.c (output_bb): Update.
13041 * graphite.c (print_global_statistics): Update.
13042 (print_graphite_scop_statistics): Update.
13043 * hsa-brig.c: Include basic-block.h.
13044 * hsa-dump.c: Include basic-block.h.
13045 * hsa-gen.c (T sum_slice): Update.
13046 (convert_switch_statements):Update.
13047 * hsa-regalloc.c: Include basic-block.h.
13048 * ipa-chkp.c (chkp_produce_thunks): Update.
13049 * ipa-cp.c (struct caller_statistics): Update.
13050 (init_caller_stats): Update.
13051 (gather_caller_stats): Update.
13052 (ipcp_cloning_candidate_p): Update.
13053 (good_cloning_opportunity_p): Update.
13054 (get_info_about_necessary_edges): Update.
13055 (dump_profile_updates): Update.
13056 (update_profiling_info): Update.
13057 (update_specialized_profile): Update.
13058 (perhaps_add_new_callers): Update.
13059 (decide_about_value): Update.
13060 (ipa_cp_c_finalize): Update.
13061 * ipa-devirt.c (struct odr_type_warn_count): Update.
13062 (struct decl_warn_count): Update.
13063 (struct final_warning_record): Update.
13064 (possible_polymorphic_call_targets): Update.
13065 (ipa_devirt): Update.
13066 * ipa-fnsummary.c (redirect_to_unreachable): Update.
13067 * ipa-icf.c (sem_function::merge): Update.
13068 * ipa-inline-analysis.c (do_estimate_edge_time): Update.
13069 * ipa-inline.c (compute_uninlined_call_time): Update.
13070 (compute_inlined_call_time): Update.
13071 (want_inline_small_function_p): Update.
13072 (want_inline_self_recursive_call_p): Update.
13073 (edge_badness): Update.
13074 (lookup_recursive_calls): Update.
13075 (recursive_inlining): Update.
13076 (inline_small_functions): Update.
13077 (dump_overall_stats): Update.
13078 (dump_inline_stats): Update.
13079 * ipa-profile.c (ipa_profile_generate_summary): Update.
13080 (ipa_propagate_frequency): Update.
13081 (ipa_profile): Update.
13082 * ipa-prop.c (ipa_make_edge_direct_to_target): Update.
13083 * ipa-utils.c (ipa_merge_profiles): Update.
13084 * loop-doloop.c (doloop_modify): Update.
13085 * loop-unroll.c (report_unroll): Update.
13086 (unroll_loop_runtime_iterations): Update.
13087 * lto-cgraph.c (lto_output_edge): Update.
13088 (lto_output_node): Update.
13089 (input_node): Update.
13090 (input_edge): Update.
13091 (merge_profile_summaries): Update.
13092 * lto-streamer-in.c (input_cfg): Update.
13093 * lto-streamer-out.c (output_cfg): Update.
13094 * mcf.c (create_fixup_graph): Update.
13095 (adjust_cfg_counts): Update.
13096 (sum_edge_counts): Update.
13097 * modulo-sched.c (sms_schedule): Update.
13098 * postreload-gcse.c (eliminate_partially_redundant_load): Update.
13099 * predict.c (maybe_hot_count_p): Update.
13100 (probably_never_executed): Update.
13101 (dump_prediction): Update.
13102 (combine_predictions_for_bb): Update.
13103 (propagate_freq): Update.
13104 (handle_missing_profiles): Update.
13105 (counts_to_freqs): Update.
13106 (rebuild_frequencies): Update.
13107 (force_edge_cold): Update.
13108 * predict.h: Include profile-count.h
13109 (maybe_hot_count_p, counts_to_freqs): UPdate.
13110 * print-rtl-function.c: Do not include cfg.h
13111 * print-rtl.c: Include basic-block.h
13112 * profile-count.c: New file.
13113 * profile-count.h: New file.
13114 * profile.c (is_edge_inconsistent): Update.
13115 (correct_negative_edge_counts): Update.
13116 (is_inconsistent): Update.
13117 (set_bb_counts): Update.
13118 (read_profile_edge_counts): Update.
13119 (compute_frequency_overlap): Update.
13120 (compute_branch_probabilities): Update; Initialize and deinitialize
13122 (branch_prob): Update.
13123 * profile.h (bb_gcov_counts, edge_gcov_counts): New.
13124 (edge_gcov_count): New.
13125 (bb_gcov_count): New.
13126 * shrink-wrap.c (try_shrink_wrapping): Update.
13127 * tracer.c (better_p): Update.
13128 * trans-mem.c (expand_transaction): Update.
13129 (ipa_tm_insert_irr_call): Update.
13130 (ipa_tm_insert_gettmclone_call): Update.
13131 * tree-call-cdce.c: Update.
13132 * tree-cfg.c (gimple_duplicate_sese_region): Update.
13133 (gimple_duplicate_sese_tail): Update.
13134 (gimple_account_profile_record): Update.
13135 (execute_fixup_cfg): Update.
13136 * tree-inline.c (copy_bb): Update.
13137 (copy_edges_for_bb): Update.
13138 (initialize_cfun): Update.
13139 (freqs_to_counts): Update.
13140 (copy_cfg_body): Update.
13141 (expand_call_inline): Update.
13142 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): Update.
13143 * tree-ssa-loop-ivcanon.c (unloop_loops): Update.
13144 (try_unroll_loop_completely): Update.
13145 (try_peel_loop): Update.
13146 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Update.
13147 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Update.
13148 * tree-ssa-loop-split.c (connect_loops): Update.
13149 * tree-ssa-loop-unswitch.c (hoist_guard): Update.
13150 * tree-ssa-reassoc.c (branch_fixup): Update.
13151 * tree-ssa-tail-merge.c (replace_block_by): Update.
13152 * tree-ssa-threadupdate.c (create_block_for_threading): Update.
13153 (compute_path_counts): Update.
13154 (update_profile): Update.
13155 (recompute_probabilities): Update.
13156 (update_joiner_offpath_counts): Update.
13157 (estimated_freqs_path): Update.
13158 (freqs_to_counts_path): Update.
13159 (clear_counts_path): Update.
13160 (ssa_fix_duplicate_block_edges): Update.
13161 (duplicate_thread_path): Update.
13162 * tree-switch-conversion.c (case_bit_test_cmp): Update.
13163 (struct switch_conv_info): Update.
13164 * tree-tailcall.c (decrease_profile): Update.
13165 * tree-vect-loop-manip.c (slpeel_add_loop_guard): Update.
13166 * tree-vect-loop.c (scale_profile_for_vect_loop): Update.
13167 * value-prof.c (check_counter): Update.
13168 (gimple_divmod_fixed_value): Update.
13169 (gimple_mod_pow2): Update.
13170 (gimple_mod_subtract): Update.
13171 (gimple_ic_transform): Update.
13172 (gimple_stringop_fixed_value): Update.
13173 * value-prof.h (gimple_ic): Update.
13175 2017-06-02 Carl Love <cel@us.ibm.com>
13177 * config/rs6000/rs6000-c: Add support for built-in functions
13178 vector double vec_doublee (vector signed int);
13179 vector double vec_doublee (vector unsigned int);
13180 vector double vec_doublee (vector float);
13181 vector double vec_doubleh (vector signed int);
13182 vector double vec_doubleh (vector unsigned int);
13183 vector double vec_doubleh (vector float);
13184 vector double vec_doublel (vector signed int);
13185 vector double vec_doublel (vector unsigned int);
13186 vector double vec_doublel (vector float);
13187 vector double vec_doubleo (vector signed int);
13188 vector double vec_doubleo (vector unsigned int);
13189 vector double vec_doubleo (vector float);.
13190 * config/rs6000/rs6000-builtin.def: Add definitions for DOUBLEE,
13191 DOUBLEO, DOUBLEH, DOUBLEL, UNS_DOUBLEO, UNS_DOUBLEE, UNS_DOUBLEH,
13193 * config/rs6000/altivec.md: Add code generator for doublee<mode>2,
13194 unsdoubleev4si2, doubleo<mode>2, unsdoubleov4si2, doubleh<mode>2,
13195 unsdoublehv4si2, doublel<mode>2, unsdoublelv4si2, add mode attribute
13197 * config/rs6000/altivec.h: Add define for vec_doublee, vec_doubleo,
13198 vec_doublel, vec_doubleh.
13199 * doc/extend.texi: Update the built-in documentation file for the
13200 new built-in functions.
13202 2017-06-02 David Malcolm <dmalcolm@redhat.com>
13205 * ipa-inline-analysis.c (free_growth_caches): Set
13206 edge_removal_hook_holder to NULL after removing it.
13208 2017-06-02 Sudakshina Das <sudi.das@arm.com>
13210 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
13211 comparision with zero.
13213 2017-06-02 Will Schmidt <will_schmidt@vnet.ibm.com>
13214 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling
13215 for early expansion of vec_min and vec_max builtins.
13216 (builtin_function_type): Add min/max unsigned variants to those
13217 identified as having unsigned arguments.
13219 2017-06-02 Olivier Hainque <hainque@adacore.com>
13221 * config/vx-common.h (DWARF_UNWIND_INFO): Switch #define to 1.
13223 2017-06-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13225 * config/aarch64/aarch64-simd.md (*aarch64_simd_vec_copy_lane<mode>):
13226 Use VALL_F16 iterator rather than VALL.
13228 2017-06-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13230 * config/aarch64/aarch64.c (aarch64_split_compare_and_swap):
13231 Emit CBNZ inside loop when doing a strong exchange and comparing
13232 against zero. Generate the CC flags after the loop.
13234 2017-06-02 David Edelsohn <dje.gcc@gmail.com>
13236 * dwarf2out.c (DWARF_INITIAL_LENGTH_SIZE_STR): New.
13237 (dl_section_ref): New.
13238 (dwarf2out_finish): Copy debug_line_section_label to dl_section_ref.
13239 On AIX, append an expression to subtract the size of the
13240 section length to dl_section_ref.
13242 2017-06-02 Will Schmidt <will_schmidt@vnet.ibm.com>
13244 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling
13245 for early expansion of vector absolute builtins.
13247 2017-06-02 Richard Biener <rguenther@suse.de>
13249 * tree-vect-slp.c (vect_detect_hybrid_slp_2): Match up
13250 what we consider a relevant use stmt with vect_detect_hybrid_slp_stmts.
13252 2017-06-02 Richard Biener <rguenther@suse.de>
13254 PR tree-optimization/80948
13255 * tree-tailcall.c (find_tail_calls): Track stmts to move in
13256 stmt order as well.
13258 2017-06-02 Richard Biener <rguenther@suse.de>
13260 * tree-vect-loop.c (vect_analyze_loop_operations): Not relevant
13262 * tree-vect-stmts.c (process_use): Do not mark backedge defs
13263 for inductions as relevant.
13265 2017-06-02 Richard Biener <rguenther@suse.de>
13267 * tree-vect-loop.c (get_initial_def_for_induction): Inline into ...
13268 (vectorizable_induction): ... this. Remove dead code.
13270 2017-06-02 Eric Botcazou <ebotcazou@adacore.com>
13272 * builtins. (expand_builtin_alloca): Remove second parameter and
13273 infer its value from the first parameter instead.
13274 (expand_builtin) <BUILT_IN_ALLOCA>: Adjust call to above.
13276 2017-06-02 Jakub Jelinek <jakub@redhat.com>
13278 PR rtl-optimization/80903
13279 * loop-doloop.c (add_test): Unshare sequence.
13281 2017-06-02 Bernd Edlinger <bernd.edlinger@hotmail.de>
13283 * doc/invoke.texi: Document the -Wsizeof-pointer-div warning.
13285 2017-06-01 Bernd Edlinger <bernd.edlinger@hotmail.de>
13287 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers): Make
13289 (xlogue_layout::get_stack_space_used, xlogue_layout::s_instances,
13290 xlogue_layout::get_instance, logue_layout::xlogue_layout,
13291 sp_valid_at, fp_valid_at, choose_basereg): Formatting.
13292 (xlogue_layout::get_stub_rtx): Make static.
13293 (xlogue_layout::get_stub_name): Avoid const-cast, make static.
13294 (xlogue_layout::compute_stub_managed_regs): Rename to...
13295 (xlogue_layout::count_stub_managed_regs): ...this.
13296 (xlogue_layout::is_stub_managed_reg): New function.
13297 (xlogue_layout::m_stub_names): Rename to...
13298 (xlogue_layout::s_stub_names): ...this, make static.
13299 (xlogue_layout::STUB_INDEX_OFFSET, xlogue_layout::MIN_REGS,
13300 xlogue_layout::MAX_REGS, xlogue_layout::MAX_EXTRA_REGS,
13301 xlogue_layout::VARIANT_COUNT, xlogue_layout::STUB_NAME_MAX_LEN,
13302 xlogue_layout::s_stub_names): Instantiate statics.
13303 (stub_managed_regs): Remove.
13304 (ix86_save_reg): Use xlogue_layout::compute_stub_managed_regs.
13305 (disable_call_ms2sysv_xlogues): Rename to...
13306 (warn_once_call_ms2sysv_xlogues): ...this, and warn only once.
13307 (ix86_initial_elimination_offset, ix86_expand_call): Fix call_ms2sysv
13309 (ix86_static_chain): Make sure that ix86_static_chain_on_stack can't
13310 change after reload_completed.
13311 (ix86_can_use_return_insn_p): Use the ix86_frame data structure
13313 (ix86_expand_prologue): Likewise.
13314 (ix86_expand_epilogue): Likewise.
13315 (ix86_expand_split_stack_prologue): Likewise.
13316 (ix86_compute_frame_layout): Remove frame parameter ...
13317 (TARGET_COMPUTE_FRAME_LAYOUT): ... and export it as a target hook.
13318 (ix86_finalize_stack_realign_flags): Call ix86_compute_frame_layout
13320 (ix86_init_machine_status): Don't set use_fast_prologue_epilogue_nregs.
13321 (ix86_frame): Move from here ...
13322 * config/i386/i386.h (ix86_frame): ... to here.
13323 (machine_function): Remove use_fast_prologue_epilogue_nregs, cache the
13324 complete ix86_frame data structure instead. Remove some_ld_name.
13326 2017-06-01 Pierre-Marie de Rodat <derodat@adacore.com>
13328 * dwarf2out.c (dwarf2out_late_global_decl): Add locations for
13329 symbols that hold a DECL_VALUE_EXPR.
13331 2017-06-01 Martin Jambor <mjambor@suse.cz>
13333 PR tree-optimization/80898
13334 * tree-sra.c (process_subtree_disqualification): Removed.
13335 (disqualify_candidate): Do not acll
13336 process_subtree_disqualification.
13337 (subtree_mark_written_and_enqueue): New function.
13338 (propagate_all_subaccesses): Set grp_write of LHS subtree if the
13339 RHS has been disqualified and re-queue LHS if necessary. Apart
13340 from that, ignore disqualified RHS.
13342 2017-06-01 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
13344 * config/s390/s390.c (s390_emit_epilogue): Disable early return
13345 address fetch for z10 or later.
13347 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13349 * config/arc/arc.md (tst_movb): Add guard when splitting.
13351 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13353 * config/arc/arc.c (arc_can_eliminate): Test against
13354 arc_frame_pointer_needed.
13356 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13358 * config/arc/arc.c (arc_expand_prologue): Emit a special barrier
13359 to prevent store reordering.
13360 * config/arc/arc.md (UNSPEC_ARC_STKTIE): Define.
13361 (type): Add block type.
13362 (stack_tie): Define special instruction to be used in
13365 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13367 * config/arc/arc.md (commutative_binary_comparison): Remove 'I'
13368 constraint. It is not valid for the pattern.
13369 (noncommutative_binary_comparison): Likewise.
13371 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13373 * config/arc/simdext.md (movv2hi_insn): Change predicate to avoid
13376 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13378 * config/arc/arc.c (arc_conditional_register_usage): Allow r30 to
13379 be used by the reg-alloc.
13381 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13383 * config/arc/arc.md (mulsi3): Avoid use of hard registers before
13384 reg-alloc when having mul64 or mul32x16 instructions.
13385 (mulsidi3): Likewise.
13386 (umulsidi3): Likewise.
13387 (mulsi32x16): New pattern.
13388 (mulsi64): Likewise.
13389 (mulsidi64): Likewise.
13390 (umulsidi64): Likewise.
13391 (MUL32x16_REG): Define.
13392 (mul64_600): Use MUL32x16_REG.
13393 (mac64_600): Likewise.
13394 (umul64_600): Likewise.
13395 (umac64_600): Likewise.
13397 2017-06-01 Claudiu Zissulescu <claziss@synopsys.com>
13399 * config/arc/arc.md (mulsi3_700): Make it commutative.
13401 2017-06-01 Jose E. Marchesi <jose.marchesi@oracle.com>
13403 * config/sparc/sparc.md (*zero_extendsidi2_insn_sp64): Set insn
13405 (*sign_extendsidi2_insn): Likewise for movstosw.
13407 2017-06-01 Pierre-Marie de Rodat <derodat@adacore.com>
13409 * dwarf2out.c (get_discr_value): Call the get_debug_type hook on
13410 the type of the input discriminant value. Convert the
13411 discriminant value of signedness vary.
13413 2017-06-01 Volker Reichelt <v.reichelt@netcologne.de>
13415 * doc/invoke.texi (-Wcatch-value): Document new shortcut.
13416 Add to -Wall section.
13418 2017-06-01 Richard Biener <rguenther@suse.de>
13420 PR middle-end/66313
13421 * fold-const.c (fold_plusminus_mult_expr): If the factored
13422 factor may be zero use a wrapping type for the inner operation.
13423 * tree-tailcall.c (independent_of_stmt_p): Pass in to_move bitmap
13424 and handle moved defs.
13425 (process_assignment): Properly guard the unary op case. Return a
13426 tri-state indicating that moving the stmt before the call may allow
13427 to continue. Pass through to_move.
13428 (find_tail_calls): Handle moving unrelated defs before
13431 2017-05-31 Segher Boessenkool <segher@kernel.crashing.org>
13434 * config/rs6000/vector.md (*vector_uneq<mode>): Write the nor in the
13435 splitter result in the canonical way.
13437 2017-05-31 Uros Bizjak <ubizjak@gmail.com>
13439 * config/i386/i386.md (*zero_extendsidi2): Enable alternative (?r, *Yj)
13440 also for 32bit target. Update insn attributes.
13441 (zero-extendsidi2 splitter): Allow all registers for operand 1.
13443 2017-05-31 Sebastian Peryt <sebastian.peryt@intel.com>
13445 * config/i386/avx512fintrin.h (_mm_mask_max_sd)
13446 (_mm_maskz_max_sd, _mm_mask_max_ss, _mm_maskz_max_ss)
13447 (_mm_mask_min_sd, _mm_maskz_min_sd, _mm_mask_min_ss)
13448 (_mm_maskz_min_ss): New intrinsics.
13450 2017-05-31 Martin Liska <mliska@suse.cz>
13452 * tree-vect-loop.c (vect_create_epilog_for_reduction):
13453 Change comment style to one we normally use.
13454 (vectorizable_reduction): Likewise.
13455 (vectorizable_induction): Likewise.
13456 * tree-vect-stmts.c (vectorizable_mask_load_store): Likewise.
13457 (vectorizable_call): Likewise.
13458 (vectorizable_simd_clone_call): Likewise.
13459 (vectorizable_conversion): Likewise.
13460 (vectorizable_assignment): Likewise.
13461 (vectorizable_shift): Likewise.
13462 (vectorizable_operation): Likewise.
13463 (vectorizable_store): Likewise.
13464 (vectorizable_load): Likewise.
13465 * tree-vectorizer.h: Likewise.
13467 2017-05-31 Alexander Monakov <amonakov@ispras.ru>
13469 * passes.c (emergency_dump_function): New.
13470 * tree-pass.h (emergency_dump_function): Declare.
13471 * plugin.c (plugins_internal_error_function): Remove.
13472 * plugin.h (plugins_internal_error_function): Remove declaration.
13473 * toplev.c (internal_error_function): New static function. Use it...
13474 (general_init): ...here.
13476 2017-05-31 Graham Markall <graham.markall@embecosm.com>
13478 * config/arc/arc.c (arc_print_operand): Handle constant operands.
13479 (arc_rtx_costs): Add costs for new patterns.
13480 * config/arc/arc.md: Additional *add_n and *sub_n patterns.
13481 * config/arc/predicates.md: Add _1_2_3_operand predicate.
13483 2017-05-31 Richard Sandiford <richard.sandiford@linaro.org>
13485 * tree-ssa-strlen.c (get_next_strinfo): New function.
13486 (get_stridx_plus_constant): Use it.
13487 (zero_length_string): Likewise.
13488 (adjust_related_strinfos): Likewise.
13489 (adjust_last_stmt): Likewise.
13491 2017-05-31 Richard Biener <rguenther@suse.de>
13494 * config/i386/i386.c (ix86_expand_builtin): Remove assert
13495 for arg being an SSA name when expanding IX86_BUILTIN_BNDRET.
13497 2017-05-31 Richard Sandiford <richard.sandiford@linaro.org>
13499 * tree-vect-data-refs.c (vect_find_same_alignment_drs): Remove
13500 loop_vinfo argument and use of dependence distance vectors.
13501 Check instead whether the two references differ only in their
13502 initial value and assume that they have the same alignment if the
13503 difference is a multiple of the vector alignment.
13504 (vect_analyze_data_refs_alignment): Update call accordingly.
13506 2017-05-31 Martin Liska <mliska@suse.cz>
13509 * config/i386/cpuid.h: Fix typo in a comment in cpuid.h.
13511 2017-05-31 Bin Cheng <bin.cheng@arm.com>
13513 * tree-vect-loop-manip.c (create_intersect_range_checks_index)
13514 (create_intersect_range_checks): Move from ...
13515 * tree-data-ref.c (create_intersect_range_checks_index)
13516 (create_intersect_range_checks): ... to here.
13517 (create_runtime_alias_checks): New function factored from ...
13518 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks): ...
13519 here. Call above function.
13520 * tree-data-ref.h (create_runtime_alias_checks): New function.
13522 2017-05-31 Bin Cheng <bin.cheng@arm.com>
13524 * tree-data-ref.c (prune_runtime_alias_test_list): Relax minimal
13525 segment length for dr_b and compute it in wide_int.
13527 2017-05-31 Richard Biener <rguenther@suse.de>
13529 PR tree-optimization/80906
13530 * graphite-isl-ast-to-gimple.c (copy_loop_close_phi_nodes): Get
13531 and pass through iv_map.
13532 (copy_bb_and_scalar_dependences): Adjust.
13533 (translate_pending_phi_nodes): Likewise.
13534 (copy_loop_close_phi_args): Handle code-generating IVs instead
13537 2017-05-30 David Malcolm <dmalcolm@redhat.com>
13539 * diagnostic-color.c (color_dict): Add "type-diff".
13540 (parse_gcc_colors): Update comment.
13541 * doc/invoke.texi (Diagnostic Message Formatting Options): Add
13542 -fdiagnostics-show-template-tree and -fno-elide-type.
13543 (GCC_COLORS): Add type-diff to example.
13545 (-fdiagnostics-show-template-tree): New.
13546 (-fno-elide-type): New.
13547 * pretty-print.c (pp_format): Pass quote and formatters[argno] to
13548 the pp_format_decoder callback. Call any m_format_postprocessor's
13550 (pretty_printer::pretty_printer): Initialize
13551 m_format_postprocessor.
13552 (pretty_printer::~pretty_printer): Delete any
13553 m_format_postprocessor.
13554 * pretty-print.h (printer_fn): Add bool and const char ** parameters.
13555 (class format_postprocessor): New class.
13556 (struct pretty_printer::format_decoder): Document the new parameters.
13557 (struct pretty_printer::m_format_postprocessor): New field.
13558 * tree-diagnostic.c (default_tree_printer): Update for new
13559 bool and const char ** params.
13560 * tree-diagnostic.h (default_tree_printer): Likewise.
13562 2017-05-30 Segher Boessenkool <segher@kernel.crashing.org>
13564 * config/rs6000/predicates.md (cc_reg_not_micro_cr0_operand): Delete.
13565 (lwa_operand): Delete rs6000_gen_cell_microcode test.
13566 * config/rs6000/rs6000.c (rs6000_option_override_internal): Delete
13567 rs6000_gen_cell_microcode code.
13568 (rs6000_final_prescan_insn): Delete.
13569 (rs6000_opt_vars): Delete the "gen-cell-microcode" and
13570 "warn-cell-microcode" entries.
13571 * config/rs6000/rs6000.h (FINAL_PRESCAN_INSN): Delete.
13572 * config/rs6000/rs6000.md: Delete rs6000_gen_cell_microcode tests
13573 throughout. Change cc_reg_not_micro_cr0_operand to
13574 cc_reg_not_cr0_operand throughout.
13575 (*extendhi<mode>2_noload): Delete.
13576 * config/rs6000/rs6000.opt (mgen-cell-microcode): Replace by stub.
13577 (mwarn-cell-microcode): Delete.
13578 * doc/invoke.texi (RS/6000 and PowerPC Options): Delete
13579 -mgen-cell-microcode and -mwarn-cell-microcode.
13581 2017-05-30 Uros Bizjak <ubizjak@gmail.com>
13584 * config/i386/constraints.md (Yd): New constraint.
13586 * config/i386/i386.md (*movti_internal): Add (?r, Ye)
13587 and (?Yd, r) alternatives. Update insn attributes.
13588 * config/i386/i386.md (*movti_internal): Add (?r, *Ye)
13589 and (?*Yd, r) alternatives. Update insn attributes.
13590 (double-mode inter-unit splitters): Add new GR<->XMM splitters.
13592 2017-05-30 Pierre-Marie de Rodat <derodat@adacore.com>
13594 * gimplify.c (gimplify_modify_expr): Don't create a
13595 DECL_DEBUG_EXPR link if *FROM_P does not belong to the current
13598 2017-05-30 Wilco Dijkstra <wdijkstr@arm.com>
13600 * config/arm/arm-builtins.c (arm_expand_builtin): Remove const.
13602 2017-05-30 Richard Biener <rguenther@suse.de>
13604 * tree-vectorizer.h (struct _stmt_vec_info): Add reduc_type
13605 and reduc_def fields.
13606 (STMT_VINFO_REDUC_TYPE): New define.
13607 (STMT_VINFO_REDUC_DEF): Likewise.
13608 (vect_force_simple_reduction): Adjust prototype.
13609 * tree-vect-loop.c (vect_analyze_scalar_cycles_1): Adjust.
13610 (vect_is_simple_reduction): Remove check_reduction argument.
13611 (vect_force_simple_reduction): Adjust and set
13612 STMT_VINFO_REDUC_TYPE and STMT_VINFO_REDUC_DEF.
13613 (vectorizable_reduction): Do not re-do reduction analysis
13614 but use STMT_VINFO_REDUC_TYPE and STMT_VINFO_REDUC_DEF.
13615 * tree-parloops.c (gather_scalar_reductions): Adjust.
13617 2017-05-30 Richard Biener <rguenther@suse.de>
13619 PR middle-end/80901
13620 * cfgexpand.c (expand_gimple_cond): Match up loop fixup with
13623 2017-05-24 Robin Dapp <rdapp@linux.vnet.ibm.com>
13625 * tree-vect-data-refs.c (vect_get_peeling_costs_all_drs):
13626 Introduce unknown_misalignment parameter and remove vf.
13627 (vect_peeling_hash_get_lowest_cost):
13628 Pass unknown_misalignment parameter.
13629 (vect_enhance_data_refs_alignment):
13630 Fix unsupportable data ref treatment.
13632 2017-05-30 Robin Dapp <rdapp@linux.vnet.ibm.com>
13634 * tree-vect-data-refs.c (vect_get_data_access_cost):
13635 Workaround for SLP handling.
13636 (vect_enhance_data_refs_alignment):
13637 Compute costs for doing no peeling at all, compare to the best
13638 peeling costs so far and avoid peeling if cheaper.
13640 2017-05-30 Robin Dapp <rdapp@linux.vnet.ibm.com>
13642 * tree-vect-data-refs.c (vect_peeling_hash_choose_best_peeling):
13643 Return peeling info and set costs to zero for unlimited cost
13645 (vect_enhance_data_refs_alignment): Also inspect all datarefs
13646 with unknown misalignment. Compute and costs for unknown
13647 misalignment, compare them to the costs for known misalignment
13648 and choose the cheapest for peeling.
13650 2017-05-30 Robin Dapp <rdapp@linux.vnet.ibm.com>
13652 * tree-vect-data-refs.c (vect_update_misalignment_for_peel): Rename.
13653 (vect_get_peeling_costs_all_drs): Create function.
13654 (vect_peeling_hash_get_lowest_cost):
13655 Use vect_get_peeling_costs_all_drs.
13656 (vect_peeling_supportable): Create function.
13657 (vect_enhance_data_refs_alignment): Use vect_peeling_supportable.
13659 2017-05-30 Robin Dapp <rdapp@linux.vnet.ibm.com>
13661 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Create
13662 DR_HAS_NEGATIVE_STEP.
13663 (vect_update_misalignment_for_peel): Define DR_MISALIGNMENT.
13664 (vect_enhance_data_refs_alignment): Use.
13665 (vect_duplicate_ssa_name_ptr_info): Use.
13666 * tree-vectorizer.h (dr_misalignment): Use.
13667 (known_alignment_for_access_p): Use.
13669 2017-05-30 Jozef Lawrynowicz <jozef.l@somniumtech.com>
13672 * config/msp430/msp430.c (gen_prefix): Return NULL when section name is
13674 (has_section_name): New function.
13676 2017-05-30 Martin Liska <mliska@suse.cz>
13679 * auto-profile.c (get_function_decl_from_block): Fix
13682 2017-05-30 Richard Biener <rguenther@suse.de>
13684 PR middle-end/80876
13685 * cfgexpand.c (expand_gimple_cond): Fixup preserving loops again.
13687 2017-05-30 Martin Liska <mliska@suse.cz>
13689 * dumpfile.c: Use newly added macro DUMP_FILE_INFO.
13690 * dumpfile.h (struct dump_file_info): Remove ctors.
13692 2017-05-30 Martin Liska <mliska@suse.cz>
13694 * predict.def: Fix GNU coding style.
13696 2017-05-29 Max Filippov <jcmvbkbc@gmail.com>
13698 * config/xtensa/xtensa.c (xtensa_initial_elimination_offset):
13699 Mark 'to' argument with ATTRIBUTE_UNUSED.
13701 2017-05-29 Max Filippov <jcmvbkbc@gmail.com>
13703 * config/xtensa/xtensa.c (xtensa_emit_call): Use
13704 HOST_WIDE_INT_PRINT_HEX instead of 0x%lx format string.
13705 (print_operand): Use HOST_WIDE_INT_PRINT_DEC instead of %ld
13708 2017-05-29 Eric Botcazou <ebotcazou@adacore.com>
13710 * doc/install.texi (Options specification): Restore entry of
13711 --enable-sjlj-exceptions.
13713 2017-05-27 Michael Eager <eager@eagercon.com>
13716 2016-01-21 Ajit Agarwal <ajitkum@xilinx.com>
13718 See https://gcc.gnu.org/ml/gcc/2017-05/msg00221.html.
13720 * config/microblaze/microblaze.h
13721 (FIXED_REGISTERS): Update in macro.
13722 (CALL_USED_REGISTERS): Update in macro.
13724 2017-05-27 François-Xavier Coudett <fxcoudert@gcc.gnu.org>
13726 * doc/install.texi: Add links to macOS binary distributions.
13728 2017-05-27 Jakub Jelinek <jakub@redhat.com>
13732 2017-05-25 Marc Glisse <marc.glisse@inria.fr>
13734 * match.pd ((A +- CST1) +- CST2): Allow some conversions.
13736 2017-05-26 Martin Liska <mliska@suse.cz>
13738 * dumpfile.h (enum dump_kind): Renumber TDF_* flags to be contiguous.
13740 2017-05-26 Martin Liska <mliska@suse.cz>
13742 * cfg.c (check_bb_profile): Do not use TDF_COMMENT and print
13743 always leading ';; '.
13744 (dump_bb_info): Likewise.
13745 (brief_dump_cfg): Likewise.
13746 * cfgrtl.c (print_rtl_with_bb): Do not use TDF_COMMENT.
13747 * dumpfile.c: Remove usage of TDF_VERBOSE.
13748 * dumpfile.h (enum dump_kind): Likewise.
13749 (dump_gimple_bb_header): Do not use TDF_COMMENT.
13750 * print-tree.c (debug_verbose): Remove.
13751 * tree-cfg.c (gimple_dump_cfg): Do not use TDF_COMMENT.
13752 (dump_function_to_file): Remove dumps guarded with TDF_VERBOSE.
13753 * tree-diagnostic.c (default_tree_printer): Replace
13754 TDF_DIAGNOSTIC with TDF_SLIM.
13756 2017-05-26 Bin Cheng <bin.cheng@arm.com>
13758 * tree-vect-loop-manip.c (create_intersect_range_checks_index): Pass
13759 in parameter loop, rather than loop_vinfo.
13760 (create_intersect_range_checks): Ditto.
13761 (vect_create_cond_for_alias_checks): Update call to above functions.
13763 2017-05-26 Bin Cheng <bin.cheng@arm.com>
13765 PR tree-optimization/80815
13766 * tree-data-ref.c (prune_runtime_alias_test_list): Simplify condition
13767 for merging runtime alias checks. Handle negative DR_STEPs.
13769 2017-05-26 Bin Cheng <bin.cheng@arm.com>
13771 * tree-vect-data-refs.c (Operator==, comp_dr_with_seg_len_pair):
13773 * tree-data-ref.c (Operator==, comp_dr_with_seg_len_pair): To here.
13774 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Factor
13775 out code pruning runtime alias checks.
13776 * tree-data-ref.c (prune_runtime_alias_test_list): New function
13777 factored out from above.
13778 * tree-vectorizer.h (struct dr_with_seg_len, dr_with_seg_len_pair_t):
13780 * tree-data-ref.h (struct dr_with_seg_len, dr_with_seg_len_pair_t):
13782 (prune_runtime_alias_test_list): New decalaration.
13784 2017-05-26 Bin Cheng <bin.cheng@arm.com>
13786 * tree-vect-data-refs.c (compare_tree): Rename and move ...
13787 * tree-data-ref.c (data_ref_compare_tree): ... to here.
13788 * tree-data-ref.h (data_ref_compare_tree): New decalaration.
13789 * tree-vect-data-refs.c (dr_group_sort_cmp): Update uses.
13790 (operator==, comp_dr_with_seg_len_pair): Ditto.
13791 (vect_prune_runtime_alias_test_list): Ditto.
13793 2017-05-26 Martin Liska <mliska@suse.cz>
13796 * params.def: Bound partial-inlining-entry-probability param.
13798 2017-05-26 Marek Polacek <polacek@redhat.com>
13801 * fold-const.c (fold_binary_loc) <case MULT_EXPR>: Check if OP1
13804 2017-05-26 Richard Biener <rguenther@suse.de>
13806 PR tree-optimization/80842
13807 * tree-ssa-ccp.c (set_lattice_value): Always meet with the old
13810 2017-05-26 Richard Biener <rguenther@suse.de>
13812 PR tree-optimization/80844
13813 * tree-vectorizer.c (adjust_simduid_builtins): Propagate results.
13815 2017-05-25 Sebastian Peryt <sebastian.peryt@intel.com>
13817 * doc/md.texi (Machine Constraints): Update x86 family
13818 machine constraints section to match 'config/i386/constraints.md'.
13820 2017-05-25 Volker Reichelt <v.reichelt@netcologne.de>
13822 * doc/invoke.texi (-Wcatch-value=): Document new warning option.
13824 2017-05-25 Nathan Sidwell <nathan@acm.org>
13826 * doc/invoke.texi (--enable-languages): Update documentation.
13828 2017-05-25 Martin Liska <mliska@suse.cz>
13830 * dumpfile.c: Add TDF_FOLDING.
13831 * dumpfile.h (enum dump_kind): Likewise.
13832 * genmatch.c (dt_simplify::gen_1): Use it.
13834 2017-05-25 Marc Glisse <marc.glisse@inria.fr>
13836 * match.pd (view_convert (convert@0 @1)): Handle zero-extension.
13838 2017-05-25 Marc Glisse <marc.glisse@inria.fr>
13840 * match.pd ((A +- CST1) +- CST2): Allow some conversions.
13841 * tree.c (drop_tree_overflow): Handle COMPLEX_CST and VECTOR_CST.
13843 2017-05-25 Marc Glisse <marc.glisse@inria.fr>
13845 * fold-const.c (fold_binary_loc) [(A & C) == D]: Remove transformation.
13846 * match.pd (X == C): Rewrite it here.
13847 (with_possible_nonzero_bits, with_possible_nonzero_bits2,
13848 with_certain_nonzero_bits2): New predicates.
13849 * tree-ssanames.c (get_nonzero_bits): Handle INTEGER_CST.
13851 2017-05-24 Nathan Sidwell <nathan@acm.org>
13853 * lto-streamer-in.c (lto_input_data_block): Adjust T const cast to
13856 * auto-profile.c (afdo_propagate): Adjust T const cast to avoid
13859 2017-05-24 Segher Boessenkool <segher@kernel.crashing.org>
13861 * config/powerpcspe: New port. Files are copied from the rs6000
13862 port, with "rs6000" in filenames replaced by "powerpcspe".
13864 2017-05-24 Wilco Dijkstra <wdijkstr@arm.com>
13866 PR rtl-optimization/80754
13867 * lra-remat.c (do_remat): Add overlap checks for dst_regno.
13869 2017-05-24 Sheldon Lobo <smlobo@sheldon.us.oracle.com>
13871 * config/sparc/sparc.md (length): Return the correct value for -mflat
13872 sibcalls to match output_sibcall.
13874 2017-05-24 Segher Boessenkool <segher@kernel.crashing.org>
13878 * config/rs6000/rs6000.c (struct machine_function): Add new field
13880 (rs6000_get_separate_components): Init that field, use it.
13881 (rs6000_components_for_bb): Use the field.
13883 2017-05-24 Thomas Preud'homme <thomas.preudhomme@arm.com>
13885 * config/arm/arm.c (arm_expand_prologue): Fix typo in comment.
13887 2017-05-24 Peter Bergner <bergner@vnet.ibm.com>
13889 PR middle-end/80823
13890 * tree-cfg.c (group_case_labels_stmt): Delete increment of "i";
13892 2017-05-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
13895 * config/s390/s390.c (s390_check_qrst_address): Check incoming
13896 address against address_operand predicate.
13897 * config/s390/s390.md ("*indirect_jump"): Swap alternatives.
13899 2017-05-24 Eric Botcazou <ebotcazou@adacore.com>
13901 * var-tracking.c (track_expr_p): Do not return 0 for tracked record
13902 parameters passed indirectly.
13904 2017-05-23 Uros Bizjak <ubizjak@gmail.com>
13906 * config/i386/i386.md (*movdi_internal): Remove SSE4
13907 alternative 18 (?r, *v). Update insn attributes.
13908 (*movsi_internal): Remove SSE4 alternative 13 (?r, *v).
13909 Update insn attributes.
13910 (*zero_extendsidi2): Remove SSE4 alternative (?r, *x).
13911 Update insn attributes.
13912 * config/i386/sse.md (vec_extract<ssevecmodelower>_0): Remove SSE4
13913 alternative 1 (r, v). Remove isa attribute.
13914 * config/i386/i386.c (dimode_scalar_chain::make_vector_copies):
13915 Always move value through stack for !TARGET_INTER_UNIT_MOVES_TO_VEC
13916 and !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13918 2017-05-23 Tom de Vries <tom@codesourcery.com>
13920 * doc/sourcebuild.texi (Directives, Verify compiler message): Document
13923 2017-05-23 Jan Hubicka <hubicka@ucw.cz>
13925 * cgraphunit.c (symbol_table::process_new_functions): Update.
13926 * ipa-fnsummary.c (pass_data_inline_parameters): Remove.
13927 (inline_generate_summary): Rename to ...
13928 (ipa_fn_summary_generate): ... this one.
13929 (inline_read_summary): Rename to ...
13930 (ipa_fn_summary_read): ... this one.
13931 (inline_write_summary): Rename to ...
13932 (ipa_fn_summary_write): ... this one.
13933 (inline_free_summary): Rename to ...
13934 (ipa_free_fn_summary): ... this one.
13935 (pass_data_local_fn_summary, pass_local_fn_summary,
13936 make_pass_local_fn_summary, pass_data_ipa_free_fn_summary,
13937 pass_ipa_free_fn_summary, make_pass_ipa_free_fn_summary,
13938 pass_data_ipa_fn_summary, pass_ipa_fn_summary,
13939 make_pass_ipa_fn_summary): New.
13940 * ipa-fnsummary.h (inline_generate_summary, inline_read_summary,
13941 inline_write_summary, inline_free_summary): Remove.
13942 (ipa_free_fn_summary) : New.
13943 * ipa-inline.c (ipa_inline): Update.
13944 (pass_ipa_inline): Do not generate summaries.
13945 * ipa.c (pass_data_ipa_free_fn_summary, pass_ipa_free_fn_summary):
13947 * passes.def: Replace pass_inline_parameters by pass_local_fn_summary
13948 and add pass_ipa_fn_summary.
13949 * tree-pass.h (make_pass_ipa_fn_summary, make_pass_local_fn_summary):
13951 (make_pass_inline_parameters): Remove.
13953 2017-05-23 Thomas Schwinge <thomas@codesourcery.com>
13955 * omp-low.c (struct omp_context): Remove "default_kind" member.
13958 * omp-offload.c (execute_oacc_device_lower): Remove the
13959 parallelism dimensions function attributes for unparallelized
13960 OpenACC kernels constructs.
13962 2017-05-23 Martin Liska <mliska@suse.cz>
13964 * cgraph.c (cgraph_node::get_create): Use symtab_node::dump_{asm_,}name
13966 (cgraph_edge::make_speculative): Likewise.
13967 (cgraph_edge::resolve_speculation): Likewise.
13968 (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
13969 (cgraph_node::dump): Likewise.
13970 * cgraph.h: Likewise.
13971 * cgraphunit.c (analyze_functions): Likewise.
13972 (symbol_table::compile): Likewise.
13973 * ipa-cp.c (print_all_lattices): Likewise.
13974 (determine_versionability): Likewise.
13975 (initialize_node_lattices): Likewise.
13976 (ipcp_verify_propagated_values): Likewise.
13977 (estimate_local_effects): Likewise.
13978 (update_profiling_info): Likewise.
13979 (create_specialized_node): Likewise.
13980 (perhaps_add_new_callers): Likewise.
13981 (decide_about_value): Likewise.
13982 (decide_whether_version_node): Likewise.
13983 (identify_dead_nodes): Likewise.
13984 (ipcp_store_bits_results): Likewise.
13985 * ipa-devirt.c (dump_targets): Likewise.
13986 (ipa_devirt): Likewise.
13987 * ipa-icf.c (sem_item::dump): Likewise.
13988 (sem_function::equals): Likewise.
13989 (sem_variable::equals): Likewise.
13990 (sem_item_optimizer::read_section): Likewise.
13991 (sem_item_optimizer::execute): Likewise.
13992 (congruence_class::dump): Likewise.
13993 * ipa-inline-analysis.c (dump_ipa_call_summary): Likewise.
13994 (dump_inline_summary): Likewise.
13995 (estimate_node_size_and_time): Likewise.
13996 (inline_analyze_function): Likewise.
13997 * ipa-inline-transform.c (inline_call): Likewise.
13998 * ipa-inline.c (report_inline_failed_reason): Likewise.
13999 (want_early_inline_function_p): Likewise.
14000 (edge_badness): Likewise.
14001 (update_edge_key): Likewise.
14002 (inline_small_functions): Likewise.
14003 * ipa-profile.c (ipa_profile): Likewise.
14004 * ipa-prop.c (ipa_print_node_jump_functions): Likewise.
14005 (ipa_make_edge_direct_to_target): Likewise.
14006 (remove_described_reference): Likewise.
14007 (ipa_impossible_devirt_target): Likewise.
14008 (propagate_controlled_uses): Likewise.
14009 (ipa_print_node_params): Likewise.
14010 (ipcp_transform_function): Likewise.
14011 * ipa-pure-const.c (pure_const_read_summary): Likewise.
14012 (propagate_pure_const): Likewise.
14013 * ipa-reference.c (generate_summary): Likewise.
14014 (read_write_all_from_decl): Likewise.
14015 (propagate): Likewise.
14016 (ipa_reference_read_optimization_summary): Likewise.
14017 * ipa-utils.c (ipa_merge_profiles): Likewise.
14018 * ipa.c (walk_polymorphic_call_targets): Likewise.
14019 (symbol_table::remove_unreachable_nodes): Likewise.
14020 (ipa_single_use): Likewise.
14021 * passes.c (execute_todo): Likewise.
14022 * predict.c (drop_profile): Likewise.
14023 * symtab.c (symtab_node::get_dump_name): New function.
14024 (symtab_node::dump_name): Likewise.
14025 (symtab_node::dump_asm_name): Likewise.
14026 (symtab_node::dump_references): Likewise.
14027 (symtab_node::dump_referring): Likewise.
14028 (symtab_node::dump_base): Likewise.
14029 (symtab_node::debug_symtab): Likewise.
14030 * tree-sra.c (convert_callers_for_node): Likewise.
14031 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
14032 * value-prof.c (init_node_map): Likewise.
14034 2017-05-23 Martin Liska <mliska@suse.cz>
14036 * cgraph.h: Move symtab_node::dump_table to symbol_table::dump
14037 and symtab_node::debug_symtab to symbol_table::debug.
14038 * cgraphunit.c (analyze_functions): Use the renamed function.
14039 (symbol_table::compile): Likewise.
14040 * ipa-cp.c (ipcp_verify_propagated_values): Likewise.
14041 * ipa-icf.c (sem_item_optimizer::execute): Likewise.
14042 * passes.c (execute_todo): Likewise.
14043 * symtab.c (symbol_table::dump): New function.
14044 * tree-ssa-structalias.c (ipa_pta_execute): Use the renamed function.
14046 2017-05-23 Jan Hubicka <hubicka@ucw.cz>
14048 * ipa-fnsummary.c (estimate_node_size_and_time): Do not sanity check
14049 that nonconst implies exec.
14051 2017-05-23 Jan Hubicka <hubicka@ucw.cz>
14053 * ipa-inline-analysis.c (cgraph_2edge_hook_list, cgraph_edge_hook_list,
14054 inline_edge_removal_hook, inline_edge_duplication_hook): Remove.
14055 (inline_edge_summary_vec): Turn into ...
14056 (ipa_call_summaries): ... this one.
14057 (redirect_to_unreachable, edge_set_predicate,
14058 evaluate_properties_for_edge, inline_summary_alloc,
14059 reset_ipa_call_summary, reset_inline_summary,
14060 inline_summary_t::duplicate): Update.
14061 (inline_edge_duplication_hook): Turn to ...
14062 (ipa_call_summary_t::duplicate): ... this one.
14063 (inline_edge_removal_hook): Turn to ...
14064 (ipa_call_summary_t::remove): ... this one.
14065 (dump_inline_edge_summary): Turn to ...
14066 (dump_ipa_call_summary): ... this one.
14067 (estimate_function_body_sizes): Update.
14068 (inline_update_callee_summaries): Update.
14069 (remap_edge_change_prob): Update.
14070 (remap_edge_summaries): Update.
14071 (inline_merge_summary): Update.
14072 (do_estimate_edge_time): Update.
14073 (inline_generate_summary): Update.
14074 (inline_read_section): Update.
14075 (inline_read_summary): Update.
14076 (inline_free_summary): Update.
14077 * ipa-inline.c (can_inline_edge_p): Update.
14078 (compute_inlined_call_time): Update.
14079 (want_inline_small_function_p): Update.
14080 (edge_badness): Update.
14081 (early_inliner): Update.
14082 * ipa-inline.h (inline_edge_summary): Turn to ...
14083 (ipa_call_summary): ... this one.
14084 (ipa_call_summary_t): New class.
14085 (inline_edge_summary_t, inline_edge_summary_vec): Remove.
14086 (ipa_call_summaries): New.
14087 (inline_edge_summary): Remove.
14088 (estimate_edge_growth): Update.
14089 * ipa-profile.c (ipa_propagate_frequency_1): Update.
14090 * ipa-prop.c (ipa_make_edge_direct_to_target): Update.
14091 * ipa-split.c (execute_split_functions): Update.
14092 * ipa.c (symbol_table::remove_unreachable_nodes): Update.
14094 2017-05-23 Tom de Vries <tom@codesourcery.com>
14096 * doc/sourcebuild.texi (Effective-Target Keywords, Other hardware
14097 attributes): Document rdrand effective target.
14099 2017-05-23 Tom de Vries <tom@codesourcery.com>
14101 * doc/sourcebuild.texi (Effective-Target Keywords, Other hardware
14102 attributes): Sort alphabetically.
14104 2017-05-23 Georg-Johann Lay <avr@gjlay.de>
14106 * config/avr/genmultilib.awk: Use gsub instead of gensub.
14108 2017-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14111 * config/rs6000/vsx.md (vsx_splat_<mode>, VSX_D iterator): Split
14112 V2DF/V2DI splat into two separate patterns, one that handles
14113 registers, and the other that only handles memory. Drop support
14114 for splatting from a GPR on ISA 2.07 and then splitting the
14115 splat into direct move and splat.
14116 (vsx_splat_<mode>_reg): Likewise.
14117 (vsx_splat_<mode>_mem): Likewise.
14119 2017-05-22 Segher Boessenkool <segher@kernel.crashing.org>
14121 * cfgcleanup.c (bb_is_just_return): Allow CLOBBERs.
14123 2017-05-22 Jakub Jelinek <jakub@redhat.com>
14125 PR middle-end/80809
14126 * omp-low.c (finish_taskreg_remap): New function.
14127 (finish_taskreg_scan): If unit size of ctx->record_type
14128 is non-constant, unshare the size expression and replace
14129 decls in it with possible outer var refs.
14131 PR middle-end/80809
14132 * gimplify.c (omp_add_variable): For GOVD_DEBUG_PRIVATE use
14133 GOVD_SHARED rather than GOVD_PRIVATE with it.
14134 (gimplify_adjust_omp_clauses_1, gimplify_adjust_omp_clauses): Expect
14135 GOVD_SHARED rather than GOVD_PRIVATE with GOVD_DEBUG_PRIVATE.
14137 PR middle-end/80853
14138 * omp-low.c (lower_reduction_clauses): Pass OMP_CLAUSE_PRIVATE
14139 as last argument to build_outer_var_ref for pointer bases of array
14140 section reductions.
14142 2017-05-19 Martin Sebor <msebor@redhat.com>
14144 * print-tree.c (print_node): Print DECL_READ_P flag.
14146 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
14148 * Makefile.in: Add ipa-fnsummary.o and ipa-fnsummary.h
14149 * auto-profile.c: Replace ipa-inline.h by ipa-fnsummary.h
14150 * cgraph.c: Likewise.
14151 * cgraphunit.c: Likewise.
14152 * gengtype.c: Likewise.
14153 * ipa-cp.c: Likewise.
14154 * ipa-devirt.c: Likewise.
14155 * ipa-icf.c: Likewise.
14156 * ipa-predicate.c: Likewise.
14157 * ipa-profile.c: Likewise.
14158 * ipa-prop.c: Likewise.
14159 * ipa-split.c: Likewise.
14161 * ipa-inline-analysis.c (inline_summaries, ipa_call_summaries,
14162 edge_predicate_pool, dump_inline_hints,
14163 inline_summary::account_size_time, redirect_to_unreachable,
14164 edge_set_predicate, set_hint_predicate,
14165 evaluate_conditions_for_known_args, evaluate_properties_for_edge,
14166 inline_summary_alloc, ipa_call_summary::reset, inline_summary::reset,
14167 inline_summary_t::remove, remap_hint_predicate_after_duplication,
14168 inline_summary_t::duplicate, ipa_call_summary_t::duplicate,
14169 ipa_call_summary_t::remove, initialize_growth_caches,
14170 free_growth_caches, dump_ipa_call_summary, dump_inline_summary,
14171 debug_inline_summary, dump_inline_summaries, initialize_inline_failed,
14172 mark_modified, unmodified_parm_1, unmodified_parm,
14173 unmodified_parm_or_parm_agg_item, eliminated_by_inlining_prob,
14174 set_cond_stmt_execution_predicate, set_switch_stmt_execution_predicate,
14175 compute_bb_predicates, will_be_nonconstant_expr_predicate,
14176 will_be_nonconstant_predicate, record_modified_bb_info,
14177 get_minimal_bb, record_modified, param_change_prob,
14178 phi_result_unknown_predicate, predicate_for_phi_result,
14179 array_index_predicate, clobber_only_eh_bb_p, fp_expression_p,
14180 estimate_function_body_sizes, compute_inline_parameters,
14181 compute_inline_parameters_for_curren, pass_data_inline_parameters,
14182 estimate_node_size_and_time, estimate_ipcp_clone_size_and_time,
14183 inline_update_callee_summaries, remap_edge_change_prob,
14184 remap_edge_summaries, remap_hint_predicate, inline_merge_summary,
14185 inline_update_overall_summary, inline_indirect_intraprocedural_analysis,
14186 inline_analyze_function, inline_summary_t::insert,
14187 inline_generate_summary, read_ipa_call_summary, inline_read_section,
14188 inline_read_summary, write_ipa_call_summary, inline_write_summary,
14189 inline_free_summary): Move to ipa-fnsummary.h
14190 (predicate_t): Remove.
14191 * ipa-fnsummary.c: New file.
14192 * ipa-inline.h: Do not include sreal.h and ipa-predicate.h
14193 (enum inline_hints_vals, inline_hints, agg_position_info,
14194 INLINE_SIZE_SCALE, size_time_entry, inline_summary, inline_summary_t,
14195 inline_summaries, ipa_call_summary, ipa_call_summary_t,
14196 ipa_call_summaries, debug_inline_summary, dump_inline_summaries,
14197 dump_inline_summary, dump_inline_hints, inline_generate_summary,
14198 inline_read_summary, inline_write_summary, inline_free_summary,
14199 inline_analyze_function, initialize_inline_failed,
14200 inline_merge_summary, inline_update_overall_summary,
14201 compute_inline_parameters): Move to ipa-fnsummary.h
14202 * ipa-fnsummary.h: New file.
14203 * ipa-inline-transform.h: Include ipa-inline.h.
14204 * ipa-inline.c: LIkewise.
14206 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
14208 * ipa-inline.c (edge_badness): Use inlined_time instead of
14209 inline_summaries->get.
14211 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
14213 * ipa-inline.c (edge_badness): Use estimate_size_after_inlining.
14215 2017-05-22 Nathan Sidwell <nathan@acm.org>
14217 * doc/invoke.texi (fdump-translation-unit): Delete documentation.
14218 (fdump-lang): Document 'raw' option.
14219 * dumpfile.h (TDI_tu): Delete.
14220 * dumpfile.c (dump_files): Remove translation-unit.
14221 (FIRST_AUTO_NUMBERED_DUMP): Decrement.
14223 2017-05-22 Georg-Johann Lay <avr@gjlay.de>
14225 * config/avr/t-avr (AWK) [t-multilib]: Remove "-v FORMAT=Makefile"
14226 command option from $(AWK) call.
14227 * config/avr/genmultilib.awk: Simplify and rewrite so that it
14228 generates MULTILIB_REQUIRED instead of MULTILIB_EXCEPTIONS.
14229 [FORMAT]: Remove handling of variable.
14230 * config/avr/t-multilib: Regenerate.
14232 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
14234 * ipa-inline-analysis.c (inline_summary::reset): Do not reset
14236 (dump_inline_summary): Do not print self_time.
14237 (estimate_function_body_sizes): Do not set self_time.
14238 (compute_inline_parameters): Likewise.
14239 (inline_read_section, inline_write_summary): Do not stream self_time.
14240 * ipa-inline.h (inline_summary): Drop self_time.
14242 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
14244 * ipa-inline-analysis.c (account_size_time): Rename to ...
14245 (inline_summary::account_size_time): ... this one.
14246 (reset_ipa_call_summary): Turn to ...
14247 (ipa_call_summary::reset): ... this one.
14248 (reset_inline_summary): Turn to ...
14249 (inline_summary::reset): ... this one.
14250 (inline_summary_t::remove): Update.
14251 (inline_summary_t::duplicate): Update.
14252 (ipa_call_summary_t::remove): Update.
14253 (dump_inline_summary): Update.
14254 (estimate_function_body_sizes): Update.
14255 (compute_inline_parameters): Update.
14256 (estimate_node_size_and_time): Update.
14257 (inline_merge_summary): Update.
14258 (inline_update_overall_summary): Update.
14259 (inline_read_section): Update.
14260 (inline_write_summary): Update.
14261 * ipa-inline.h (inline_summary): Rename entry to size_time_table;
14262 add account_size_time and reset member functions.
14263 (ipa_call_summary): Add reset function.
14264 * ipa-predicate.h (predicate::operator &): Constify.
14266 2017-05-22 Richard Biener <rguenther@suse.de>
14268 * df-scan.c (df_insn_refs_verify): Speedup when not verifying.
14270 2017-05-19 Jason Merrill <jason@redhat.com>
14272 * tree.c (make_tree_vec_stat, grow_tree_vec_stat): Use size_t.
14274 2017-05-19 Marek Polacek <polacek@redhat.com>
14277 * fold-const.c (extract_muldiv_1) <case TRUNC_DIV_EXPR>: Add
14278 TYPE_OVERFLOW_WRAPS checks.
14280 2017-05-19 Thomas Schwinge <thomas@codesourcery.com>
14282 * tree-core.h (enum omp_clause_default_kind): Add
14283 "OMP_CLAUSE_DEFAULT_PRESENT".
14284 * tree-pretty-print.c (dump_omp_clause): Handle it.
14285 * gimplify.c (enum gimplify_omp_var_data): Add
14286 "GOVD_MAP_FORCE_PRESENT".
14287 (gimplify_adjust_omp_clauses_1): Map it to
14288 "GOMP_MAP_FORCE_PRESENT".
14289 (oacc_default_clause): Handle "OMP_CLAUSE_DEFAULT_PRESENT".
14291 * gimplify.c (oacc_default_clause): Clarify.
14293 2017-05-19 Nathan Sidwell <nathan@acm.org>
14295 LANG_HOOK_REGISTER_DUMPS
14296 * toplev.c (general_init): Call register dump lang hook.
14297 * doc/invoke.texi: Document -fdump-lang option family.
14298 * dumpfile.c (dump_files): Remove class dump here.
14299 (FIRST_AUTO_NUMBERED_DUMP): Adjust.
14300 * dumpfile.h (tree_dump_index): Remove TDI_class.
14301 * langhooks-def.h (lhd_register_dumps): Declare.
14302 (LANG_HOOKS_REGISTER_DUMPS): Define.
14303 (LANG_HOOKS_INITIALIZER): Add it.
14304 * langhooks.c (lhd_register_dumps): Define.
14305 * langhooks.h (struct lang_hooks): Add register_dumps.
14307 2017-05-19 Nathan Sidwell <nathan@acm.org>
14309 * context.h (context::set_passes): New.
14310 * context.c (context::context): Do not create pass manager.
14311 * toplev.c (general_init): Create pass manager here.
14313 2017-05-19 Segher Boessenkool <segher@kernel.crashing.org>
14315 * config/rs6000/rs6000.md (splitter to load of -1 and mask): Don't
14316 use this splitter if two add or or instructions would also work for
14317 the constant we want to generate.
14319 2017-05-19 Richard Biener <rguenther@suse.de>
14322 * genmatch.c (dt_node::gen_kids_1): Add missing scope around
14323 predicate evaluation.
14325 2017-05-19 Jan Hubicka <hubicka@ucw.cz>
14327 * ipa-inline.h (ipa_call_summary): Turn sizes into signed;
14329 * ipa-inline.c (want_inline_small_function_p): Do not cast to
14332 2017-05-19 Jan Hubicka <hubicka@ucw.cz>
14334 * ipa-inline-analysis.c (cgraph_2edge_hook_list, cgraph_edge_hook_list,
14335 inline_edge_removal_hook, inline_edge_duplication_hook): Remove.
14336 (inline_edge_summary_vec): Turn into ...
14337 (ipa_call_summaries): ... this one.
14338 (redirect_to_unreachable, edge_set_predicate,
14339 evaluate_properties_for_edge, inline_summary_alloc,
14340 reset_ipa_call_summary, reset_inline_summary,
14341 inline_summary_t::duplicate): Update.
14342 (inline_edge_duplication_hook): Turn to ...
14343 (ipa_call_summary_t::duplicate): ... this one.
14344 (inline_edge_removal_hook): Turn to ...
14345 (ipa_call_summary_t::remove): ... this one.
14346 (dump_inline_edge_summary): Turn to ...
14347 (dump_ipa_call_summary): ... this one.
14348 (estimate_function_body_sizes): Update.
14349 (inline_update_callee_summaries): Update.
14350 (remap_edge_change_prob): Update.
14351 (remap_edge_summaries): Update.
14352 (inline_merge_summary): Update.
14353 (do_estimate_edge_time): Update.
14354 (inline_generate_summary): Update.
14355 (inline_read_section): Update.
14356 (inline_read_summary): Update.
14357 (inline_free_summary): Update.
14358 * ipa-inline.c (can_inline_edge_p): Update.
14359 (compute_inlined_call_time): Update.
14360 (want_inline_small_function_p): Update.
14361 (edge_badness): Update.
14362 (early_inliner): Update.
14363 * ipa-inline.h (inline_edge_summary): Turn to ...
14364 (ipa_call_summary): ... this one.
14365 (ipa_call_summary_t): New class.
14366 (inline_edge_summary_t, inline_edge_summary_vec): Remove.
14367 (ipa_call_summaries): New.
14368 (inline_edge_summary): Remove.
14369 (estimate_edge_growth): Update.
14370 * ipa-profile.c (ipa_propagate_frequency_1): Update.
14371 * ipa-prop.c (ipa_make_edge_direct_to_target): Update.
14372 * ipa-split.c (execute_split_functions): Update.
14373 * ipa.c (symbol_table::remove_unreachable_nodes): Update.
14375 2017-05-19 Richard Biener <rguenther@suse.de>
14377 PR middle-end/80764
14378 * cfgexpand.c (expand_gimple_cond): Fix loop fixup.
14380 2017-05-18 Segher Boessenkool <segher@kernel.crashing.org>
14382 * config/rs6000/rs6000.c (struct machine_function): Add field
14383 fpr_is_wrapped_separately.
14384 (rs6000_get_separate_components): Use 64 components. Handle the
14385 new FPR components.
14386 (rs6000_components_for_bb): Handle the FPR components.
14387 (rs6000_emit_prologue_components): Handle the FPR components.
14388 (rs6000_emit_epilogue_components): Handle the FPR components.
14389 (rs6000_set_handled_components): Handle the FPR components.
14390 (rs6000_emit_prologue): Don't output prologue code for those FPRs
14391 that are already separately shrink-wrapped.
14392 (rs6000_emit_epilogue): Don't output epilogue code for those FPRs
14393 that are already separately shrink-wrapped.
14395 2017-05-18 Michael Meissner <meissner@linux.vnet.ibm.com>
14398 * config/rs6000/predicates.md (simple_offsettable_mem_operand):
14401 * config/rs6000/rs6000.md (ALTIVEC_DFORM): New iterator.
14402 (define_peephole2 for Altivec d-form load): Add peepholes to catch
14403 cases where the register allocator uses a move and an offsettable
14404 memory operation to/from a FPR register on ISA 2.06/2.07.
14405 (define_peephole2 for Altivec d-form store): Likewise.
14407 2017-05-18 Uros Bizjak <ubizjak@gmail.com>
14410 * config/i386/mmx.md (*mov<mode>_internal): Enable
14411 alternatives 11, 12, 13 and 14 also for 32bit targets.
14412 Remove alternatives 15, 16, 17 and 18.
14413 * config/i386/sse.md (vec_concatv2di): Change
14414 alternative (!x, *y) to (x, ?!*Yn).
14416 2017-05-18 Paolo Carlini <paolo.carlini@oracle.com>
14418 * dumpfile.h (enum dump_kind): Remove stray comma.
14420 2017-05-18 Jan Hubicka <hubicka@ucw.cz>
14422 * Makefile.in: Add ipa-predicate.o and ipa-predicate.h
14423 * ipa-inline-analysis.c (NUM_CONDITIONS): turn into
14424 predicate::num_conditions
14425 (IS_NOT_CONSTANT): turn into predicate::is_not_constant.
14426 (CHANGED): turn into predicate::changed.
14427 (agg_position_info): Move to ipa-predicate.h
14428 (add_condition, predicate::add_clause, predicate::operator &=,
14429 predicate::or_with, predicate::evaluate, predicate::probability,
14430 dump_condition, dump_clause, predicate::dump,
14431 predicate::remap_after_duplication, predicate::remap_after_inlining,
14432 predicate::stream_in, predicate::stream_out): Move to ipa-predicate.c
14433 (evaluate_conditions_for_known_args): Update.
14434 (set_cond_stmt_execution_predicate): Update.
14435 * ipa-inline.h: Include ipa-predicate.h
14436 (condition, inline_param_summary, conditions, agg_position_info,
14437 predicate): Move to ipa-predicate.h
14438 * ipa-predicate.c: New file.
14439 * ipa-predicate.h: New file.
14441 2017-05-18 Wilco Dijkstra <wdijkstr@arm.com>
14443 * final.c (leaf_function_p): Check we are not in a sequence.
14445 2017-05-18 Martin Liska <mliska@suse.cz>
14447 * cfgrtl.c (rtl_verify_edges): Remove usage of TDF_RTL.
14448 * dumpfile.c (dump_register): Use new enum dump_kind.
14449 (get_dump_file_name): Likewise.
14450 (dump_enable_all): Likewise.
14451 (dump_switch_p_1): Likewise.
14452 (enable_rtl_dump_file): Remove usage of TDF_RTL.
14453 * dumpfile.h (enum dump_kind): New enum type.
14454 (struct dump_file_info): Create constructor and
14455 format fields and comments.
14456 * passes.c (pass_manager::register_one_dump_file):
14458 * statistics.c (statistics_early_init): Likewise.
14459 * tree-ssa-loop-prefetch.c (dump_mem_details): Replace
14460 TDF_TREE with TDF_SLIM.
14461 (gather_memory_references_ref): Likewise.
14463 2017-05-18 Martin Liska <mliska@suse.cz>
14465 * vec.h (struct vnull): Use it.
14467 2017-05-18 Jan Hubicka <hubicka@ucw.cz>
14469 * ipa-inline-analysis.c (predicate_conditions): Move to ipa-inline.h
14470 (true_predicate, false_predicate, true_predicate_p,
14471 false_predicate_p): Remove.
14472 (single_cond_predicate, not_inlined_predicate): Turn to member function
14474 (add_condition): Update.
14475 (add_clause): Turn to...
14476 (predicate::add_clause): ... this one; update; allow passing NULL
14478 (and_predicates): Turn to ...
14479 (predicate::operator &=): ... this one.
14480 (predicates_equal_p): Move to predicate::operator == in ipa-inline.h
14481 (or_predicates): Turn to ...
14482 (predicate::or_with): ... this one.
14483 (evaluate_predicate): Turn to ...
14484 (predicate::evaluate): ... this one.
14485 (predicate_probability): Turn to ...
14486 (predicate::probability): ... this one.
14487 (dump_condition): Update.
14488 (dump_predicate): Turn to ...
14489 (predicate::dump): ... this one.
14490 (account_size_time): Update.
14491 (edge_set_predicate): Update.
14492 (set_hint_predicate): UPdate.
14493 (evaluate_conditions_for_known_args): Update.
14494 (evaluate_properties_for_edge): Update.
14495 (remap_predicate_after_duplication): Turn to...
14496 (predicate::remap_after_duplication): ... this one.
14497 (remap_hint_predicate_after_duplication): Update.
14498 (inline_summary_t::duplicate): UPdate.
14499 (dump_inline_edge_summary): Update.
14500 (dump_inline_summary): Update.
14501 (set_cond_stmt_execution_predicate): Update.
14502 (set_switch_stmt_execution_predicate): Update.
14503 (compute_bb_predicates): Update.
14504 (will_be_nonconstant_expr_predicate): Update.
14505 (will_be_nonconstant_predicate): Update.
14506 (phi_result_unknown_predicate): Update.
14507 (predicate_for_phi_result): Update.
14508 (array_index_predicate): Update.
14509 (estimate_function_body_sizes): Update.
14510 (estimate_node_size_and_time): Update.
14511 (estimate_ipcp_clone_size_and_time): Update.
14512 (remap_predicate): Rename to ...
14513 (predicate::remap_after_inlining): ... this one.
14514 (remap_hint_predicate): Update.
14515 (inline_merge_summary): Update.
14516 (inline_update_overall_summary): Update.
14517 (estimate_size_after_inlining): Update.
14518 (read_predicate): Rename to ...
14519 (predicate::stream_in): ... this one.
14520 (read_inline_edge_summary): Update.
14521 (write_predicate): Rename to ...
14522 (predicate::stream_out): ... this one.
14523 (write_inline_edge_summary): Update.
14524 * ipa-inline.h (MAX_CLAUSES): Turn to predicate::max_clauses.
14525 (clause_t): Turn to uint32_t
14526 (predicate): Turn to class; implement constructor and operators
14528 (size_time_entry): Update.
14529 (inline_summary): Update.
14530 (inline_edge_summary): Update.
14532 2017-05-18 Marc Glisse <marc.glisse@inria.fr>
14534 * fold-const.c (fold_binary_loc): Move transformation...
14535 * match.pd (C - X CMP X): ... here.
14537 2017-05-18 Sheldon Lobo <sheldon.lobo@oracle.com>
14539 * config/sparc/sparc.c (sparc_option_override): Set function
14540 alignment for -mcpu=niagara7 to 64 to match the I$ line.
14541 * config/sparc/sparc.h (BRANCH_COST): Set the SPARC M7 branch
14543 * config/sparc/sparc.h (BRANCH_COST): Set the SPARC T4 branch
14545 * config/sparc/sol2.h: Fix a ASM_CPU32_DEFAULT_SPEC typo.
14547 2017-05-18 Marek Polacek <polacek@redhat.com>
14550 * ubsan.c (instrument_null): Unwrap ADDR_EXPRs.
14551 (pass_ubsan::execute): Call gimple_assign_single_p instead of
14552 gimple_assign_load_p.
14554 2017-05-17 Segher Boessenkool <segher@kernel.crashing.org>
14556 PR middle-end/80692
14557 * real.c (do_compare): Give decimal_do_compare preference over
14558 comparing just the signs.
14560 2017-05-17 Uros Bizjak <ubizjak@gmail.com>
14562 * doc/md.texi (Canonicalization of Instructions): Describe the
14563 canonical form of instructions that inherently set a condition
14566 2017-05-17 Peter Bergner <bergner@vnet.ibm.com>
14568 PR middle-end/80775
14569 * tree-cfg.c: Move deletion of unreachable case statements to after
14570 the merging of consecutive case labels.
14572 2017-05-17 Thomas Preud'homme <thomas.preudhomme@arm.com>
14574 * config/arm/arm.c (cmse_nonsecure_call_clear_caller_saved): Refer
14575 readers to __gnu_cmse_nonsecure_call libcall for saving, clearing and
14576 restoring of callee-saved registers.
14578 2017-05-17 Eric Botcazou <ebotcazou@adacore.com>
14580 * compare-elim.c (equivalent_reg_at_start): Adjust outdated comment.
14581 * config/visium/visium.c (single_set_and_flags): Likewise.
14582 * config/visium/visium.md (Substitutions): Likewise.
14584 2017-05-17 Martin Liska <mliska@suse.cz>
14586 * cfg.c: Introduce dump_flags_t type and
14587 use it instead of int type.
14589 * cfghooks.c: Likewise.
14590 * cfghooks.h (struct cfg_hooks): Likewise.
14591 * cfgrtl.c: Likewise.
14592 * cfgrtl.h: Likewise.
14593 * cgraph.c (cgraph_node::get_body): Likewise.
14594 * coretypes.h: Likewise.
14595 * domwalk.c: Likewise.
14596 * domwalk.h: Likewise.
14597 * dumpfile.c (struct dump_option_value_info): Likewise.
14598 (dump_enable_all): Likewise.
14599 (dump_switch_p_1): Likewise.
14600 (opt_info_switch_p): Likewise.
14601 * dumpfile.h (enum tree_dump_index): Likewise.
14602 (struct dump_file_info): Likewise.
14603 * genemit.c: Likewise.
14604 * generic-match-head.c: Likewise.
14605 * gengtype.c (open_base_files): Likewise.
14606 * gimple-pretty-print.c: Likewise.
14607 * gimple-pretty-print.h: Likewise.
14608 * graph.c (print_graph_cfg): Likewise.
14609 * graphite-scop-detection.c (dot_all_sese): Likewise.
14610 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
14611 * loop-unroll.c (report_unroll): Likewise.
14612 * passes.c (pass_manager::register_one_dump_file): Likewise.
14613 * print-tree.c: Likewise.
14614 * statistics.c: Likewise.
14615 * tree-cfg.c: Likewise.
14616 * tree-cfg.h: Likewise.
14617 * tree-dfa.c: Likewise.
14618 * tree-dfa.h: Likewise.
14619 * tree-dump.c (dump_function): Likewise.
14620 * tree-dump.h (struct dump_info): Likewise.
14621 * tree-pretty-print.c: Likewise.
14622 * tree-pretty-print.h: Likewise.
14623 * tree-ssa-live.c: Likewise.
14624 * tree-ssa-live.h: Likewise.
14625 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Likewise.
14626 * tree-vect-loop.c: Likewise.
14627 * tree-vect-slp.c: Likewise.
14629 2017-05-16 James Greenhalgh <james.greenhalgh@arm.com>
14630 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14632 PR tree-optimization/80457
14633 * tree-vect-stmts.c (vect_model_simple_cost): Model the cost
14634 of all arguments to a statement as scalar_to_vec operations.
14635 (vectorizable_call): Adjust call to vect_model_simple_cost for
14637 (vectorizable_conversion): Likewise.
14638 (vectorizable_assignment): Likewise.
14639 (vectorizable_shift): Likewise.
14640 (vectorizable_operation): Likewise.
14641 (vectorizable_comparison): Likewise.
14642 (vect_is_simple_cond): Record the def types for operands.
14643 (vectorizable_condition): Likewise, call vect_model_simple_cost.
14644 * tree-vectorizer.h (vect_model_simple_cost): Add new parameter
14645 for statement argument count.
14647 2017-05-16 Carl Love <cel@us.ibm.com>
14649 * config/rs6000/rs6000-c: Add support for built-in functions
14650 vector unsigned long long vec_bperm (vector unsigned long long,
14651 vector unsigned char)
14652 vector signed long long vec_mule (vector signed int,
14654 vector unsigned long long vec_mule (vector unsigned int,
14655 vector unsigned int)
14656 vector signed long long vec_mulo (vector signed int,
14658 vector unsigned long long vec_mulo (vector unsigned int,
14659 vector unsigned int)
14660 vector signed char vec_sldw (vector signed char,
14661 vector signed char,
14663 vector unsigned char vec_sldw (vector unsigned char,
14664 vector unsigned char,
14666 vector signed short vec_sldw (vector signed short,
14667 vector signed short,
14669 vector unsigned short vec_sldw (vector unsigned short,
14670 vector unsigned short,
14672 vector signed int vec_sldw (vector signed int,
14675 vector unsigned int vec_sldw (vector unsigned int,
14676 vector unsigned int,
14678 vector signed long long vec_sldw (vector signed long long,
14679 vector signed long long,
14681 vector unsigned long long vec_sldw (vector unsigned long long,
14682 vector unsigned long long,
14684 * config/rs6000/rs6000-c: Add support for built-in functions
14685 * config/rs6000/rs6000-builtin.def: Add definition for SLDW.
14686 * config/rs6000/altivec.h: Add defintion for vec_sldw.
14687 * doc/extend.texi: Update the built-in documentation for the
14688 new built-in functions.
14690 2017-05-16 Marek Polacek <polacek@redhat.com>
14694 * tree.c (save_expr): Don't fold the expression.
14696 2017-05-16 Uros Bizjak <ubizjak@gmail.com>
14698 * config/i386/i386.md (*movsi_internal): Split (?rm,*y) alternative
14699 to (?r,*Yn) and (?m,*y) alternatives, and (?*y,rm) to (?*Ym,r)
14700 and (?*y,m). Update insn attributes.
14702 2017-05-16 Martin Liska <mliska@suse.cz>
14704 * cgraph.c (cgraph_edge::resolve_speculation): Add default value for
14705 flags argument of print_gimple_stmt, print_gimple_expr,
14706 print_generic_stmt and print_generic_expr.
14707 * cgraphclones.c (symbol_table::materialize_all_clones): Likewise.
14708 * coretypes.h: Likewise.
14709 * except.c (dump_eh_tree): Likewise.
14710 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Likewise.
14711 * gimple-pretty-print.h: Likewise.
14712 * gimple-ssa-backprop.c (dump_usage_prefix): Likewise.
14713 (backprop::push_to_worklist): Likewise.
14714 (backprop::pop_from_worklist): Likewise.
14715 (backprop::process_use): Likewise.
14716 (backprop::intersect_uses): Likewise.
14717 (note_replacement): Likewise.
14718 * gimple-ssa-store-merging.c
14719 (pass_store_merging::terminate_all_aliasing_chains): Likewise.
14720 (imm_store_chain_info::coalesce_immediate_stores): Likewise.
14721 (pass_store_merging::execute): Likewise.
14722 * gimple-ssa-strength-reduction.c (dump_candidate): Likewise.
14723 (ssa_base_cand_dump_callback): Likewise.
14724 (dump_incr_vec): Likewise.
14725 (replace_refs): Likewise.
14726 (replace_mult_candidate): Likewise.
14727 (create_add_on_incoming_edge): Likewise.
14728 (create_phi_basis): Likewise.
14729 (insert_initializers): Likewise.
14730 (all_phi_incrs_profitable): Likewise.
14731 (introduce_cast_before_cand): Likewise.
14732 (replace_one_candidate): Likewise.
14733 * gimplify.c (gimplify_expr): Likewise.
14734 * graphite-isl-ast-to-gimple.c (is_valid_rename): Likewise.
14735 (set_rename): Likewise.
14736 (rename_uses): Likewise.
14737 (copy_loop_phi_nodes): Likewise.
14738 (add_close_phis_to_merge_points): Likewise.
14739 (copy_loop_close_phi_args): Likewise.
14740 (copy_cond_phi_args): Likewise.
14741 (graphite_copy_stmts_from_block): Likewise.
14742 (translate_pending_phi_nodes): Likewise.
14743 * graphite-poly.c (print_pdr): Likewise.
14744 (dump_gbb_cases): Likewise.
14745 (dump_gbb_conditions): Likewise.
14746 (print_scop_params): Likewise.
14747 * graphite-scop-detection.c (build_cross_bb_scalars_def): Likewise.
14748 (build_cross_bb_scalars_use): Likewise.
14749 (gather_bbs::before_dom_children): Likewise.
14750 * hsa-dump.c (dump_hsa_immed): Likewise.
14751 * ipa-cp.c (print_ipcp_constant_value): Likewise.
14752 (get_replacement_map): Likewise.
14753 * ipa-inline-analysis.c (dump_condition): Likewise.
14754 (estimate_function_body_sizes): Likewise.
14755 * ipa-polymorphic-call.c (check_stmt_for_type_change): Likewise.
14756 (ipa_polymorphic_call_context::get_dynamic_type): Likewise.
14757 * ipa-prop.c (ipa_dump_param): Likewise.
14758 (ipa_print_node_jump_functions_for_edge): Likewise.
14759 (ipa_modify_call_arguments): Likewise.
14760 (ipa_modify_expr): Likewise.
14761 (ipa_dump_param_adjustments): Likewise.
14762 (ipa_dump_agg_replacement_values): Likewise.
14763 (ipcp_modif_dom_walker::before_dom_children): Likewise.
14764 * ipa-pure-const.c (check_stmt): Likewise.
14765 (pass_nothrow::execute): Likewise.
14766 * ipa-split.c (execute_split_functions): Likewise.
14767 * omp-offload.c (dump_oacc_loop_part): Likewise.
14768 (dump_oacc_loop): Likewise.
14769 * trans-mem.c (tm_log_emit): Likewise.
14770 (tm_memopt_accumulate_memops): Likewise.
14771 (dump_tm_memopt_set): Likewise.
14772 (dump_tm_memopt_transform): Likewise.
14773 * tree-cfg.c (gimple_verify_flow_info): Likewise.
14774 (print_loop): Likewise.
14775 * tree-chkp-opt.c (chkp_print_addr): Likewise.
14776 (chkp_gather_checks_info): Likewise.
14777 (chkp_get_check_result): Likewise.
14778 (chkp_remove_check_if_pass): Likewise.
14779 (chkp_use_outer_bounds_if_possible): Likewise.
14780 (chkp_reduce_bounds_lifetime): Likewise.
14781 * tree-chkp.c (chkp_register_addr_bounds): Likewise.
14782 (chkp_mark_completed_bounds): Likewise.
14783 (chkp_register_incomplete_bounds): Likewise.
14784 (chkp_mark_invalid_bounds): Likewise.
14785 (chkp_maybe_copy_and_register_bounds): Likewise.
14786 (chkp_build_returned_bound): Likewise.
14787 (chkp_get_bound_for_parm): Likewise.
14788 (chkp_build_bndldx): Likewise.
14789 (chkp_get_bounds_by_definition): Likewise.
14790 (chkp_generate_extern_var_bounds): Likewise.
14791 (chkp_get_bounds_for_decl_addr): Likewise.
14792 * tree-chrec.c (chrec_apply): Likewise.
14793 * tree-data-ref.c (dump_data_reference): Likewise.
14794 (dump_subscript): Likewise.
14795 (dump_data_dependence_relation): Likewise.
14796 (analyze_overlapping_iterations): Likewise.
14797 * tree-inline.c (expand_call_inline): Likewise.
14798 (tree_function_versioning): Likewise.
14799 * tree-into-ssa.c (dump_defs_stack): Likewise.
14800 (dump_currdefs): Likewise.
14801 (dump_names_replaced_by): Likewise.
14802 (dump_update_ssa): Likewise.
14803 (update_ssa): Likewise.
14804 * tree-object-size.c (pass_object_sizes::execute): Likewise.
14805 * tree-parloops.c (build_new_reduction): Likewise.
14806 (try_create_reduction_list): Likewise.
14807 (ref_conflicts_with_region): Likewise.
14808 (oacc_entry_exit_ok_1): Likewise.
14809 (oacc_entry_exit_single_gang): Likewise.
14810 * tree-pretty-print.h: Likewise.
14811 * tree-scalar-evolution.c (set_scalar_evolution): Likewise.
14812 (get_scalar_evolution): Likewise.
14813 (add_to_evolution): Likewise.
14814 (get_loop_exit_condition): Likewise.
14815 (analyze_evolution_in_loop): Likewise.
14816 (analyze_initial_condition): Likewise.
14817 (analyze_scalar_evolution): Likewise.
14818 (instantiate_scev): Likewise.
14819 (number_of_latch_executions): Likewise.
14820 (gather_chrec_stats): Likewise.
14821 (final_value_replacement_loop): Likewise.
14822 (scev_const_prop): Likewise.
14823 * tree-sra.c (dump_access): Likewise.
14824 (disqualify_candidate): Likewise.
14825 (create_access): Likewise.
14826 (reject): Likewise.
14827 (maybe_add_sra_candidate): Likewise.
14828 (create_access_replacement): Likewise.
14829 (analyze_access_subtree): Likewise.
14830 (analyze_all_variable_accesses): Likewise.
14831 (sra_modify_assign): Likewise.
14832 (initialize_constant_pool_replacements): Likewise.
14833 (find_param_candidates): Likewise.
14834 (decide_one_param_reduction): Likewise.
14835 (replace_removed_params_ssa_names): Likewise.
14836 * tree-ssa-ccp.c (ccp_fold_stmt): Likewise.
14837 * tree-ssa-copy.c (dump_copy_of): Likewise.
14838 (copy_prop_visit_cond_stmt): Likewise.
14839 * tree-ssa-dce.c (mark_operand_necessary): Likewise.
14840 * tree-ssa-dom.c (pass_dominator::execute): Likewise.
14841 (record_equivalences_from_stmt): Likewise.
14842 * tree-ssa-dse.c (compute_trims): Likewise.
14843 (delete_dead_call): Likewise.
14844 (delete_dead_assignment): Likewise.
14845 * tree-ssa-forwprop.c (forward_propagate_into_gimple_cond): Likewise.
14846 (forward_propagate_into_cond): Likewise.
14847 (pass_forwprop::execute): Likewise.
14848 * tree-ssa-ifcombine.c (ifcombine_ifandif): Likewise.
14849 * tree-ssa-loop-im.c (invariantness_dom_walker::before_dom_children):
14851 (move_computations_worker): Likewise.
14852 (execute_sm): Likewise.
14853 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Likewise.
14854 (remove_exits_and_undefined_stmts): Likewise.
14855 (remove_redundant_iv_tests): Likewise.
14856 * tree-ssa-loop-ivopts.c (dump_use): Likewise.
14857 (adjust_iv_update_pos): Likewise.
14858 * tree-ssa-math-opts.c (bswap_replace): Likewise.
14859 * tree-ssa-phiopt.c (factor_out_conditional_conversion): Likewise.
14860 (value_replacement): Likewise.
14861 * tree-ssa-phiprop.c (phiprop_insert_phi): Likewise.
14862 * tree-ssa-pre.c (print_pre_expr): Likewise.
14863 (get_representative_for): Likewise.
14864 (create_expression_by_pieces): Likewise.
14865 (insert_into_preds_of_block): Likewise.
14866 (eliminate_insert): Likewise.
14867 (eliminate_dom_walker::before_dom_children): Likewise.
14868 (eliminate): Likewise.
14869 (remove_dead_inserted_code): Likewise.
14870 * tree-ssa-propagate.c (substitute_and_fold): Likewise.
14871 * tree-ssa-reassoc.c (get_rank): Likewise.
14872 (eliminate_duplicate_pair): Likewise.
14873 (eliminate_plus_minus_pair): Likewise.
14874 (eliminate_not_pairs): Likewise.
14875 (undistribute_ops_list): Likewise.
14876 (eliminate_redundant_comparison): Likewise.
14877 (update_range_test): Likewise.
14878 (optimize_range_tests_var_bound): Likewise.
14879 (optimize_vec_cond_expr): Likewise.
14880 (rewrite_expr_tree): Likewise.
14881 (rewrite_expr_tree_parallel): Likewise.
14882 (linearize_expr): Likewise.
14883 (break_up_subtract): Likewise.
14884 (linearize_expr_tree): Likewise.
14885 (attempt_builtin_powi): Likewise.
14886 (attempt_builtin_copysign): Likewise.
14887 (transform_stmt_to_copy): Likewise.
14888 (transform_stmt_to_multiply): Likewise.
14889 (dump_ops_vector): Likewise.
14890 * tree-ssa-sccvn.c (vn_nary_build_or_lookup_1): Likewise.
14891 (print_scc): Likewise.
14892 (set_ssa_val_to): Likewise.
14893 (visit_reference_op_store): Likewise.
14894 (visit_use): Likewise.
14895 (sccvn_dom_walker::before_dom_children): Likewise.
14896 (run_scc_vn): Likewise.
14897 * tree-ssa-scopedtables.c (avail_exprs_stack::lookup_avail_expr):
14899 (expr_hash_elt::print): Likewise.
14900 (const_and_copies::pop_to_marker): Likewise.
14901 (const_and_copies::record_const_or_copy_raw): Likewise.
14902 * tree-ssa-structalias.c (compute_dependence_clique): Likewise.
14903 * tree-ssa-uninit.c (collect_phi_def_edges): Likewise.
14904 (dump_predicates): Likewise.
14905 (find_uninit_use): Likewise.
14906 (warn_uninitialized_phi): Likewise.
14907 (pass_late_warn_uninitialized::execute): Likewise.
14908 * tree-ssa.c (verify_vssa): Likewise.
14909 (verify_ssa): Likewise.
14910 (maybe_optimize_var): Likewise.
14911 * tree-vrp.c (dump_value_range): Likewise.
14912 (dump_all_value_ranges): Likewise.
14913 (dump_asserts_for): Likewise.
14914 (register_edge_assert_for_2): Likewise.
14915 (vrp_visit_cond_stmt): Likewise.
14916 (vrp_visit_switch_stmt): Likewise.
14917 (vrp_visit_stmt): Likewise.
14918 (vrp_visit_phi_node): Likewise.
14919 (simplify_cond_using_ranges_1): Likewise.
14920 (fold_predicate_in): Likewise.
14921 (evrp_dom_walker::before_dom_children): Likewise.
14922 (evrp_dom_walker::push_value_range): Likewise.
14923 (evrp_dom_walker::pop_value_range): Likewise.
14924 (execute_early_vrp): Likewise.
14926 2017-05-16 Richard Biener <rguenther@suse.de>
14928 * dwarf2out.c (loc_list_from_tree_1): Do not create
14929 DW_OP_GNU_variable_value for DECL_IGNORED_P decls.
14931 2017-05-16 Richard Biener <rguenther@suse.de>
14933 * dwarf2out.c (resolve_variable_value_in_expr): Lookup DIE
14935 (note_variable_value_in_expr): If we resolved the decl ref
14936 do not push to the stack.
14938 2017-05-16 Matthew Wahab <matthew.wahab@arm.com>
14940 * config/arm/arm_neon.h (vadd_f16): Use standard arithmetic
14941 operations in fast-math mode.
14942 (vaddq_f16): Likewise.
14943 (vmul_f16): Likewise.
14944 (vmulq_f16): Likewise.
14945 (vsub_f16): Likewise.
14946 (vsubq_f16): Likewise.
14947 * config/arm/neon.md (add<mode>3): New.
14949 (fma:<VH:mode>3): New. Also remove outdated comment.
14952 2017-05-16 Martin Liska <mliska@suse.cz>
14956 * ipa-devirt.c (warn_types_mismatch): Fix typo.
14957 (odr_types_equivalent_p): Likewise.
14959 2017-05-15 Sylvestre Ledru <sylvestre@debian.org>
14961 * plugin.c (try_init_one_plugin): Fix ressource leaks (CID 726637).
14963 2017-05-15 Uros Bizjak <ubizjak@gmail.com>
14966 * config/i386.i386.md (*zero_extendsidi2): Do not penalize
14967 non-interunit SSE move alternatives with '?'.
14968 (zero-extendsidi peephole2): New peephole to skip intermediate
14969 general register in SSE zero-extend sequence.
14971 2017-05-15 Jeff Law <law@redhat.com>
14973 * reorg.c (relax_delay_slots): Create a new variable to hold
14974 the temporary target rather than clobbering TARGET_LABEL.
14976 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Add
14977 missing argument to extract_bit_field call.
14978 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Likewise.
14980 2017-05-15 Martin Liska <mliska@suse.cz>
14983 * gcc.c (process_command): Do not allow empty argument of -o option.
14985 2017-05-15 Renlin Li <renlin.li@arm.com>
14987 * config/aarch64/aarch64-protos.h (aarch64_expand_call): Declare.
14988 * config/aarch64/aarch64.c (aarch64_expand_call): Define.
14989 * config/aarch64/constraints.md (Usf): Add long call check.
14990 * config/aarch64/aarch64.md (call): Use aarch64_expand_call.
14991 (call_value): Likewise.
14992 (sibcall): Likewise.
14993 (sibcall_value): Likewise.
14995 (call_value_insn): New.
14996 (sibcall_insn): Update rtx pattern.
14997 (sibcall_value_insn): Likewise.
14998 (call_internal): Remove.
14999 (call_value_internal): Likewise.
15000 (sibcall_internal): Likewise.
15001 (sibcall_value_internal): Likewise.
15002 (call_reg): Likewise.
15003 (call_symbol): Likewise.
15004 (call_value_reg): Likewise.
15005 (call_value_symbol): Likewise.
15007 2017-05-14 Krister Walfridsson <krister.walfridsson@gmail.com>
15010 * config/netbsd.h (NETBSD_LIBGCC_SPEC): Always add -lgcc.
15012 2017-05-14 Uros Bizjak <ubizjak@gmail.com>
15014 * config/i386.i386.c (ix86_cc_modes_compatible): CCNOmode is
15015 compatible with CCGOCmode and with CCZmode.
15017 2017-05-14 Martin Sebor <msebor@redhat.com>
15019 PR middle-end/77671
15020 * gimple-fold.c (gimple_fold_builtin_sprintf): Make extern.
15021 (gimple_fold_builtin_snprintf): Same.
15022 * gimple-fold.h (gimple_fold_builtin_sprintf): Declare.
15023 (gimple_fold_builtin_snprintf): Same.
15024 * gimple-ssa-sprintf.c (get_format_string): Correct the detection
15025 of character types.
15026 (is_call_safe): New function.
15027 (try_substitute_return_value): Call it.
15028 (try_simplify_call): New function.
15029 (pass_sprintf_length::handle_gimple_call): Call it.
15031 2017-05-14 Martin Sebor <msebor@redhat.com>
15033 PR middle-end/80669
15034 * builtins.c (expand_builtin_stpncpy): Simplify.
15036 2017-05-14 Daniel Santos <daniel.santos@pobox.com>
15038 * config/i386/i386.opt: Add option -mcall-ms2sysv-xlogues.
15039 * config/i386/i386.h
15040 (x86_64_ms_sysv_extra_clobbered_registers): Change type to unsigned.
15041 (NUM_X86_64_MS_CLOBBERED_REGS): New macro.
15042 (struct machine_function): Add new members call_ms2sysv,
15043 call_ms2sysv_pad_in, call_ms2sysv_pad_out and call_ms2sysv_extra_regs.
15044 (struct machine_frame_state): New fields sp_realigned and
15045 sp_realigned_offset.
15046 * config/i386/i386.c
15047 (enum xlogue_stub): New enum.
15048 (enum xlogue_stub_sets): New enum.
15049 (class xlogue_layout): New class.
15050 (struct ix86_frame): New fields stack_realign_allocate_offset,
15051 stack_realign_offset and outlined_save_offset. Modify comments to
15052 detail stack layout when using out-of-line stubs.
15053 (ix86_target_string): Add -mcall-ms2sysv-xlogues option.
15054 (ix86_option_override_internal): Add sorry() for TARGET_SEH and
15055 -mcall-ms2sysv-xlogues.
15056 (stub_managed_regs): New static variable.
15057 (ix86_save_reg): Add new parameter ignore_outlined to optionally omit
15058 registers managed by out-of-line stub.
15059 (disable_call_ms2sysv_xlogues): New function.
15060 (ix86_compute_frame_layout): Modify re-alignment calculations, disable
15061 m->call_ms2sysv when appropriate and compute frame layout for
15063 (sp_valid_at, fp_valid_at): New inline functions.
15064 (choose_basereg): New function.
15065 (choose_baseaddr): Add align parameter, use choose_basereg and modify
15067 (ix86_emit_save_reg_using_mov, ix86_emit_restore_sse_regs_using_mov):
15068 Use align parameter of choose_baseaddr to generated aligned SSE movs
15070 (pro_epilogue_adjust_stack): Modify to track
15071 machine_frame_state::sp_realigned.
15072 (ix86_nsaved_regs): Modify to accommodate changes to ix86_save_reg.
15073 (ix86_nsaved_sseregs): Likewise.
15074 (ix86_emit_save_regs): Likewise.
15075 (ix86_emit_save_regs_using_mov): Likewise.
15076 (ix86_emit_save_sse_regs_using_mov): Likewise.
15077 (get_scratch_register_on_entry): Likewise.
15078 (gen_frame_set): New function.
15079 (gen_frame_load): Likewise.
15080 (gen_frame_store): Likewise.
15081 (emit_outlined_ms2sysv_save): Likewise.
15082 (emit_outlined_ms2sysv_restore): Likewise.
15083 (ix86_expand_prologue): Modify stack re-alignment code and call
15084 emit_outlined_ms2sysv_save when appropriate.
15085 (ix86_emit_leave): Clear machine_frame_state::sp_realigned. Add
15086 parameter rtx_insn *insn, which allows the function to be used to only
15087 generate the notes.
15088 (ix86_expand_epilogue): Modify validity checks of frame and stack
15089 pointers, and call emit_outlined_ms2sysv_restore when appropriate.
15090 (ix86_expand_call): Modify to enable m->call_ms2sysv when appropriate.
15091 * config/i386/predicates.md
15092 (save_multiple): New predicate.
15093 (restore_multiple): Likewise.
15094 * config/i386/sse.md
15095 (save_multiple<mode>): New pattern.
15096 (save_multiple_realign<mode>): Likewise.
15097 (restore_multiple<mode>): Likewise.
15098 (restore_multiple_and_return<mode>): Likewise.
15099 (restore_multiple_leave_return<mode>): Likewise.
15100 * Makefile.in: Export HOSTCXX and HOSTCXXFLAGS to site.exp
15102 2017-05-14 Julia Koval <julia.koval@intel.com>
15104 * config/i386/i386-builtin-types.def (VOID_FTYPE_INT_INT64): New type.
15105 * config/i386/i386-builtin.def (__builtin_ia32_xgetbv)
15106 (__builtin_ia32_xsetbv): New builtins.
15107 * config/i386/i386.c (ix86_expand_special_args_builtin):
15109 (ix86_expand_builtin): Special expand for new intrinsics.
15110 * config/i386/i386.md (UNSPECV_XGETBV, UNSPECV_XSETBV): New.
15111 (xsetbv, xsetbv_rex64, xgetbv, xgetbv_rex64): New insn patterns.
15112 * config/i386/xsaveintrin.h (_xsetbv, _getbv): New intrinsics.
15114 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15116 * cfganal.c (inverted_post_order_compute): Change argument type
15118 * cfganal.h (inverted_post_order_compute): Adjust prototype.
15119 * df-core.c (rest_of_handle_df_initialize): Adjust.
15120 (rest_of_handle_df_finish): Likewise.
15121 (df_analyze_1): Likewise.
15122 (df_analyze): Likewise.
15123 (loop_inverted_post_order_compute): Change argument to be a vec *.
15124 (df_analyze_loop): Adjust.
15125 (df_get_n_blocks): Likewise.
15126 (df_get_postorder): Likewise.
15127 * df.h (struct df_d): Change field to be a vec.
15128 * lcm.c (compute_laterin): Adjust.
15129 (compute_available): Likewise.
15130 * lra-lives.c (lra_create_live_ranges_1): Likewise.
15131 * tree-ssa-dce.c (remove_dead_stmt): Likewise.
15132 * tree-ssa-pre.c (compute_antic): Likewise.
15134 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15136 * cfganal.c (connect_infinite_loops_to_exit): Adjust.
15137 (depth_first_search::depth_first_search): Change structure init
15138 function to this constructor.
15139 (depth_first_search::add_bb): Rename function to this member.
15140 (depth_first_search::execute): Likewise.
15141 (flow_dfs_compute_reverse_finish): Adjust.
15143 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15145 * ddg.c (find_nodes_on_paths): Use auto_sbitmap.
15146 (longest_simple_path): Likewise.
15147 * shrink-wrap.c (spread_components): Likewise.
15148 (disqualify_problematic_components): Likewise.
15149 (emit_common_heads_for_components): Likewise.
15150 (emit_common_tails_for_components): Likewise.
15151 (insert_prologue_epilogue_for_components): Likewise.
15153 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15155 * tree-ssa-dse.c (dse_dom_walker): Make m_live_byes a
15158 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15160 * df-core.c (df_set_blocks): Start using auto_bitmap.
15161 (df_compact_blocks): Likewise.
15162 * df-problems.c (df_rd_confluence_n): Likewise.
15163 * df-scan.c (df_insn_rescan_all): Likewise.
15164 (df_process_deferred_rescans): Likewise.
15165 (df_update_entry_block_defs): Likewise.
15166 (df_update_exit_block_uses): Likewise.
15167 (df_entry_block_bitmap_verify): Likewise.
15168 (df_exit_block_bitmap_verify): Likewise.
15169 (df_scan_verify): Likewise.
15170 * lra-constraints.c (lra_constraints): Likewise.
15171 (undo_optional_reloads): Likewise.
15172 (lra_undo_inheritance): Likewise.
15173 * lra-remat.c (calculate_gen_cands): Likewise.
15174 (do_remat): Likewise.
15175 * lra-spills.c (assign_spill_hard_regs): Likewise.
15176 (spill_pseudos): Likewise.
15177 * tree-ssa-pre.c (bitmap_set_and): Likewise.
15178 (bitmap_set_subtract_values): Likewise.
15180 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15182 * haifa-sched.c (estimate_shadow_tick): Replace manual bitmap
15183 management with auto_bitmap.
15184 (fix_inter_tick): Likewise.
15185 (fix_recovery_deps): Likewise.
15186 * ira.c (add_store_equivs): Likewise.
15187 (find_moveable_pseudos): Likewise.
15188 (split_live_ranges_for_shrink_wrap): Likewise.
15189 * print-rtl.c (rtx_reuse_manager::rtx_reuse_manager): Likewise.
15190 (rtx_reuse_manager::seen_def_p): Likewise.
15191 (rtx_reuse_manager::set_seen_def): Likewise.
15192 * print-rtl.h (class rtx_reuse_manager): Likewise.
15194 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15196 * bt-load.c (combine_btr_defs): Use auto_bitmap to manage bitmap
15198 (migrate_btr_def): Likewise.
15199 * cfgloop.c (get_loop_body_in_bfs_order): Likewise.
15200 * df-core.c (loop_post_order_compute): Likewise.
15201 (loop_inverted_post_order_compute): Likewise.
15202 * hsa-common.h: Likewise.
15203 * hsa-gen.c (hsa_bb::~hsa_bb): Likewise.
15204 * init-regs.c (initialize_uninitialized_regs): Likewise.
15205 * ipa-inline.c (resolve_noninline_speculation): Likewise.
15206 (inline_small_functions): Likewise.
15207 * ipa-reference.c (ipa_reference_write_optimization_summary): Likewise.
15208 * ira.c (combine_and_move_insns): Likewise.
15209 (build_insn_chain): Likewise.
15210 * loop-invariant.c (find_invariants): Likewise.
15211 * lower-subreg.c (propagate_pseudo_copies): Likewise.
15212 * predict.c (tree_predict_by_opcode): Likewise.
15213 (predict_paths_leading_to): Likewise.
15214 (predict_paths_leading_to_edge): Likewise.
15215 (estimate_loops_at_level): Likewise.
15216 (estimate_loops): Likewise.
15217 * shrink-wrap.c (try_shrink_wrapping): Likewise.
15218 (spread_components): Likewise.
15219 * tree-cfg.c (remove_edge_and_dominated_blocks): Likewise.
15220 * tree-loop-distribution.c (rdg_build_partitions): Likewise.
15221 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
15222 * tree-ssa-coalesce.c (coalesce_ssa_name): Likewise.
15223 * tree-ssa-phionlycprop.c (pass_phi_only_cprop::execute): Likewise.
15224 * tree-ssa-pre.c (remove_dead_inserted_code): Likewise.
15225 * tree-ssa-sink.c (nearest_common_dominator_of_uses): Likewise.
15226 * tree-ssa-threadupdate.c (compute_path_counts): Likewise.
15227 (mark_threaded_blocks): Likewise.
15228 (thread_through_all_blocks): Likewise.
15229 * tree-ssa.c (verify_ssa): Likewise.
15230 (execute_update_addresses_taken): Likewise.
15231 * tree-ssanames.c (verify_ssaname_freelists): Likewise.
15233 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15235 * cfganal.c (mark_dfs_back_edges): Replace manual stack with
15237 (post_order_compute): Likewise.
15238 (inverted_post_order_compute): Likewise.
15239 (pre_and_rev_post_order_compute_fn): Likewise.
15241 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15243 * genrecog.c (int_set::int_set): Explicitly construct our
15244 auto_vec base class.
15245 * vec.h (auto_vec::auto_vec): New constructor.
15247 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15249 * bitmap.h (class auto_bitmap): New constructor taking
15250 bitmap_obstack * argument.
15252 2017-05-13 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
15254 * bitmap.h (class auto_bitmap): Change type of m_bits to
15255 bitmap_head, and adjust ctor / dtor and member operators.
15257 2017-05-13 Uros Bizjak <ubizjak@gmail.com>
15259 * compare-elim.c (equivalent_reg_at_start): Return NULL_RTX
15260 when returned register mode doesn't match original mode.
15262 2017-05-12 Jeff Law <law@redhat.com>
15263 Jakub Jelinek <jakub@redhat.com>
15265 * config/mn10300/mn10300.c (mn10300_match_ccmode): Fix where
15266 we look for cc setter after the compare-elim changes.
15267 * config/mn10300/mn10300.md (addsi3_flags): Fix order of patterns
15268 within the vector to match what compare-elim now expects.
15269 (subsi3_flags, andsi3_flags, iorsi3_flags): Likewise.
15270 (xorsi3_flags, one_cmplsi2_flags): Likewise.
15272 * config/rx/rx.c (rx_match_ccmode): Fix where we look cc setter
15273 after the compare-elim changes.
15274 * config/rx/rx.md (abssi2_flags): Fix order of patterns within
15275 the vector to match what compare-elim now expects.
15276 (addsi3_flags, adc_flags, addsi3_flags peepholes): Likewise.
15277 (andsi3_flags, negsi2_flags, one_cmplsi2_flags): Likewise.
15278 (iorsi3_flags, rotlsi3_flags, rotrsi3_flags): Likewise.
15279 (ashrsi3_flags, lshrsi3_flags, ashlsi3_flags): Likewise.
15280 (ssaddsi3, subsi3_flags, sbb_flags, xorsi3_flags): Likewise.
15282 * config/visium/visium.c (single_set_and_flags): Fix where
15283 we look for cc setter after the compare-elim changes.
15284 * config/visium/visium.md (flags_subst_logic): Fix order of patterns
15285 with the vector to match what compare-elim now expects.
15286 (flags_subst_arith, add<mode>3_insn_set_carry): Likewise.
15287 (add<mode>3_insn_set_overflow, addsi3_insn_set_carry): Likewise.
15288 (addsi3_insn_set_overflow, sub<mode>3_insn_set_carry): Likewise.
15289 (sub<mode>3_insn_set_overflow, subsi3_insn_set_carry): Likewise.
15290 (subsi3_insn_set_overflow, negsi2_insn_set_carry): Likewise.
15291 (neg<mode>2_insn_set_overflow): Likewise.
15293 2017-05-12 Jim Wilson <jim.wilson@linaro.org>
15295 PR middle-end/79794
15296 * expmed.c (extract_bit_field_1): Add alt_rtl argument. Before
15297 maybe_expand_insn call, set ops[0].target. If still set after call,
15298 set alt_rtl. Add extra arg to recursive calls.
15299 (extract_bit_field): Add alt_rtl argument. Pass to
15301 * expmed.h (extract_bit_field): Fix prototype.
15302 * expr.c (emit_group_load_1, copy_blkmode_from_reg)
15303 (copy_blkmode_to_reg, read_complex_part, store_field): Pass extra NULL
15304 to extract_bit_field_calls.
15305 (expand_expr_real_1): Pass alt_rtl to expand_expr_real instead of 0.
15306 Pass alt_rtl to extract_bit_field calls.
15307 * calls.c (store_unaligned_arguments_into_psuedos)
15308 load_register_parameters): Pass extra NULL to extract_bit_field calls.
15309 * optabs.c (maybe_legitimize_operand): Clear op->target when call
15311 * optabs.h (struct expand_operand): Add target bitfield.
15313 2017-05-12 Uros Bizjak <ubizjak@gmail.com>
15315 * compare-elim.c (try_eliminate_compare): Canonicalize
15316 operation with embedded compare to
15317 [(set (reg:CCM) (compare:CCM (operation) (immediate)))
15318 (set (reg) (operation)].
15320 * config/i386/i386.c (TARGET_FLAGS_REGNUM): New define.
15322 2017-05-12 Uros Bizjak <ubizjak@gmail.com>
15325 * config/i386/i386.c (ix86_rtx_cost) [case PLUS]: Ignore the
15326 cost of adding a carry flag for ADC instruction.
15327 [case MINUS]: Ignore the cost of subtracting a carry flag
15328 for SBB instruction.
15330 2017-05-12 Steven Munroe <munroesj@gcc.gnu.org>
15332 * config.gcc (powerpc*-*-*): Add bmi2intrin.h, bmiintrin.h,
15334 * config/rs6000/bmiintrin.h: New file.
15335 * config/rs6000/bmi2intrin.h: New file.
15336 * config/rs6000/x86intrin.h: New file.
15338 2017-05-12 Jeff Law <law@redhat.com>
15340 * tree-vrp.c (vrp_dom_walker::before_dom_children): Push unwinding
15343 2017-05-12 Peter Bergner <bergner@vnet.ibm.com>
15345 PR middle-end/80707
15346 * tree-cfg.c: Remove cfg edges of unreachable case statements.
15348 2017-05-12 Will Schmidt <will_schmidt@vnet.ibm.com>
15350 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling for
15351 early expansion of vector divide builtins.
15352 (builtin_function_type): Add VSX_BUILTIN_UDIV_V2DI to the list of
15353 builtins identified as having unsigned arguments.
15355 2017-05-12 Will Schmidt <will_schmidt@vnet.ibm.com>
15357 * config/rs6000/rs6000.c (gimple-fold.h): New #include.
15358 (rs6000_gimple_fold_builtin): Add handling for early GIMPLE
15359 expansion of vector logical operations (and, andc, or, xor,
15362 2017-05-12 Will Schmidt <will_schmidt@vnet.ibm.com>
15364 * gimple-fold.c (create_tmp_reg_or_ssa_name): Remove static declaration.
15365 * gimple-fold.h (create_tmp_reg_or_ssa_name): New prototype.
15367 2017-05-12 Will Schmidt <will_schmidt@vnet.ibm.com>
15369 * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling for
15370 early GIMPLE expansion of vector multiplies.
15372 2017-05-12 Prakhar Bahuguna <prakhar.bahuguna@arm.com>
15374 * config/arm/arm.md (movsi): Add TARGET_32BIT in addition to the
15375 TARGET_HAVE_MOVT conditional.
15376 (movt splitter): Likewise.
15378 2017-05-12 Richard Biener <rguenther@suse.de>
15380 * tree-ssa-sccvn.h (has_VN_INFO): Declare.
15381 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
15382 Fold all stmts not inplace.
15384 2017-05-12 Richard Biener <rguenther@suse.de>
15386 PR tree-optimization/80713
15387 * tree-ssa-pre.c (remove_dead_inserted_code): Clear
15388 inserted_exprs bit for not removed stmts.
15390 2017-05-12 Thomas Schwinge <thomas@codesourcery.com>
15392 PR middle-end/69921
15393 * tree-parloops.c (create_parallel_loop): Set "oacc kernels
15394 parallelized" attribute for parallelized OpenACC kernels.
15395 * omp-offload.c (execute_oacc_device_lower): Use it.
15397 * omp-expand.c (expand_omp_target) <GF_OMP_TARGET_KIND_OACC_KERNELS>:
15398 Set "oacc kernels" attribute.
15399 * omp-general.c (oacc_set_fn_attrib): Remove is_kernel formal
15400 parameter. Adjust all users.
15401 (oacc_fn_attrib_kernels_p): Remove function.
15402 * omp-offload.c (execute_oacc_device_lower): Look for "oacc
15403 kernels" attribute instead of calling oacc_fn_attrib_kernels_p.
15404 * tree-ssa-loop.c (gate_oacc_kernels): Likewise.
15405 * tree-parloops.c (create_parallel_loop): If oacc_kernels_p,
15406 assert "oacc kernels" attribute is set.
15408 2017-05-11 Carl Love <cel@us.ibm.com>
15410 * config/rs6000/rs6000-c: Add support for built-in functions
15411 vector unsigned char vec_popcnt (vector signed char)
15412 vector unsigned char vec_popcnt (vector unsigned char)
15413 vector unsigned short vec_popcnt (vector signed short)
15414 vector unsigned short vec_popcnt (vector unsigned short)
15415 vector unsigned int vec_popcnt (vector signed int)
15416 vector unsigned int vec_popcnt (vector unsigned int)
15417 vector unsigned long long vec_popcnt (vector signed long long)
15418 vector unsigned long long vec_popcnt (vector unsigned long long)
15419 vector signed long long vec_slo (vector signed long long,
15420 vector signed char)
15421 vector signed long long vec_slo (vector signed long long,
15422 vector unsigned char)
15423 vector unsigned long long vec_slo (vector unsigned long long,
15424 vector signed char)
15425 vector unsigned long long vec_slo (vector unsigned long long,
15426 vector unsigned char)
15427 * config/rs6000/rs6000-builtin.def: Add definitions for VPOPCNTUB,
15428 VPOPCNTUH, VPOPCNTUW, and VPOPCNTUD overloads.
15429 * config/rs6000/altivec.h: Add define for vec_popcnt, vec_popcntb,
15430 vec_popcnth, vec_popcntw and vec_popcntd built-in functions.
15431 * doc/extend.texi: Update the built-in documentation file for the
15432 new built-in functions.
15434 2017-05-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15436 * attribs.h (sorted_attr_string): Move machine independent
15437 functions for target clone support from the i386 port to common
15438 code. Rename ix86_function_versions to common_function_versions.
15439 Rename make_name to make_unique_name.
15440 (common_function_versions): Likewise.
15441 (make_unique_name): Likewise.
15442 (make_dispatcher_decl): Likewise.
15443 (is_function_default_version): Likewise.
15444 * attribs.c (attr_strcmp): Likewise.
15445 (sorted_attr_string): Likewise.
15446 (common_function_versions): Likewise.
15447 (make_unique_name): Likewise.
15448 (make_dispatcher_decl): Likewise.
15449 (is_function_default_version): Likewise.
15450 * config/i386/i386.c (attr_strcmp): Likewise.
15451 (sorted_attr_string): Likewise.
15452 (ix86_function_versions): Likewise.
15453 (make_name): Likewise.
15454 (make_dispatcher_decl): Likewise.
15455 (is_function_default_version): Likewise.
15456 (TARGET_OPTION_FUNCTION_VERSIONS): Update target function hook.
15458 2017-05-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15461 * config/rs6000/rs6000.c (rs6000_builtin_vectorization_cost):
15462 Account for direct move costs for vec_construct of integer
15465 2017-05-11 Uros Bizjak <ubizjak@gmail.com>
15468 * config/i386/sync.md (UNSPEC_LDX_ATOMIC): New unspec.
15469 (UNSPEC_STX_ATOMIC): Ditto.
15470 (loaddi_via_sse): New insn.
15471 (storedi_via_sse): Ditto.
15472 (atomic_loaddi_fpu): Emit loaddi_via_sse and storedi_via_sse.
15473 Update corresponding peephole2 patterns.
15474 (atomic_storedi_fpu): Ditto.
15476 2017-05-11 Julia Koval <julia.koval@intel.com>
15478 * config/i386/avx512fintrin.h (_mm_mask_rsqrt14_sd)
15479 (_mm_maskz_rsqrt14_sd, _mm_mask_rsqrt14_ss, _mm_maskz_rsqrt14_ss):
15481 * config/i386/i386-builtin.def (__builtin_ia32_rsqrt14sd_mask)
15482 (__builtin_ia32_rsqrt14ss_mask): New builtins.
15483 * config/i386/sse.md (rsqrt14_<mode>_mask): New pattern.
15485 2017-05-11 Nathan Sidwell <nathan@acm.org>
15487 * graphite-poly.c: Include dumpfile.h.
15489 * dumpfle.h (dump_function): Declare here ...
15490 * tree-dump.h (dump_function): ... not here.
15491 * dumpfile.c: #include tree-cfg.h.
15492 (dump_function): Move here from ...
15493 * tree-dump.c (dump_function): ... here.
15494 * gimplify.c: #include splay-tree.h, not tree-dump.h.
15495 * graphite-poly.c: Don't include tree-dump.h.
15496 * cgraphclones.c: Include dumpfile.h not tree-dump.h.
15497 * print-tree.c: Likewise.
15498 * stor-layout.c: Likewise.
15499 * tree-nested.c: Likewise.
15501 * dumpfile.c (dump_start): Use TDF_FLAGS.
15502 (dump_enable_all): Fix TDF_KIND check thinko.
15504 2017-05-11 Kelvin Nilsen <kelvin@gcc.gnu.org>
15506 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15507 array entries to represent two legal parameterizations of the
15508 overloaded __builtin_cmpb function, as represented by the
15509 P6_OV_BUILTIN_CMPB constant.
15510 (altivec_resolve_overloaded_builtin): Add special case handling
15511 for the __builtin_cmpb function, as represented by the
15512 P6_OV_BUILTIN_CMPB constant.
15513 * config/rs6000/rs6000-builtin.def (BU_P6_2): New macro.
15514 (BU_P6_64BIT_2): New macro.
15515 (BU_P6_OVERLOAD_2): New macro
15516 (CMPB_32): Add 32-bit compare-bytes support for 32-bit only targets.
15517 (CMPB): Add 64-bit compare-bytes support for 32-bit and 64-bit targets.
15518 (CMPB): Add overload support to represent both 32-bit and 64-bit
15519 compare-bytes function.
15520 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
15521 support for TARGET_CMPB.
15522 * config/rs6000/rs6000.h: Add support for RS6000_BTM_CMPB.
15523 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Add
15524 documentation of the __builtin_cmpb overloaded built-in function.
15526 2017-05-11 Richard Biener <rguenther@suse.de>
15528 PR tree-optimization/80705
15529 * tree-vect-data-refs.c (vect_analyze_data_refs): DECL_NONALIASED
15530 bases are not vectorizable.
15532 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15534 * tree-ssa-loop-ivopts.c (determine_set_costs): Skip non-interger
15535 when counting register pressure.
15537 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15539 * tree-ssa-loop-ivopts.c (compare_cost_pair): New.
15540 (iv_ca_more_deps): Renamed to ...
15541 (iv_ca_compare_deps): ... this.
15542 (iv_ca_extend): Extend iv_ca if NEW_CP is cheaper than OLD_CP.
15544 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15546 * tree-ssa-loop-ivopts.c (find_interesting_uses): Move inv vars dump
15548 (determine_group_iv_costs): ... here.
15549 (find_inv_vars_cb): Record inv var if it's not recorded before.
15551 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15553 * tree-ssa-loop-ivopts.c (allow_ip_end_pos_p): Refine comments.
15554 (get_shiftadd_cost): Ditto.
15556 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15558 * tree-ssa-address.c: Include header file.
15559 (move_hint_to_base): Return TRUE if BASE_HINT is moved to memory
15561 (add_to_parts): Refactor.
15562 (addr_to_parts): New parameter. Update use of move_hint_to_base.
15563 (create_mem_ref): Update use of addr_to_parts. Re-associate addr
15566 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15568 PR tree-optimization/53090
15569 * tree-ssa-loop-ivopts.c (enum comp_iv_rewrite): New enum value
15571 (extract_cond_operands): Detect condition with IV on both sides
15572 and return COMP_IV_EXPR_2.
15573 (find_interesting_uses_cond): Add iv_use for both IVs in condition.
15574 (rewrite_use_compare): Simplify by removing call to function
15575 extract_cond_operands.
15577 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15579 * tree-ssa-loop-ivopts.c (enum comp_iv_rewrite): New.
15580 (extract_cond_operands): Detect condition comparing against non-
15581 invariant bound and return appropriate enum value.
15582 (find_interesting_uses_cond): Update use of extract_cond_operands.
15583 Handle its return value accordingly.
15584 (determine_group_iv_cost_cond, rewrite_use_compare): Ditto.
15586 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15588 * tree-ssa-loop-ivopts.c (rewrite_use_nonlinear_expr): Re-associate
15589 nonlinear iv_use computation in loop invariant sensitive way.
15591 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15593 * tree-ssa-loop-ivopts.c (relate_compare_use_with_all_cands): New.
15594 (find_iv_candidates): Call relate_compare_use_with_all_cands.
15596 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15598 * tree-ssa-loop-ivopts.c (struct iv_cand): New field inv_exprs.
15599 (dump_cand): Support iv_cand.inv_exprs.
15600 (add_candidate_1): Record invariant exprs in iv_cand.inv_exprs
15602 (iv_ca_set_no_cp, iv_ca_set_cp, free_loop_data): Support
15605 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15607 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p): Move
15609 * tree-ssa-address.c (multiplier_allowed_in_address_p): ... to here
15610 as local function. Include necessary header files.
15611 * tree-ssa-loop-ivopts.h (multiplier_allowed_in_address_p): Delete.
15613 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15615 * tree-ssa-loop-ivopts.c (autoinc_possible_for_pair): Simplify.
15617 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15619 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Handle more
15620 operators: TRUNC_DIV_EXPR, BIT_AND_EXPR, BIT_IOR_EXPR, LSHIFT_EXPR,
15621 RSHIFT_EXPR and BIT_NOT_EXPR.
15623 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15625 * tree-ssa-loop-ivopts.c (get_loop_invariant_expr): Simplify.
15626 (adjust_setup_cost): New parameter supporting round up adjustment.
15627 (struct address_cost_data): Delete.
15628 (force_expr_to_var_cost): Don't bound cost with spill_cost.
15629 (split_address_cost, ptr_difference_cost): Delete.
15630 (difference_cost, compare_aff_trees, record_inv_expr): Delete.
15631 (struct ainc_cost_data): New struct.
15632 (get_address_cost_ainc): New function.
15633 (get_address_cost, get_computation_cost): Reimplement.
15634 (determine_group_iv_cost_address): Record inv_expr for all uses of
15636 (determine_group_iv_cost_cond): Call get_loop_invariant_expr.
15637 (iv_ca_has_deps): Reimplemented to ...
15638 (iv_ca_more_deps): ... this. Check if NEW_CP introduces more deps
15640 (iv_ca_extend): Call iv_ca_more_deps.
15642 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15644 * tree-ssa-address.c (struct mem_address): Move to header file.
15645 (valid_mem_ref_p, move_fixed_address_to_symbol): Make it global.
15646 * tree-ssa-address.h (struct mem_address): Move from C file.
15647 (valid_mem_ref_p, move_fixed_address_to_symbol): Declare.
15649 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15651 * tree-affine.h (aff_combination_type): New interface.
15652 (aff_combination_zero_p): Remove static.
15653 (aff_combination_const_p): New interface.
15654 (aff_combination_singleton_var_p): New interfaces.
15656 2017-05-11 Richard Biener <rguenther@suse.de>
15658 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
15659 Skip unreachable blocks and destinations.
15660 (eliminate): Move stmt removal and fixup ...
15661 (fini_eliminate): ... here. Skip inserted exprs.
15662 (pass_pre::execute): Move fini_pre after fini_eliminate.
15663 * tree-ssa-tailmerge.c: Include tree-cfgcleanup.h.
15664 (tail_merge_optimize): Run cleanup_tree_cfg if requested by
15665 PRE to get rid of dead code that has invalid SSA form and
15666 split critical edges again.
15668 2017-05-11 Bin Cheng <bin.cheng@arm.com>
15670 * rtlanal.c (rtx_cost): Handle TRUNCATE between tieable modes.
15672 2017-05-11 Richard Biener <rguenther@suse.de>
15674 * passes.c (execute_function_todo): Verify loops if they are
15675 said to be up-to-date.
15676 * cfgexpand.c (pass_expand::execute): Discard loops for -dx.
15677 * trans-mem.c (pass_tm_edges::execute): Mark loops for fixup.
15679 2017-05-10 John David Anglin <danglin@gcc.gnu.org>
15682 * config/pa/pa.c (pa_assemble_integer): When outputting a SYMBOL_REF,
15683 handle calling assemble_external ourself.
15686 * config/pa/pa.c (pa_cannot_change_mode_class): Reject changes to/from
15687 modes with zero size. Enhance comment.
15689 2017-05-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15691 * config/rs6000/rs6000.c (altivec_init_builtins): Define POWER8
15692 built-ins for vec_xl and vec_xst with short and char pointer
15695 2017-05-10 Sebastian Peryt <sebastian.peryt@intel.com>
15697 * config/i386/avx512fintrin.h (_mm_mask_max_round_sd)
15698 (_mm_maskz_max_round_sd, _mm_mask_max_round_ss)
15699 (_mm_maskz_max_round_ss, _mm_mask_min_round_sd)
15700 (_mm_maskz_min_round_sd, _mm_mask_min_round_ss)
15701 (_mm_maskz_min_round_ss): New intrinsics.
15702 * config/i386/i386-builtin-types.def (V2DF, V2DF, V2DF, V2DF, UQI, INT)
15703 (V4SF, V4SF, V4SF, V4SF, UQI, INT): New function type aliases.
15704 * config/i386/i386-builtin.def (__builtin_ia32_maxsd_mask_round)
15705 (__builtin_ia32_maxss_mask_round, __builtin_ia32_minsd_mask_round)
15706 (__builtin_ia32_minss_mask_round): New builtins.
15707 * config/i386/i386.c (V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
15708 (V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT): Handle new types.
15709 * config/i386/sse.md (<sse>_vm<code><mode>3<round_saeonly_name>):
15711 (<sse>_vm<code><mode>3<mask_name><round_saeonly_name>): ... this.
15712 (v<maxmin_float><ssescalarmodesuffix>\t{<round_saeonly_op3>%2, %1, %0|%0, %1, %<iptr>2<round_saeonly_op3>}):
15714 (v<maxmin_float><ssescalarmodesuffix>\t{<round_saeonly_mask_op3>%2, %1, %0<mask_operand3>|%0<mask_operand3>, %1, %<iptr>2<round_saeonly_mask_op3>}):
15717 2017-05-10 Sebastian Peryt <sebastian.peryt@intel.com>
15719 * config/i386/avx512fintrin.h (_mm_mask_mul_round_sd)
15720 (_mm_maskz_mul_round_sd, _mm_mask_mul_round_ss)
15721 (_mm_maskz_mul_round_ss, _mm_mask_div_round_sd)
15722 (_mm_maskz_div_round_sd, _mm_mask_div_round_ss)
15723 (_mm_maskz_div_round_ss, _mm_mask_mul_sd, _mm_maskz_mul_sd)
15724 (_mm_mask_mul_ss, _mm_maskz_mul_ss, _mm_mask_div_sd)
15725 (_mm_maskz_div_sd, _mm_mask_div_ss, _mm_maskz_div_ss): New intrinsics.
15726 * config/i386/i386-builtin-types.def (V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
15727 (V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT): New function type aliases.
15728 * config/i386/i386-builtin.def (__builtin_ia32_divsd_mask_round)
15729 (__builtin_ia32_divss_mask_round, __builtin_ia32_mulsd_mask_round)
15730 (__builtin_ia32_mulss_mask_round): New builtins.
15731 * config/i386/i386.c (V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
15732 (V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT): Handle new types.
15733 * config/i386/sse.md (<sse>_vm<multdiv_mnemonic><mode>3<round_name>):
15735 (<sse>_vm<multdiv_mnemonic><mode>3<mask_name><round_name>): ... this.
15736 (v<multdiv_mnemonic><ssescalarmodesuffix>\t{<round_op3>%2, %1, %0|%0, %1, %<iptr>2<round_op3>}):
15738 (v<multdiv_mnemonic><ssescalarmodesuffix>\t{<round_mask_op3>%2, %1, %0<mask_operand3>|%0<mask_operand3>, %1, %<iptr>2<round_mask_op3>}):
15741 2017-05-10 Julia Koval <julia.koval@intel.com>
15743 * config/i386/avxintrin.h (_mm256_set_m128, _mm256_set_m128d)
15744 (_mm256_set_m128i, _mm256_setr_m128, _mm256_setr_m128d)
15745 (_mm256_setr_m128i): New intrinsics.
15747 2017-05-10 Julia Koval <julia.koval@intel.com>
15749 * config/i386/avx512fintrin.h (_mm_mask_rcp14_sd)
15750 (_mm_maskz_rcp14_sd, _mm_mask_rcp14_ss)
15751 (_mm_maskz_rcp14_ss): New intrinsics.
15752 * config/i386/i386-builtin.def (__builtin_ia32_rcp14sd_mask)
15753 (__builtin_ia32_rcp14ss_mask): New builtins.
15754 * config/i386/sse.md (srcp14<mode>_mask): New pattern.
15756 2017-05-10 Peter Bergner <bergner@vnet.ibm.com>
15758 PR tree-optimization/51513
15759 * tree-cfg.c (gimple_seq_unreachable_p): New function.
15760 (assert_unreachable_fallthru_edge_p): Use it.
15761 (group_case_labels_stmt): Likewise.
15762 * tree-cfg.h: Prototype it.
15763 * stmt.c: Include cfghooks.h and tree-cfg.h.
15764 (emit_case_dispatch_table) <gap_label>: New local variable.
15765 Use it to fill dispatch table gaps.
15766 Test for default_label before updating probabilities.
15767 (expand_case) <default_label>: Remove unneeded initialization.
15768 Test for unreachable default case statement and remove its edge.
15769 Set default_label accordingly.
15770 * tree-ssa-ccp.c (optimize_unreachable): Update comment.
15772 2017-05-10 Carl Love <cel@us.ibm.com>
15774 * config/rs6000/rs6000-c: Add support for built-in functions
15775 vector signed char vec_neg (vector signed char)
15776 vector signed short int vec_neg (vector short int)
15777 vector signed int vec_neg (vector signed int)
15778 vector signed long long vec_neg (vector signed long long)
15779 vector float vec_neg (vector float)
15780 vector double vec_neg (vector double)
15781 * config/rs6000/rs6000-builtin.def: Add definitions for NEG function
15783 * config/rs6000/altivec.h: Add define for vec_neg
15784 * doc/extend.texi: Update the built-in documentation for the
15785 new built-in functions.
15787 2017-05-10 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
15789 PR tree-optimization/77644
15790 * match.pd (sqrt(x) cmp sqrt(y) -> x cmp y): New pattern.
15792 2017-05-10 Nathan Sidwell <nathan@acm.org>
15794 * dumpfile.h (TDI_lang_all): New.
15795 (TDF_KIND): New. Renumber others
15796 (TDF_LANG, TDF_TREE, TDF_RTL, TDF_IPA): Enumerate value, rather
15798 * dumpfile.c (dump_files): Mark language dumps as TDF_LANG. add
15800 (get_dump_file_name): Adjust suffix generation.
15801 (dump_enable_all): Use TDF_KIND.
15802 * doc/invoke.texi (-fdump-lang-all): Document.
15804 * dumpfile.h: Tabify.
15806 2017-05-10 Wilco Dijkstra <wdijkstr@arm.com>
15809 * config/aarch64/cortex-a57-fma-steering.c (merge_forest):
15810 Move member access before delete.
15812 2017-05-10 Alexandre Oliva <aoliva@redhat.com>
15814 * tree-inline.c (expand_call_inline): Split block at stmt
15817 2017-05-09 Michael Meissner <meissner@linux.vnet.ibm.com>
15820 * config/rs6000/rs6000.md (f32_lr): Delete mode attributes that
15821 are now unused after splitting mov{sf,sd}_hardfloat.
15822 (f32_lr2): Likewise.
15823 (f32_lm): Likewise.
15824 (f32_lm2): Likewise.
15825 (f32_li): Likewise.
15826 (f32_li2): Likewise.
15827 (f32_lv): Likewise.
15828 (f32_sr): Likewise.
15829 (f32_sr2): Likewise.
15830 (f32_sm): Likewise.
15831 (f32_sm2): Likewise.
15832 (f32_si): Likewise.
15833 (f32_si2): Likewise.
15834 (f32_sv): Likewise.
15835 (f32_dm): Likewise.
15836 (f32_vsx): Likewise.
15837 (f32_av): Likewise.
15838 (mov<mode>_hardfloat): Split into separate movsf and movsd pieces.
15839 For movsf, order stores so the VSX stores occur before the GPR
15840 store which encourages the register allocator to use a traditional
15841 FPR instead of a GPR. For movsd, order the stores so that the GPR
15842 store comes before the VSX stores to allow the power6 to work.
15843 This is due to the power6 not having a 32-bit integer store
15844 instruction from a FPR.
15845 (movsf_hardfloat): Likewise.
15846 (movsd_hardfloat): Likewise.
15848 2017-05-09 Martin Sebor <msebor@redhat.com>
15850 PR translation/80280
15851 * config/sol2-c.c (cmn_err_flag_specs): Initialize new data member
15854 PR translation/80280
15855 * config/i386/msformat-c.c (ms_printf_flag_specs): Initialize new
15856 data member added in r247778.
15857 (ms_scanf_flag_specs, ms_strftime_flag_specs): Same.
15859 2017-05-09 Nathan Sidwell <nathan@acm.org>
15861 * tree.h (tree_fits_shwi_p, tree_fits_uhwi_p): Unconditionally pure.
15863 * ipa-devirt.c (default_hash_traits<type_pair>): Skip struct-scope
15866 2017-05-09 Marek Polacek <polacek@redhat.com>
15868 * doc/invoke.texi: Fix typo.
15870 2017-05-09 Richard Biener <rguenther@suse.de>
15872 * tree-vrp.c (vrp_val_is_max): Adjust comment.
15873 (vrp_val_is_min): Likewise.
15874 (set_value_range_to_value): Likewise.
15875 (set_value_range_to_nonnegative): Likewise.
15876 (gimple_assign_nonzero_p): Likewise.
15877 (gimple_stmt_nonzero_p): Likewise.
15878 (vrp_int_const_binop): Likewise. Remove unreachable case.
15879 (adjust_range_with_scev): Adjust comments.
15880 (compare_range_with_value): Likewise.
15881 (extract_range_from_phi_node): Likewise.
15882 (test_for_singularity): Likewise.
15884 2017-05-09 Richard Biener <rguenther@suse.de>
15886 * tree-vrp.c (get_single_symbol): Add assert that we don't
15887 get overflowed constants as invariant part.
15888 (compare_values_warnv): Add comment before the TREE_NO_WARNING
15889 checks. Use wi::cmp instead of recursing for integer constants.
15890 (compare_values): Just ignore whether we assumed undefined
15891 overflow instead of failing the compare.
15892 (extract_range_for_var_from_comparison_expr): Add comment before the
15893 TREE_NO_WARNING sets.
15894 (test_for_singularity): Likewise.
15895 (extract_range_from_comparison): Do not disable optimization
15896 when we assumed undefined overflow.
15897 (extract_range_basic): Remove init of unused var.
15899 2017-05-09 Richard Biener <rguenther@suse.de>
15901 * tree-vrp.c (vrp_int_const_binop): Use wide-ints and simplify.
15902 (extract_range_from_multiplicative_op_1): Adjust.
15903 (extract_range_from_binary_expr_1): Use int_const_binop.
15905 2017-05-08 Kelvin Nilsen <kelvin@gcc.gnu.org>
15908 * config/rs6000/power6.md: Replace store_data_bypass_p calls with
15909 rs6000_store_data_bypass_p in seven define_bypass directives and
15910 in several comments.
15911 * config/rs6000/rs6000-protos.h: Add prototype for
15912 rs6000_store_data_bypass_p function.
15913 * config/rs6000/rs6000.c (rs6000_store_data_bypass_p): New
15914 function implements slightly different (rs6000-specific) semantics
15915 than store_data_bypass_p, returning false rather than aborting
15916 with assertion error when arguments do not satisfy the
15917 requirements of store data bypass.
15918 (rs6000_adjust_cost): Replace six calls of store_data_bypass_p with
15919 rs6000_store_data_bypass_p.
15921 2017-05-08 Max Filippov <jcmvbkbc@gmail.com>
15923 * config/xtensa/xtensa-protos.h
15924 (xtensa_initial_elimination_offset): New declaration.
15925 * config/xtensa/xtensa.c (xtensa_initial_elimination_offset):
15926 New function. Move its body from the INITIAL_ELIMINATION_OFFSET
15927 macro definition, add case for FRAME_POINTER_REGNUM when
15928 FRAME_GROWS_DOWNWARD.
15929 * config/xtensa/xtensa.h (FRAME_GROWS_DOWNWARD): New macro definition.
15930 (INITIAL_ELIMINATION_OFFSET): Replace body with call to
15931 xtensa_initial_elimination_offset.
15933 2017-05-08 Nathan Sidwell <nathan@acm.org>
15935 * doc/invoke.texi: Alphabetize -fdump options.
15937 2017-05-08 Martin Sebor <msebor@redhat.com>
15939 PR translation/80280
15940 * config/sol2-c.c (solaris_pragma_align): Correct quoting.
15942 2017-05-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
15944 * target.def (compute_frame_layout): New optional target hook.
15945 * doc/tm.texi.in (TARGET_COMPUTE_FRAME_LAYOUT): Add hook.
15946 * doc/tm.texi (TARGET_COMPUTE_FRAME_LAYOUT): Add documentation.
15947 * lra-eliminations.c (update_reg_eliminate): Call compute_frame_layout
15949 * reload1.c (verify_initial_elim_offsets): Likewise.
15950 * config/arm/arm.c (TARGET_COMPUTE_FRAME_LAYOUT): Define.
15951 (use_simple_return_p): Call arm_compute_frame_layout if needed.
15952 (arm_get_frame_offsets): Split up into this ...
15953 (arm_compute_frame_layout): ... and this function.
15955 2017-05-08 Richard Sandiford <richard.sandiford@arm.com>
15957 * config/aarch64/constraints.md (Usa): New constraint.
15958 * config/aarch64/aarch64.md (*movsi_aarch64, *movdi_aarch64): Use it.
15960 2017-05-08 Thomas Preud'homme <thomas.preudhomme@arm.com>
15962 * config.gcc (arm*-*-*): Set TM_MULTILIB_CONFIG from
15963 with_multilib_list after it has been checked.
15965 2017-05-08 Richard Biener <rguenther@suse.de>
15967 * tree-ssa-pre.c (bitmap_set_and): Avoid bitmap copy.
15968 (bitmap_set_subtract_values): Likewise.
15970 2017-05-08 Richard Biener <rguenther@suse.de>
15972 * tree-vrp.c (gimple_assign_nonzero_warnv_p): Rename to ...
15973 (gimple_assign_nonzero): ... this and remove strict_overflow_p
15975 (gimple_stmt_nonzero_warnv_p): Rename to ...
15976 (gimple_stmt_nonzero_p): ... this and remove strict_overflow_p
15978 (vrp_stmt_computes_nonzero): Remove strict_overflow_p argument.
15979 (extract_range_basic): Adjust, do not disable propagation on
15980 strict overflow sensitive simplification.
15981 (vrp_visit_cond_stmt): Likewise.
15983 2017-05-05 Jan Hubicka <hubicka@ucw.cz>
15985 * ipa-inline-analysis.c (estimate_function_body_sizes): Recompute
15986 body size unconditionally.
15988 2017-05-07 Jeff Law <law@redhat.com>
15991 2017-05-06 Jeff Law <law@redhat.com>
15992 PR tree-optimization/78496
15993 * tree-vrp.c (simplify_assert_expr_using_ranges): Remove debugging
15996 PR tree-optimization/78496
15997 * tree-vrp.c (simplify_assert_expr_using_ranges): New function.
15998 (simplify_stmt_using_ranges): Call it.
15999 (vrp_dom_walker::before_dom_children): Extract equivalences
16000 from an ASSERT_EXPR with an equality comparison against a
16003 2017-05-06 Jeff Law <law@redhat.com>
16005 PR tree-optimization/78496
16006 * tree-vrp.c (simplify_assert_expr_using_ranges): Remove debugging
16009 PR tree-optimization/78496
16010 * tree-vrp.c (simplify_assert_expr_using_ranges): New function.
16011 (simplify_stmt_using_ranges): Call it.
16012 (vrp_dom_walker::before_dom_children): Extract equivalences
16013 from an ASSERT_EXPR with an equality comparison against a
16016 2017-05-06 Richard Sandiford <richard.sandiford@linaro.org>
16018 * lra-constraints.c (lra_copy_reg_equiv): New function.
16019 (split_reg): Use it to copy equivalence information from the
16020 original register to the spill register.
16022 2017-05-06 Richard Sandiford <richard.sandiford@linaro.org>
16024 PR rtl-optimization/75964
16025 * simplify-rtx.c (simplify_const_relational_operation): Remove
16026 invalid handling of comparisons of integer ABS.
16028 2017-05-06 Uros Bizjak <ubizjak@gmail.com>
16030 * config/i386/i386.c (ext_80387_constant_init): Do not explicitly
16031 initialize to zero.
16032 (init_regs): Remove declaration.
16033 (function_arg_advance_32): Initialize error_p as boolean variable.
16035 2017-05-05 Nathan Sidwell <nathan@acm.org>
16037 * store-motion.c (remove_reachable_equiv_notes): Reformat long
16038 lines. Use for (;;).
16040 2017-05-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16042 * config/rs6000/rs6000.c (rs6000_vect_nonmem): New static var.
16043 (rs6000_init_cost): Initialize rs6000_vect_nonmem.
16044 (rs6000_add_stmt_cost): Update rs6000_vect_nonmem.
16045 (rs6000_finish_cost): Avoid vectorizing simple copy loops with
16046 VF=2 that require versioning.
16048 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16050 * diagnostic.h (CARET_LINE_MARGIN): Convert from macro to const
16053 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16055 * diagnostic.h (diagnostic_override_option_index): Convert from
16056 macro to inline function.
16058 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16060 * diagnostic.c (last_module_changed_p): New function.
16061 (set_last_module): New function.
16062 (diagnostic_report_current_module): Convert macro usage to
16063 the above functions.
16064 * diagnostic.h (diagnostic_context::last_module): Strengthen
16065 from const line_map * to const line_map_ordinary *.
16066 (diagnostic_last_module_changed): Delete macro.
16067 (diagnostic_set_last_module): Delete macro.
16069 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16071 * diagnostic.c (diagnostic_impl): Replace report_diagnostic
16072 with diagnostic_report_diagnostic.
16073 (diagnostic_n_impl_richloc): Likewise.
16074 * diagnostic.h (report_diagnostic): Delete macro.
16075 * rtl-error.c (diagnostic_for_asm): Replace report_diagnostic
16076 with diagnostic_report_diagnostic.
16077 * substring-locations.c (format_warning_va): Likewise.
16079 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16081 * diagnostic.c (diagnostic_report_diagnostic): Eliminate
16082 save/restor of format_spec. Move option-printing code to...
16083 (print_option_information): ...this new function, and
16084 reimplement by simply printing to the pretty_printer,
16085 rather than appending to the format string.
16087 2017-05-05 David Malcolm <dmalcolm@redhat.com>
16089 * diagnostic.c (diagnostic_report_diagnostic): Split out pragma
16090 handling logic into...
16091 (update_effective_level_from_pragmas): ...this new function.
16093 2017-05-04 Andrew Waterman <andrew@sifive.com>
16095 * config/riscv/riscv.opt (mstrict-align): New option.
16096 * config/riscv/riscv.h (STRICT_ALIGNMENT): Use it. Update comment.
16097 (SLOW_UNALIGNED_ACCESS): Define.
16098 (riscv_slow_unaligned_access): Declare.
16099 * config/riscv/riscv.c (riscv_tune_info): Add slow_unaligned_access
16101 (riscv_slow_unaligned_access): New variable.
16102 (rocket_tune_info): Set slow_unaligned_access to true.
16103 (optimize_size_tune_info): Set slow_unaligned_access to false.
16104 (riscv_cpu_info_table): Add entry for optimize_size_tune_info.
16105 (riscv_valid_lo_sum_p): Use TARGET_STRICT_ALIGN.
16106 (riscv_option_override): Set riscv_slow_unaligned_access.
16107 * doc/invoke.texi: Add -mstrict-align to RISC-V.
16109 2017-05-04 Kito Cheng <kito.cheng@gmail.com>
16111 * config/riscv/riscv.md: Unify indentation.
16113 2017-05-05 Michael Meissner <meissner@linux.vnet.ibm.com>
16118 * config/rs6000/rs6000.md (u code attribute): Add FIX and
16120 (extendsi<mode>2): Add support for doing sign extension via
16121 VUPKHSW and XXPERMDI if the value is in Altivec registers and we
16122 don't have ISA 3.0 instructions.
16123 (extendsi<mode>2 splitter): Likewise.
16124 (fix_trunc<mode>si2): If we are at ISA 2.07 (VSX small integer),
16125 generate the normal insns since SImode can now go in vector
16126 registers. Disallow the special UNSPECs needed for previous
16127 machines to hide SImode being used. Add new insns
16128 fctiw{,w}_<mode>_smallint if SImode can go in vector registers.
16129 (fix_trunc<mode>si2_stfiwx): Likewise.
16130 (fix_trunc<mode>si2_internal): Likewise.
16131 (fixuns_trunc<mode>si2): Likewise.
16132 (fixuns_trunc<mode>si2_stfiwx): Likewise.
16133 (fctiw<u>z_<mode>_smallint): Likewise.
16134 (fctiw<u>z_<mode>_mem): New combiner pattern to prevent conversion
16135 of floating point to 32-bit integer from doing a direct move to
16136 the GPR registers to do a store.
16137 (fctiwz_<mode>): Break long line.
16139 2017-05-05 Bin Cheng <bin.cheng@arm.com>
16141 * Makefile.in (GTFILES): Add tree-ssa-loop-ivopts.c.
16142 * tree-ssa-loop-ivopts.c (compute_max_addr_offset): Delete.
16143 (addr_list, addr_offset_valid_p): New.
16144 (split_address_groups): Check offset validity with above function.
16145 (gt-tree-ssa-loop-ivopts.h): Include header file.
16147 2017-05-05 Nathan Sidwell <nathan@acm.org>
16149 * config.gcc (arm*-*-*): Add missing 'fi'.
16151 2017-05-05 Steve Ellcey <sellcey@cavium.com>
16153 * doc/invoke.texi (-fopt-info): Explicitly say order of options
16154 included in -fopt-info does not matter.
16155 * doc/optinfo.texi (-fopt-info): Fix description of default
16156 behavour. Explicitly say order of options included in -fopt-info
16159 2017-05-05 Thomas Preud'homme <thomas.preudhomme@arm.com>
16161 * config.gcc: Allow combinations of aprofile and rmprofile values for
16162 --with-multilib-list.
16163 * config/arm/t-multilib: New file.
16164 * config/arm/t-aprofile: Remove initialization of MULTILIB_*
16165 variables. Remove setting of ISA and floating-point ABI in
16166 MULTILIB_OPTIONS and MULTILIB_DIRNAMES. Set architecture and FPU in
16167 MULTI_ARCH_OPTS_A and MULTI_ARCH_DIRS_A rather than MULTILIB_OPTIONS
16168 and MULTILIB_DIRNAMES respectively. Add comment to introduce all
16169 matches. Add architecture matches for marvel-pj4 and generic-armv7-a
16171 * config/arm/t-rmprofile: Likewise except for the matches changes.
16172 * doc/install.texi (--with-multilib-list): Document the combination of
16173 aprofile and rmprofile values and warn about pitfalls in doing that.
16175 2017-05-05 Wilco Dijkstra <wdijkstr@arm.com>
16177 * config/aarch64/aarch64.md (movsi_aarch64): Remove '*' from r=w.
16178 (movdi_aarch64): Likewise.
16180 2017-05-05 Jakub Jelinek <jakub@redhat.com>
16182 PR tree-optimization/80632
16183 * tree-switch-conversion.c (struct switch_conv_info): Add target_vop
16185 (build_arrays): Initialize it for virtual phis.
16186 (fix_phi_nodes): Use it for virtual phis.
16188 PR tree-optimization/80558
16189 * tree-vrp.c (extract_range_from_binary_expr_1): Optimize
16190 [x, y] op z into [x op, y op z] for op & or | if conditions
16193 2017-05-05 Andre Vieira <andre.simoesdiasvieira@arm.com>
16194 Prakhar Bahuguna <prakhar.bahuguna@arm.com>
16197 * config/arm/arm.md (use_literal_pool): Remove.
16198 (64-bit immediate split): No longer takes cost into consideration
16199 if arm_disable_literal_pool is enabled.
16200 * config/arm/arm.c (arm_tls_referenced_p): Add diagnostic if TLS is
16201 used when arm_disable_literal_pool is enabled.
16202 (arm_max_const_double_inline_cost): Remove use of
16203 arm_disable_literal_pool.
16204 (push_minipool_fix): Add assert.
16205 (arm_reorg): Add return if arm_disable_literal_pool is enabled.
16206 * config/arm/vfp.md (no_literal_pool_df_immediate): New.
16207 (no_literal_pool_sf_immediate): New.
16209 2017-05-05 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
16211 PR tree-optimization/80613
16212 * tree-ssa-dce.c (propagate_necessity): Remove cases for
16213 BUILT_IN_STRDUP and BUILT_IN_STRNDUP.
16215 2017-05-05 Richard Biener <rguenther@suse.de>
16217 * tree-ssa-pre.c (get_or_alloc_expr_for): Simplify.
16219 2017-05-05 Georg-Johann Lay <avr@gjlay.de>
16221 * config/avr/avr.md [flag_strict_overflow]: Remove any occurence
16222 of this flag from insn conditions due to removal from r247495.
16224 2017-05-05 Wilco Dijkstra <wdijkstr@arm.com>
16226 * config/arm/aarch-common.c (arm_early_load_addr_dep_ptr):
16228 (arm_early_store_addr_dep_ptr): Likewise.
16229 * config/arm/aarch-common-protos.h
16230 (arm_early_load_addr_dep_ptr): Add prototype.
16231 (arm_early_store_addr_dep_ptr): Likewise.
16232 * config/arm/cortex-a53.md: Add new bypasses.
16234 2017-05-05 Jakub Jelinek <jakub@redhat.com>
16236 * tree.c (next_type_uid): Change type to unsigned.
16237 (type_hash_canon): Decrement back next_type_uid if
16238 freeing a type node with the highest TYPE_UID. For INTEGER_TYPEs
16239 also ggc_free TYPE_MIN_VALUE, TYPE_MAX_VALUE and TYPE_CACHED_VALUES
16242 2017-05-04 Martin Sebor <msebor@redhat.com>
16244 * builtins.c: Fix a trivial typo in a comment.
16246 PR middle-end/79234
16247 * builtins.c (check_sizes): Adjust to handle reading past the end.
16248 Avoid printing excessive upper bound of ranges. Use %E to print
16249 tree nodes instead of converting them to %wu.
16250 (expand_builtin_memchr): New function.
16251 (compute_dest_size): Rename...
16252 (compute_objsize): ...to this.
16253 (expand_builtin_memcpy): Adjust.
16254 (expand_builtin_mempcpy): Adjust.
16255 (expand_builtin_strcat): Adjust.
16256 (expand_builtin_strcpy): Adjust.
16257 (check_strncat_sizes): Adjust.
16258 (expand_builtin_strncat): Adjust.
16259 (expand_builtin_strncpy): Adjust and simplify.
16260 (expand_builtin_memset): Adjust.
16261 (expand_builtin_bzero): Adjust.
16262 (expand_builtin_memcmp): Adjust.
16263 (expand_builtin): Handle memcmp.
16264 (maybe_emit_chk_warning): Check strncat just once.
16266 2017-05-04 Martin Sebor <msebor@redhat.com>
16268 PR preprocessor/79214
16269 PR middle-end/79222
16270 PR middle-end/79223
16271 * builtins.c (check_sizes): Add inlining context and issue
16272 warnings even when -Wno-system-headers is set.
16273 (check_strncat_sizes): Same.
16274 (expand_builtin_strncat): Same.
16275 (expand_builtin_memmove): New function.
16276 (expand_builtin_stpncpy): Same.
16277 (expand_builtin): Handle memmove and stpncpy.
16279 2017-05-04 Bin Cheng <bin.cheng@arm.com>
16281 * tree-ssa-loop-ivopts.c (struct cost_pair): Remove field inv_expr
16282 which is not used any more.
16284 2017-05-04 Wilco Dijkstra <wdijkstr@arm.com>
16286 * config/aarch64/aarch64.c (generic_tunings): Update prefetch model.
16288 2017-05-04 Wilco Dijkstra <wdijkstr@arm.com>
16290 * config/aarch64/aarch64.c (cortexa35_tunings): Set jump alignment to 4.
16291 (cortexa53_tunings): Likewise.
16292 (cortexa57_tunings): Likewise.
16293 (cortexa72_tunings): Likewise.
16294 (cortexa73_tunings): Likewise.
16296 2017-05-04 Wilco Dijkstra <wdijkstr@arm.com>
16298 * config/aarch64/aarch64.c (generic_tunings): Set jump alignment to 4.
16299 Set loop alignment to 8.
16301 2017-05-04 Martin Sebor <msebor@redhat.com>
16303 PR translation/80280
16304 * builtins.c (expand_builtin_object_size): Add missing quoting to
16305 %D and like directives.
16306 * hsa-gen.c (hsa_type_for_scalar_tree_type): Same.
16307 (hsa_type_for_tree_type): Same.
16308 (verify_function_arguments): Same.
16309 * symtab.c (symbol_table::change_decl_assembler_name): Same.
16310 * varasm.c (get_section): Same.
16313 2017-05-04 Martin Sebor <msebor@redhat.com>
16315 PR translation/80280
16316 * config/i386/i386.c (ix86_function_versions): Quote a %D directive.
16318 2017-05-04 Wilco Dijkstra <wdijkstr@arm.com>
16320 * config/aarch64/aarch64.c (generic_addrcost_table):
16321 Change HI/TI mode setting.
16323 2017-05-04 Martin Jambor <mjambor@suse.cz>
16325 PR tree-optimization/80622
16326 * tree-sra.c (comes_initialized_p): New function.
16327 (build_accesses_from_assign): Only set write lazily when
16328 comes_initialized_p is false.
16329 (analyze_access_subtree): Use comes_initialized_p.
16330 (propagate_subaccesses_across_link): Assert !comes_initialized_p
16331 instead of testing for PARM_DECL.
16333 2017-05-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16335 * config/aarch64/aarch64.md (prefetch); Adjust predicate and
16336 constraint on operand 0 to allow more general addressing modes.
16337 Adjust output template.
16338 * config/aarch64/aarch64.c (aarch64_address_valid_for_prefetch_p):
16340 * config/aarch64/aarch64-protos.h
16341 (aarch64_address_valid_for_prefetch_p): Declare prototype.
16342 * config/aarch64/constraints.md (Dp): New address constraint.
16343 * config/aarch64/predicates.md (aarch64_prefetch_operand): New
16346 2017-05-04 Jan Hubicka <hubicka@ucw.cz>
16348 * ipa-cp.c (perform_estimation_of_a_value): Drop base_time parameter;
16349 update use of estimate_ipcp_clone_size_and_time.
16350 (estimate_local_effects): Update use of
16351 estimate_ipcp_clone_size_and_time and perform_estimation_of_a_value.
16352 * ipa-inline.h (estimate_ipcp_clone_size_and_time): Update prototype.
16353 * ipa-inline-analysis.c (estimate_ipcp_clone_size_and_time):
16354 Return nonspecialized time.
16356 2017-05-04 Richard Biener <rguenther@suse.de>
16358 * tree-ssa-alias.c (get_continuation_for_phi): Improve looking
16359 for the last VUSE which def dominates the PHI. Directly call
16361 (get_continuation_for_phi_1): Remove.
16363 2017-05-04 Richard Sandiford <richard.sandiford@linaro.org>
16365 * tree-ssa-loop-manip.c (niter_for_unrolled_loop): Add commentary
16366 to explain the use of truncating division. Cap the number of
16367 iterations to the maximum given by nb_iterations_upper_bound,
16370 2017-05-04 Thomas Preud'homme <thomas.preudhomme@arm.com>
16372 * configure.ac (--enable-mingw-wildcard): Add new configurable feature.
16373 * configure: Regenerate.
16374 * config.in: Regenerate.
16375 * config/i386/driver-mingw32.c: new file.
16376 * config/i386/x-mingw32: Add rule to build driver-mingw32.o.
16377 * config.host: Link driver-mingw32.o on MinGW host.
16378 * doc/install.texi: Document new --enable-mingw-wildcard configure
16381 2017-05-04 Marek Polacek <polacek@redhat.com>
16383 PR tree-optimization/80612
16384 * calls.c (get_size_range): Check for INTEGRAL_TYPE_P.
16386 2017-05-04 Prakhar Bahuguna <prakhar.bahuguna@arm.com>
16387 Andre Simoes Dias Vieira <andre.simoesdiasvieira@arm.com>
16389 * config/arm/arm.md (movsi): Change TARGET_32BIT to TARGET_HAVE_MOVT.
16390 (movt splitter): Likewise.
16391 * config/arm/arm.c (arm_option_check_internal): Change arm_arch_thumb2
16392 to TARGET_HAVE_MOVT, and merge with -mslow-flash-data check.
16393 (const_ok_for_arm): Change else to else if (TARGET_THUMB2) and add else
16394 block for Thumb-1 with MOVT.
16395 (thumb2_legitimate_address_p): Move code block ...
16396 (can_avoid_literal_pool_for_label_p): ... into this new function.
16397 (thumb1_legitimate_address_p): Add check for TARGET_HAVE_MOVT and
16399 (thumb_legitimate_constant_p): Add conditional on TARGET_HAVE_MOVT
16400 * doc/invoke.texi (-mpure-code): Change "ARMv7-M targets" for
16401 "M-profile targets with the MOVT instruction".
16403 2017-05-04 Prakhar Bahuguna <prakhar.bahuguna@arm.com>
16405 * config/arm/arm-builtins.c (arm_init_builtins): Rename
16406 __builtin_arm_ldfscr to __builtin_arm_get_fpscr, and rename
16407 __builtin_arm_stfscr to __builtin_arm_set_fpscr.
16409 2017-05-04 Martin Liska <mliska@suse.cz>
16411 * tree-vrp.c (simplify_cond_using_ranges_2): Remove unused
16412 variable cond_code.
16414 2017-05-04 Richard Biener <rguenther@suse.de>
16416 * tree.c (array_at_struct_end_p): Handle arrays at struct
16417 end with flexarrays more conservatively. Refactor and treat
16418 arrays of arrays or aggregates more strict. Fix
16419 VIEW_CONVERT_EXPR handling. Remove allow_compref argument.
16420 * tree.c (array_at_struct_end_p): Adjust prototype.
16421 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
16422 * gimple-fold.c (get_range_strlen): Likewise.
16423 * tree-chkp.c (chkp_may_narrow_to_field): Likewise.
16425 2017-05-04 Richard Biener <rguenther@suse.de>
16427 PR tree-optimization/31130
16428 * tree-vrp.c (needs_overflow_infinity): Remove as always returning
16430 (supports_overflow_infinity): Likewise.
16431 (is_negative_overflow_infinity): Likewise.
16432 (is_positive_overflow_infinity): Likewise.
16433 (is_overflow_infinity): Likewise.
16434 (stmt_overflow_infinity): Likewise.
16435 (overflow_infinity_range_p): Likewise.
16436 (usable_range_p): Remove as always returning true.
16437 (make_overflow_infinity): Remove.
16438 (negative_overflow_infinity): Likewise.
16439 (positive_overflow_infinity): Likewise.
16440 (avoid_overflow_infinity): Likewise.
16441 (set_value_range): Adjust accordingly.
16442 (set_value_range_to_nonnegative): Likewise, remove now unused
16443 overflow_infinity arg.
16444 (vrp_operand_equal_p): Adjust.
16445 (update_value_range): Likewise.
16446 (range_int_cst_singleton_p): Likewise.
16447 (operand_less_p): Likewise.
16448 (compare_values_warnv): Likewise.
16449 (extract_range_for_var_from_comparison_expr): Likewise.
16450 (vrp_int_const_binop): Likewise.
16451 (zero_nonzero_bits_from_vr): Likewise.
16452 (extract_range_from_multiplicative_op_1): Likewise.
16453 (extract_range_from_binary_expr_1): Likewise.
16454 (extract_range_from_unary_expr): Likewise.
16455 (extract_range_from_comparison): Likewise.
16456 (extract_range_basic): Likewise.
16457 (adjust_range_with_scev): Likewise.
16458 (compare_ranges): Likewise.
16459 (compare_range_with_value): Likewise.
16460 (dump_value_range): Likewise.
16461 (test_for_singularity): Likewise, remove strict_overflow_p parameter
16463 (simplify_cond_using_ranges): Adjust.
16465 2017-05-04 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
16467 * brig-builtins.def: Added a builtin for class_f64.
16468 * builtin-types.def: Added a builtin type needed by class_f64.
16470 2017-05-03 Jason Merrill <jason@redhat.com>
16472 * timevar.def: Add TV_CONSTEXPR.
16474 2017-05-03 David Malcolm <dmalcolm@redhat.com>
16476 * common.opt (fdiagnostics-parseable-fixits): Fix typo.
16478 2017-05-03 Martin Jambor <mjambor@suse.cz>
16480 * ipa-prop.c (ipa_update_after_lto_read): Removed.
16481 * ipa-prop.h (ipa_update_after_lto_read): Remove declaration.
16482 * ipa-cp.c (ipcp_propagate_stage): Do not call
16483 ipa_update_after_lto_read.
16484 * ipa-inline.c (ipa_inline): Likewise.
16486 2017-05-03 Martin Jambor <mjambor@suse.cz>
16488 * ipa-prop.h (ipa_edge_args): Make a class. Mark with for_user GTY
16489 tag. Added a default constructor and a destructor.
16490 (ipa_edge_args_sum_t): New class;
16491 (ipa_edge_args_sum): Declare.
16492 (ipa_edge_args_vector): Remove declaration.
16493 (IPA_EDGE_REF): Use ipa_edge_args_sum.
16494 (ipa_free_edge_args_substructures): Remove declaration.
16495 (ipa_check_create_edge_args): Use ipa_edge_args_sum.
16496 (ipa_edge_args_info_available_for_edge_p): Likewise.
16497 * ipa-prop.c (ipa_edge_args_vector): Removed.
16498 (edge_removal_hook_holder): Likewise.
16499 (edge_duplication_hook_holder): Likewise.
16500 (ipa_edge_args_sum): New variable.
16501 (ipa_propagate_indirect_call_infos): Test ipa_edge_args_sum instead of
16502 ipa_edge_args_vector.
16503 (ipa_free_edge_args_substructures): Likewise.
16504 (ipa_free_all_edge_args): Free ipa_edge_args_sum instead of
16505 ipa_edge_args_vector.
16506 (ipa_edge_removal_hook): Turned into method
16507 ipa_edge_args_sum_t::remove.
16508 (ipa_edge_duplication_hook): Turned into method
16509 ipa_edge_args_sum_t::duplicate.
16510 (ipa_register_cgraph_hooks): Create ipa_edge_args_sum instead of
16511 registering edge hooks.
16512 (ipa_unregister_cgraph_hooks): Do not unregister edge hooks.
16513 * ipa-inline-analysis.c (estimate_function_body_sizes): Test
16514 ipa_edge_args_sum instead of ipa_edge_args_vector.
16515 * ipa-profile.c (ipa_profile): Likewise.
16517 2017-05-03 Martin Jambor <mjambor@suse.cz>
16519 * symbol-summary.h (function_summary): New method exists.
16520 (function_summary::symtab_removal): Deallocate through release.
16521 (call_summary): New class.
16522 (gt_ggc_mx): New overload.
16523 (gt_pch_nx): Likewise.
16524 (gt_pch_nx): Likewise.
16526 2017-05-03 Jeff Law <law@redhat.com>
16528 PR tree-optimization/78496
16529 * tree-vrp.c (simplify_cond_using_ranges_1): Renamed
16530 from simplify_cond_using_ranges. Split off code to walk
16531 backwards through casts into ...
16532 (simplify_cond_using_ranges_2): New function.
16533 (simplify_stmt_using_ranges): Call simplify_cond_using_ranges_1.
16534 (execute_vrp): After identifying jump threads, call
16535 simplify_cond_using_ranges_2.
16537 2017-05-03 Jan Hubicka <hubicka@ucw.cz>
16540 * ipa-inline.h (inline_summary): Add ctor.
16541 (create_ggc): Do not use ggc_cleared_alloc.
16543 2017-05-03 Jeff Downs <heydowns@somuchpressure.net>
16544 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16546 * gcc.c (handle_braces): Support escaping in switch matching
16548 * doc/invoke.texi (Spec Files): Document it.
16549 Remove superfluous @code markup in items.
16551 2017-05-03 David Malcolm <dmalcolm@redhat.com>
16553 * diagnostic-show-locus.c (struct column_range): New struct.
16554 (get_affected_columns): New function.
16555 (get_printed_columns): New function.
16556 (struct correction): New struct.
16557 (correction::ensure_capacity): New function.
16558 (correction::ensure_terminated): New function.
16559 (struct line_corrections): New struct.
16560 (line_corrections::~line_corrections): New dtor.
16561 (line_corrections::add_hint): New function.
16562 (layout::print_trailing_fixits): Reimplement in terms of the new
16564 (selftest::test_overlapped_fixit_printing): New function.
16565 (selftest::diagnostic_show_locus_c_tests): Call it.
16567 2017-05-03 Nathan Sidwell <nathan@acm.org>
16569 Canonicalize canonical type hashing
16570 * tree.h (type_hash_canon_hash): Declare.
16571 * tree.c (type_hash_list, attribute_hash_list): Move into
16572 type_hash_canon_hash.
16573 (build_type_attribute_qual_variant): Break out hash code calc into
16574 type_hash_canon_hash.
16575 (type_hash_canon_hash): New. Generic type hash computation.
16576 (build_range_type_1, build_array_type_1, build_function_type,
16577 build_method_type_directly, build_offset_type, build_complex_type,
16578 make_vector_type): Call it.
16580 2017-05-03 Richard Biener <rguenther@suse.de>
16582 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16583 When all DRs have unknown misaligned do not always peel
16584 when there is a store but apply the same costing model as if
16585 there were only loads.
16587 2017-05-03 Richard Biener <rguenther@suse.de>
16590 PR tree-optimization/80492
16591 * tree-ssa-alias.c (decl_refs_may_alias_p): Handle
16592 compare_base_decls returning dont-know properly.
16594 2017-05-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
16596 * config/arm/iterators.md (CCSI): New mode iterator.
16597 (arch): New mode attribute.
16598 * config/arm/sync.md (atomic_compare_and_swap<mode>_1): Rename into ...
16599 (atomic_compare_and_swap<CCSI:arch><NARROW:mode>_1): This and ...
16600 (atomic_compare_and_swap<CCSI:arch><SIDI:mode>_1): This. Use CCSI
16601 code iterator for success result mode.
16602 * config/arm/arm.c (arm_expand_compare_and_swap): Adapt code to use
16603 the corresponding new insn generators.
16605 2017-05-03 Bin Cheng <bin.cheng@arm.com>
16608 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16609 * rtlanal.c (rtx_cost): Handle TRUNCATE between tieable modes.
16611 2017-05-03 Richard Sandiford <richard.sandiford@linaro.org>
16613 * tree-data-ref.h (SUB_CONFLICTS_IN_A): Wrap SUB argument in brackets.
16614 (SUB_CONFLICTS_IN_B, SUB_LAST_CONFLICT, SUB_DISTANCE): Likewise.
16615 (DDR_A): Wrap DDR argument in brackets.
16616 (DDR_B, DDR_AFFINE_P, DDR_ARE_DEPENDENT, DDR_SUBSCRIPTS): Likewise.
16617 (DDR_LOOP_NEST, DDR_INNER_LOOP, DDR_SELF_REFERENCE): Likewise.
16618 (DDR_REVERSED_P): Likewise.
16620 2017-05-03 Jakub Jelinek <jakub@redhat.com>
16622 PR tree-optimization/79472
16623 * tree-switch-conversion.c (struct switch_conv_info): Add
16624 contiguous_range and default_case_nonstandard fields.
16625 (collect_switch_conv_info): Compute contiguous_range and
16626 default_case_nonstandard fields, don't clear final_bb if
16627 contiguous_range and only the default case doesn't have the required
16629 (check_all_empty_except_final): Set default_case_nonstandard instead
16630 of failing if contiguous_range and the default case doesn't have empty
16632 (check_final_bb): Add SWTCH argument, don't fail if contiguous_range
16633 and only the default case doesn't have the required constants. Skip
16635 (gather_default_values): Skip virtual phis. Allow non-NULL CASE_LOW
16636 if default_case_nonstandard.
16637 (build_constructors): Build constant 1 just once. Assert that default
16638 values aren't inserted in between cases if contiguous_range. Skip
16640 (build_arrays): Skip virtual phis.
16641 (prune_bbs): Add DEFAULT_BB argument, don't remove that bb.
16642 (fix_phi_nodes): Don't add e2f phi arg if default_case_nonstandard.
16643 Handle virtual phis.
16644 (gen_inbound_check): Handle default_case_nonstandard case.
16645 (process_switch): Adjust check_final_bb caller. Call
16646 gather_default_values with the first non-default case instead of
16647 default case if default_case_nonstandard.
16649 2017-05-02 Nathan Sidwell <nathan@acm.org>
16651 * ggc-page.c (move_ptes_to_front): Replace unsigned >0 with i--
16652 check. Fix formatting.
16654 2017-05-02 Jan Hubicka <hubicka@ucw.cz>
16656 * ipa-inline-analysis.c (estimate_node_size_and_time): Allow roundoff
16657 errors when comparing specialized and unspecialized times.
16659 2017-05-02 David Malcolm <dmalcolm@redhat.com>
16661 * diagnostic-show-locus.c
16662 (layout::should_print_annotation_line_p): Make private.
16663 (layout::print_annotation_line): Make private.
16664 (layout::annotation_line_showed_range_p): Make private.
16665 (layout::show_ruler): Make private.
16666 (layout::print_source_line): Make private. Pass in line and
16667 line_width, rather than calling location_get_source_line. Drop
16669 (layout::print_leading_fixits): New method.
16670 (layout::print_any_fixits): Rename to...
16671 (layout::print_trailing_fixits): ...this, and make private.
16672 Don't print newline fixits.
16673 (diagnostic_show_locus): Move logic for printing one row into...
16674 (layout::print_line): ...this new function. Move the
16675 location_get_source_line call and error-handling from
16676 print_source_line to here. Call print_leading_fixits, and rename
16677 print_any_fixits to print_trailing_fixits.
16678 (selftest::test_fixit_insert_containing_newline): Update now that
16679 newlines are partially supported.
16680 (selftest::test_fixit_insert_containing_newline_2): New test.
16681 (selftest::test_fixit_replace_containing_newline): Update comments.
16682 (selftest::diagnostic_show_locus_c_tests): Call the new test.
16683 * edit-context.c (class added_line): New class.
16684 (class edited_line): Describe newline handling in comment.
16685 (edited_line::actually_edited_p): New method.
16686 (edited_line::print_content): Delete redundant decl.
16687 (edited_line::m_predecessors): New field.
16688 (edited_file::print_content): Call edited_line::print_content.
16689 (edited_file::print_diff): Update to support newlines.
16690 (edited_file::print_diff_hunk): Likewise.
16691 (edited_file::print_run_of_changed_lines): New function.
16692 (edited_file::print_diff_line): Convert to...
16693 (print_diff_line): ...this.
16694 (edited_file::get_effective_line_count): New function.
16695 (edited_line::edited_line): Initialize new field m_predecessors.
16696 (edited_line::~edited_line): Clean up m_predecessors.
16697 (edited_line::apply_fixit): Handle newlines.
16698 (edited_line::get_effective_line_count): New function.
16699 (edited_line::print_content): New function.
16700 (edited_line::print_diff_lines): New function.
16701 (selftest::test_applying_fixits_insert_containing_newline): New
16703 (selftest::test_applying_fixits_replace_containing_newline): New
16705 (selftest::insert_line): New function.
16706 (selftest::test_applying_fixits_multiple_lines): Add example of
16708 (selftest::edit_context_c_tests): Call the new tests.
16710 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16712 * tree-ssa-loop-ivopts.c (get_scaled_computation_cost_at): Delete
16713 parameter cand. Update dump information.
16714 (get_computation_cost): Update uses.
16716 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16718 * tree-ssa-loop-ivopts.c (get_computation_aff_1): New.
16719 (get_computation_aff): Reorder parameters. Use get_computation_aff_1.
16720 (get_computation_at, rewrite_use_address): Update use of
16721 get_computation_aff.
16723 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16725 * tree-ssa-loop-ivopts.c (get_computation_at): Reorder parameters.
16726 (get_computation): Delete.
16727 (get_computation_cost): Implement like get_computation_cost_at.
16728 Use get_computation_at.
16729 (get_computation_cost_at): Delete.
16730 (rewrite_use_nonlinear_expr): Use get_computation_at.
16731 (rewrite_use_compare, remove_unused_ivs): Ditto.
16733 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16735 * tree-ssa-loop-ivopts.c (rewrite_use_address): Simple refactor.
16737 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16739 * tree-ssa-loop-ivopts.c (struct iv_ca): Rename n_regs to n_invs.
16740 (ivopts_global_cost_for_size): Rename parameter and update uses.
16741 (iv_ca_recount_cost): Update uses.
16742 (iv_ca_set_remove_invs, iv_ca_set_no_cp): Record invariants and
16743 candidates seperately in n_invs and n_cands.
16744 (iv_ca_set_add_invs, iv_ca_set_cp, iv_ca_new): Ditto.
16746 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16748 * tree-ssa-loop-ivopts.c (struct walk_tree_data): New.
16749 (find_inv_vars_cb): New.
16750 (find_depends): Renamed to ...
16751 (find_inv_vars): ... this.
16752 (add_candidate_1, force_var_cost): Call find_inv_vars.
16753 (split_address_cost, determine_group_iv_cost_cond): Ditto.
16755 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16757 * tree-ssa-loop-ivopts.c (struct cost_pair): Rename depends_on to
16758 inv_vars. Add inv_exprs.
16759 (struct iv_cand): Rename depends_on to inv_vars.
16760 (struct ivopts_data): Rename max_inv_id/n_invariant_uses to
16761 max_inv_var_id/n_inv_var_uses. Move max_inv_expr_id around.
16762 Refactor field used_inv_exprs from has_map to array n_inv_expr_uses.
16763 (dump_cand): Dump inv_vars.
16764 (tree_ssa_iv_optimize_init): Support inv_vars and inv_exprs.
16765 (record_invariant, find_depends, add_candidate_1): Ditto.
16766 (set_group_iv_cost, force_var_cost): Ditto.
16767 (split_address_cost, ptr_difference_cost, difference_cost): Ditto.
16768 (get_computation_cost_at, get_computation_cost): Ditto.
16769 (determine_group_iv_cost_generic): Ditto.
16770 (determine_group_iv_cost_address): Ditto.
16771 (determine_group_iv_cost_cond, autoinc_possible_for_pair): Ditto.
16772 (determine_group_iv_costs): Ditto.
16773 (iv_ca_recount_cost): Update call to ivopts_global_cost_for_size.
16774 (iv_ca_set_remove_invariants): Renamed to ...
16775 (iv_ca_set_remove_invs): ... this. Support inv_vars and inv_exprs.
16776 (iv_ca_set_no_cp): Use iv_ca_set_remove_invs.
16777 (iv_ca_set_add_invariants): Renamed to ...
16778 (iv_ca_set_add_invs): ... this. Support inv_vars and inv_exprs.
16779 (iv_ca_set_cp): Use iv_ca_set_add_invs.
16780 (iv_ca_has_deps): Support inv_vars and inv_exprs.
16781 (iv_ca_new, iv_ca_free, iv_ca_dump, free_loop_data): Ditto.
16782 (create_new_ivs): Remove useless dump.
16784 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16786 * tree-ssa-loop-ivopts.c (get_computation_cost_at): Remove pseudo
16788 (determine_group_iv_cost_cond, determine_iv_cost): Ditto.
16789 (iv_ca_set_no_cp, create_new_iv): Ditto.
16791 2017-05-02 Bin Cheng <bin.cheng@arm.com>
16793 * rtlanal.c (rtx_cost): Handle TRUNCATE between tieable modes.
16795 2017-05-02 Xi Ruoyao <ryxi@stu.xidian.edu.cn>
16797 * tree.h (EXPR_CILK_SPAWN): Use macro TREE_CHECK2 instead of
16798 function tree_check2.
16800 2017-05-02 Martin Liska <mliska@suse.cz>
16802 * doc/gcov.texi: Add missing preposition.
16803 * gcov.c (function_info::function_info): Properly fill up
16804 all member variables.
16806 2017-05-02 Tamar Christina <tamar.christina@arm.com>
16808 * expr.c (expand_expr_real_2): Re-cost if previous costs are the same.
16810 2017-05-02 Tamar Christina <tamar.christina@arm.com>
16812 * simplify-rtx.c (simplify_binary_operation_1): Add LSHIFTRT case.
16814 2017-05-02 Martin Liska <mliska@suse.cz>
16817 * lto-streamer-in.c (lto_read_tree_1): Remove
16818 LTO_STREAMER_DEBUG.
16819 * lto-streamer.c (struct tree_hash_entry): Likewise.
16820 (struct tree_entry_hasher): Likewise.
16821 (tree_entry_hasher::hash): Likewise.
16822 (tree_entry_hasher::equal): Likewise.
16823 (lto_streamer_init): Likewise.
16824 (lto_orig_address_map): Likewise.
16825 (lto_orig_address_get): Likewise.
16826 (lto_orig_address_remove): Likewise.
16827 * lto-streamer.h: Likewise.
16828 * tree-streamer-in.c (streamer_alloc_tree): Likewise.
16829 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
16831 2017-05-02 Sebastian Peryt <sebastian.peryt@intel.com>
16833 * config/i386/avx512fintrin.h (_mm_mask_add_round_sd)
16834 (_mm_maskz_add_round_sd, _mm_mask_add_round_ss)
16835 (mm_maskz_add_round_ss, _mm_mask_sub_round_sd)
16836 (mm_maskz_sub_round_sd, _mm_mask_sub_round_ss)
16837 (mm_maskz_sub_round_ss, _mm_mask_add_sd)
16838 (mm_maskz_add_sd, _mm_mask_add_ss, _mm_maskz_add_ss)
16839 (mm_mask_sub_sd, _mm_maskz_sub_sd, _mm_mask_sub_ss)
16840 (mm_maskz_sub_ss): New intrinsics.
16841 * config/i386/i386-builtin-types.def (V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
16842 (V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT): New function type aliases.
16843 * config/i386/i386-builtin.def (__builtin_ia32_addsd_mask_round)
16844 (__builtin_ia32_addss_mask_round, __builtin_ia32_subsd_mask_round)
16845 (__builtin_ia32_subss_mask_round): New builtins.
16846 * config/i386/i386.c (V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
16847 (V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT): Handle new types.
16848 * config/i386/sse.md (<sse>_vm<plusminus_insn><mode>3<round_name>):
16850 (<sse>_vm<plusminus_insn><mode>3<mask_name><round_name>): ... this.
16851 (v<plusminus_mnemonic><ssescalarmodesuffix>\t{<round_op3>%2, %1, %0|%0, %1, %<iptr>2<round_op3>}):
16853 (v<plusminus_mnemonic><ssescalarmodesuffix>\t{<round_mask_op3>%2, %1, %0<mask_operand3>|%0<mask_operand3>, %1, %<iptr>2<round_mask_op3>}):
16856 2017-05-02 Martin Jambor <mjambor@suse.cz>
16858 PR tree-optimization/78687
16859 * tree-sra.c (access): New field parent.
16860 (process_subtree_disqualification): New function.
16861 (disqualify_candidate): Call it.
16862 (build_accesses_from_assign): Reset write flag if creating an
16864 (build_access_subtree): Fill in parent field and also prpagate
16865 down grp_write flag.
16866 (create_artificial_child_access): New parameter set_grp_write, set
16867 grp_write to its value.
16868 (propagate_subaccesses_across_link): Also propagate grp_write flag
16870 (propagate_all_subaccesses): Push the closest parent back to work
16871 queue if add_access_to_work_queue returned true.
16873 2017-05-02 Richard Biener <rguenther@suse.de>
16875 * common.opt (fstrict-overflow): Alias negative to fwrapv.
16876 * doc/invoke.texi (fstrict-overflow): Remove all traces of
16877 -fstrict-overflow documentation.
16878 * tree.h (TYPE_OVERFLOW_UNDEFINED): Do not test flag_strict_overflow.
16879 (POINTER_TYPE_OVERFLOW_UNDEFINED): Test !flag_wrapv instead of
16880 flag_strict_overflow.
16881 * ipa-inline.c (can_inline_edge_p): Do not test flag_strict_overflow.
16882 * lto-opts.c (lto_write_options): Do not stream it.
16883 * lto-wrapper.c (merge_and_complain): Do not handle it.
16884 * opts.c (default_options_table): Do not set -fstrict-overflow.
16885 (finish_options): Likewise do not clear it when sanitizing.
16886 * simplify-rtx.c (simplify_const_relational_operation): Do not
16887 test flag_strict_overflow.
16889 2017-05-02 Uros Bizjak <ubizjak@gmail.com>
16891 * config/alpha/alpha.md (*add<mode>3_ieee): Merge to add<mode>3
16892 using enabled attribute.
16893 (*sub<mode>3_ieee): Merge to sub<mode>3 using enabled attribute.
16894 (*mul<mode>3_ieee): Merge to mul<mode>3 using enabled attribute.
16895 (*div<mode>3_ieee): Merge to div<mode>3 using enabled attribute.
16896 (*sqrt<mode>2_ieee): Merge to sqrt<mode>2 using enabled attribute.
16897 (*fix_truncdfdi_ieee): Merge to *fix_truncdfdi2 using enabled attribute.
16898 (*fix_truncsfdi_ieee): Merge to *fix_truncsfdi2 using enabled attribute.
16899 (*floatdisf_ieee): Merge to floatdisf2 using enabled attribute.
16900 (*floatdidf_ieee): Merge to floatdidf2 using enabled attribute.
16901 (*truncdfsf2_ieee): Merge to truncdfsf2 using enabled attribute.
16902 (*cmpdf_ieee): Merge to *cmpdf_internal using enabled attribute.
16904 2017-05-02 Uros Bizjak <ubizjak@gmail.com>
16906 * config/i386/i386.c (ix86_code_end): Use {FIRST,LAST}_INT_REG.
16908 2017-05-02 Richard Biener <rguenther@suse.de>
16910 PR tree-optimization/80591
16912 2017-04-10 Richard Biener <rguenther@suse.de>
16914 * tree-ssa-structalias.c (find_func_aliases): Properly handle
16917 2017-05-02 Richard Biener <rguenther@suse.de>
16919 PR tree-optimization/80549
16920 * tree-cfgcleanup.c (mfb_keep_latches): New helper.
16921 (cleanup_tree_cfg_noloop): Create forwarders to known loop
16922 headers if they do not have a preheader.
16924 2017-05-02 Martin Liska <mliska@suse.cz>
16927 * common.opt: Fix typo.
16928 * doc/invoke.texi: Likewise.
16930 2017-05-01 Jan Beulich <jbeulich@suse.com>
16932 * config/i386/sse.md (xop_vpermil2<mode>3): Do not allow operand
16933 swapping, add (x,x,m,x,n) alternative.
16935 2017-05-01 Nathan Sidwell <nathan@acm.org>
16937 * calls.c (combine_pending_stack_adjustment_and_call): Remove
16938 unnecessary unadjusted_alignment check.
16940 2017-05-01 Xi Ruoyao <ryxi@stu.xidian.edu.cn>
16943 * cilk_common.c (expand_builtin_cilk_detach): Move pedigree
16945 * gimplify.c (gimplify_cilk_detach): New function.
16946 (gimplify_call_expr, gimplify_modify_expr): Call it as needed.
16947 * tree-core.h: Document EXPR_CILK_SPAWN.
16948 * tree.h (EXPR_CILK_SPAWN): Define.
16950 2017-05-01 David Malcolm <dmalcolm@redhat.com>
16952 * diagnostic-show-locus.c (layout::get_expanded_location): Rewrite
16953 to use new fixit_hint representation, using the "replace" logic.
16954 (get_line_span_for_fixit_hint): Likewise.
16955 (layout::print_any_fixits): Likewise.
16956 (selftest::test_one_liner_many_fixits): Rename to...
16957 (selftest::test_one_liner_many_fixits_1): ...this, and update
16958 comment and expected output to reflect that the multiple fix-it
16959 hints are now consolidated into one insertion.
16960 (selftest::test_one_liner_many_fixits_2): New test.
16961 (selftest::test_diagnostic_show_locus_one_liner): Update for
16963 (selftest::test_fixit_consolidation): Update for fix-it API
16965 * diagnostic.c (print_parseable_fixits): Likewise.
16966 * edit-context.c (edited_line::m_line_events): Convert from
16967 auto_vec <line_event *> to auto_vec <line_event>.
16968 (class line_event): Convert from abstract base class to a concrete
16969 class, taking over the role of replace_event.
16970 (class insert_event): Delete.
16971 (class replace_event): Rename to class line_event. Convert to
16973 (edit_context::add_fixits): Reimplement.
16974 (edit_context::apply_insert): Delete.
16975 (edit_context::apply_replace): Rename to...
16976 (edit_context::apply_fixit): ...this. Convert to half-open range.
16977 (edited_file::apply_insert): Delete.
16978 (edited_file::apply_replace): Rename to...
16979 (edited_file::apply_fixit): ...this.
16980 (edited_line::~edited_line): Drop deletion of events.
16981 (edited_line::apply_insert): Delete.
16982 (edited_line::apply_replace): Rename to...
16983 (edited_line::apply_fixit): ...this. Convert to half-open range.
16984 Update for change to type of m_line_events.
16985 * edit-context.h (edit_context::apply_insert): Delete.
16986 (edit_context::apply_replace): Rename to...
16987 (edit_context::apply_fixit): ...this.
16989 2017-05-01 Martin Sebor <msebor@redhat.com>
16991 * gimple-ssa-sprintf.c (format_integer): Set knownrange when it's
16994 2017-05-01 Uros Bizjak <ubizjak@gmail.com>
16997 * config/i386/cpuid.h (__get_cpuid): Always return 0 when
16998 __get_cpuid_max returns 0.
16999 (__get_cpuid_count): Ditto.
17001 2017-05-01 Eric Botcazou <ebotcazou@adacore.com>
17003 * tree.c (substitute_in_expr) <tcc_vl_exp>: Also inline a call if the
17004 replacement expression is another instance of one of its arguments.
17006 2017-05-01 Jakub Jelinek <jakub@redhat.com>
17009 * rtlanal.c (reg_set_p): If reg is a stack_pointer_rtx, also
17010 check for stack push/pop autoinc.
17011 * config/i386/i386.c (ix86_agi_dependent): Return false
17012 if the only reason why modified_in_p returned true is that
17013 addr is SP based and set_insn is a push or pop.
17015 2017-04-29 Jan Hubicka <hubicka@ucw.cz>
17017 * ipa-inline.c (compute_inlined_call_time): Remove now unnecesary
17020 2017-04-29 Jan Hubicka <hubicka@ucw.cz>
17023 * ipa-inline-analysis.c (dump_predicate): Add optional parameter NL.
17024 (account_size_time): Use two predicates - exec_pred and
17026 (evaluate_conditions_for_known_args): Compute both clause and
17028 (evaluate_properties_for_edge): Evaulate both clause and nonspec_clause.
17029 (inline_summary_t::duplicate): Update.
17030 (estimate_function_body_sizes): Caluculate exec and nonconst predicates
17032 (compute_inline_parameters): Likewise.
17033 (estimate_edge_size_and_time): Update caluclation of time.
17034 (estimate_node_size_and_time): Compute both time and nonspecialized
17036 (estimate_ipcp_clone_size_and_time): Update.
17037 (inline_merge_summary): Update.
17038 (do_estimate_edge_time): Update.
17039 (do_estimate_edge_size): Update.
17040 (do_estimate_edge_hints): Update.
17041 (inline_read_section, inline_write_summary): Stream both new predicates.
17042 * ipa-inline.c (compute_uninlined_call_time): Take uninlined_call_time
17044 (compute_inlined_call_time): Cleanup.
17045 (big_speedup_p): Update.
17046 (edge_badness): Update.
17047 * ipa-inline.h (INLINE_TIME_SCALE): Remove.
17048 (size_time_entry): Replace predicate by exec_predicate and
17049 nonconst_predicate.
17050 (edge_growth_cache_entry): Cache both time nad nonspecialized time.
17051 (estimate_edge_time): Return also nonspec_time.
17052 (reset_edge_growth_cache): Update.
17054 2017-04-29 Jakub Jelinek <jakub@redhat.com>
17056 PR rtl-optimization/80491
17057 * ifcvt.c (noce_process_if_block): When looking for x setter
17058 with missing else_bb, don't check only the insn right before
17059 cond_earliest, but look for the last insn that x is modified in
17060 within the same bb.
17062 PR rtl-optimization/80491
17063 * alias.c (memory_modified_in_insn_p): Return true for CALL_INSNs.
17065 2017-04-29 Marc Glisse <marc.glisse@inria.fr>
17067 PR tree-optimization/80487
17068 * tree-ssa-alias.c (stmt_kills_ref_p): Handle stpncpy and strncpy.
17070 2017-04-29 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
17072 PR tree-optimization/79697
17073 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Check if callee
17074 is BUILT_IN_STRDUP, BUILT_IN_STRNDUP, BUILT_IN_REALLOC.
17075 (propagate_necessity): Check if def_callee is BUILT_IN_STRDUP or
17077 * gimple-fold.c (gimple_fold_builtin_realloc): New function.
17078 (gimple_fold_builtin): Call gimple_fold_builtin_realloc.
17080 2017-04-28 Martin Sebor <msebor@redhat.com>
17082 PR tree-optimization/80523
17083 * gimple-ssa-sprintf.c (target_to_host_charmap): New global variable.
17084 (init_target_to_host_charmap, target_to_host, target_strtol10): New
17086 (maybe_warn, format_directive, parse_directive): Use new functions.
17087 (pass_sprintf_length::execute): Call init_target_to_host_charmap.
17089 2017-04-28 Marc Glisse <marc.glisse@inria.fr>
17091 * match.pd (X+Z OP Y+Z, X-Z OP Y-Z, Z-X OP Z-Y): New transformations.
17093 2017-04-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
17095 * configure.ac (SYSTEM_HEADER_DIR, BUILD_SYSTEM_HEADER_DIR,
17096 target_header_dir): Set correctly.
17097 * configure: Regenerated.
17098 * Makefile.in (BUILD_SYSTEM_HEADER_DIR): New make variabe.
17099 (LIMITS_H_TEST, if_multiarch, stmp-fixinc): Use BUILD_SYSTEM_HEADER_DIR
17100 instead of SYSTEM_HEADER_DIR.
17102 2017-04-28 Jan Hubicka <hubicka@ucw.cz>
17104 * ipa-cp.c (perform_estimation_of_a_value): Turn time to sreal.
17105 (estimate_local_effects): Likewise.
17106 * ipa-inline.c (compute_inlined_call_time, want_inline_small_function_p,
17107 edge_badness, inline_small_functions, dump_overall_stats): LIkewise.
17108 * ipa-inline.h (edge_growth_cache_entry, estimate_time_after_inlining,
17109 estimate_ipcp_clone_size_and_time, do_estimate_edge_time,
17110 do_estimate_edge_time, estimate_edge_time): Likewise.
17111 * ipa-inline-analysis.c (estimate_node_size_and_time,
17112 estimate_ipcp_clone_size_and_time, do_estimate_edge_time): Likewise.
17113 (estimate_time_after_inlining): Remove.
17115 2017-04-28 Martin Liska <mliska@suse.cz>
17117 * doc/gcov.texi: Enhance documentation of gcov.
17119 2017-04-28 Martin Liska <mliska@suse.cz>
17121 * doc/gcov.texi: Sort options in alphabetic order.
17122 * doc/gcov-dump.texi: Likewise.
17123 * doc/gcov-tool.texi: Likewise.
17124 * gcov.c (print_usage): Likewise.
17125 * gcov-dump.c (print_usage): Likewise.
17126 * gcov-tool.c (print_merge_usage_message): Likewise.
17127 (print_rewrite_usage_message): Likewise.
17128 (print_overlap_usage_message): Likewise.
17130 2017-04-28 Martin Liska <mliska@suse.cz>
17132 PR gcov-profile/53915
17133 * gcov.c (format_gcov): Print 'NAN %' when top > bottom.
17135 2017-04-28 Martin Liska <mliska@suse.cz>
17137 PR gcov-profile/79891
17138 * gcov.c (add_line_counts): Assign BBs to lines just if the BB
17139 is marked by compiler as living on a line.
17140 (get_cycles_count): Remove usage of the union.
17141 (output_intermediate_file): Likewise.
17142 (find_source): Fix GNU coding style.
17143 (accumulate_line_counts): Remove old non-all block mode.
17144 (output_lines): Remove usage of the union.
17145 * profile.c (output_location): Include all BBs, even if
17146 belonging to a same line (and file) as a previous BB.
17148 2017-04-28 Martin Liska <mliska@suse.cz>
17150 * gcov.c (process_args): Handle new argument 'w'.
17151 (read_graph_file): Assign ID to BBs.
17152 (output_branch_count): Display BB # if verbose flag is set.
17153 (output_lines): Likewise for arcs.
17154 (print_usage): Add '--verbose' option help.
17155 * doc/gcov.texi: Document --verbose (-w) option.
17157 2017-04-28 Martin Liska <mliska@suse.cz>
17159 * gcov.c (struct block_location_info): New struct.
17160 (process_file): Fill up the new structure.
17161 (read_graph_file): Replace usage of encoding by the newly added
17163 (add_line_counts): Likewise.
17164 (accumulate_line_counts): Remove usage of the union.
17165 (function_info::function_info): New function.
17166 (function_info::~function_info): Likewise.
17167 (process_file): Call delete instead of release_function.
17168 (release_function): Release the function.
17169 (release_structures): Call delete instead of release_function.
17170 (solve_flow_graph): Replace usage of num_blocks.
17171 (find_exception_blocks): Likewise.
17172 (output_lines): Fix GNU coding style.
17174 2017-04-28 Martin Liska <mliska@suse.cz>
17177 * coverage.c (coverage_remove_note_file): New function.
17178 * coverage.h: Declare the function.
17179 * toplev.c (finalize): Clean if an error has been seen.
17181 2017-04-28 Martin Liska <mliska@suse.cz>
17183 PR gcov-profile/80031
17184 * gcov-dump.c (tag_blocks): Just print number of basic blocks.
17185 * gcov-io.h (GCOV_TAG_BLOCKS_NUM): Remove unused macro.
17186 * gcov.c (read_graph_file): Read just number of blocks.
17187 * profile.c (branch_prob): Do not stream 0 flags per a basic
17190 2017-04-28 Martin Liska <mliska@suse.cz>
17192 * gcov-dump.c (tag_*): Add new argument to declarations.
17193 (dump_gcov_file): Likewise.
17194 (tag_blocks): Add and use new argument depth.
17195 (tag_arcs): Likewise.
17196 (tag_lines): Likewise.
17197 (tag_counters): Likewise.
17198 (tag_summary): Likewise.
17199 (dump_working_sets): Use depth to do a proper indentation.
17201 2017-04-28 Jakub Jelinek <jakub@redhat.com>
17204 * cgraph.h (symtab_node::debug_symtab): No longer inline.
17205 * symtab.c (symtab_node::debug_symtab): Move definition here.
17207 2017-04-28 Richard Biener <rguenther@suse.de>
17209 * lto-streamer.h (LTO_major_version): Bump to 7.
17211 2017-04-28 Richard Biener <rguenther@suse.de>
17213 * tree-vrp.c (assert_info): New struct.
17214 (add_assert_info): New helper.
17215 (register_edge_assert_for_2): Refactor to add asserts to a vector
17217 (register_edge_assert_for_1): Likewise.
17218 (register_edge_assert_for): Likewise.
17219 (finish_register_edge_assert_for): New helper actually registering
17220 asserts where live on edge.
17221 (find_conditional_asserts): Adjust.
17222 (find_switch_asserts): Likewise.
17223 (evrp_dom_walker::try_find_new_range): Generalize.
17224 (evrp_dom_walker::before_dom_children): Use register_edge_assert_for.
17226 2017-04-27 Marek Polacek <polacek@redhat.com>
17229 * fold-const.c (fold_binary_loc) <case EQ_EXPR, NE_EXPR>: Convert
17230 arg10 and arg11 to itype.
17232 2017-04-27 Jonathan Wakely <jwakely@redhat.com>
17234 * doc/extend.texi (Object Size Checking): Improve grammar.
17236 2017-04-27 Richard Earnshaw <rearnsha@arm.com>
17239 * config/aarch64/aarch64.c (aarch64_emit_approx_sqrt): Ensure
17240 that the logic for permitting reciprocal estimates matches that
17243 2017-04-27 Jakub Jelinek <jakub@redhat.com>
17246 * tree.c (type_cache_hasher::equal): Only compare
17247 TYPE_TYPELESS_STORAGE flag on non-aggregate element types.
17248 (build_array_type_1): Only hash TYPE_TYPELESS_STORAGE flag on
17249 non-aggregate element types.
17250 * tree.h (TYPE_TYPELESS_STORAGE): Fix comment typo, add more details
17251 about the flag on ARRAY_TYPEs in the comment, formatting fix.
17253 2017-04-27 Richard Biener <rguenther@suse.de>
17255 PR middle-end/80533
17256 * emit-rtl.c (set_mem_attributes_minus_bitpos): When
17257 stripping ARRAY_REFs from MEM_EXPR make sure we're not
17258 keeping a reference to a trailing array.
17260 2017-04-27 Richard Biener <rguenther@suse.de>
17262 PR middle-end/80539
17263 * tree-chrec.c (chrec_fold_plus_poly_poly): Deal with not
17264 being in loop-closed SSA form conservatively.
17265 (chrec_fold_multiply_poly_poly): Likewise.
17267 2017-04-27 Tamar Christina <tamar.christina@arm.com>
17269 PR middle-end/79665
17270 * expr.c (expand_expr_real_2): Move TRUNC_MOD_EXPR, FLOOR_MOD_EXPR,
17271 CEIL_MOD_EXPR, ROUND_MOD_EXPR cases.
17273 2017-04-27 Jakub Jelinek <jakub@redhat.com>
17276 * config/aarch64/aarch64.c (struct aarch64_fn_arg_alignment): Remove.
17277 (aarch64_function_arg_alignment): Return unsigned int again, but still
17278 ignore TYPE_FIELDS chain decls other than FIELD_DECLs.
17279 (aarch64_layout_arg): Adjust aarch64_function_arg_alignment caller.
17280 Don't emit -Wpsabi note.
17281 (aarch64_function_arg_boundary): Likewise.
17282 (aarch64_gimplify_va_arg_expr): Adjust aarch64_function_arg_alignment
17285 2017-04-26 Nathan Sidwell <nathan@acm.org>
17287 * tree.h (crc32_unsigned_n): Declare.
17288 (crc32_unsigned, crc32_unsigned): Make inline.
17289 * tree.c (crc32_unsigned_bits): Replace with ...
17290 (crc32_unsigned_n): ... this.
17291 (crc32_unsigned, crc32_byte): Remove.
17292 (crc32_string): Remove unnecessary braces.
17294 2017-04-25 Jan Hubicka <hubicka@ucw.cz>
17296 * ipa-cp.c (estimate_local_effects): Convert sreal to int.
17297 * ipa-inline-analysis.c (MAX_TIME): Remove.
17298 (account_size_time): Use sreal for time.
17299 (dump_inline_summary): Update.
17300 (estimate_function_body_sizes): Update.
17301 (estimate_edge_size_and_time): Update.
17302 (estimate_calls_size_and_time): Update.
17303 (estimate_node_size_and_time): Update.
17304 (inline_merge_summary): Update.
17305 (inline_update_overall_summary): Update.
17306 (estimate_time_after_inlining): Update.
17307 (inline_read_section): Update.
17308 (inline_write_summary): Update.
17309 * ipa-inline.c (compute_uninlined_call_time): Update.
17310 (compute_inlined_call_time): Update.
17311 (recursive_inlining): Update.
17312 (inline_small_functions): Update.
17313 (dump_overall_stats): Update.
17314 * ipa-inline.h: Include sreal.h.
17315 (size_time_entry): Turn time to sreal.
17316 (inline_summary): Turn self_time nad time to sreal.
17318 2017-04-25 Jan Hubicka <hubicka@ucw.cz>
17320 * sreal.c: Include backend.h, tree.h, gimple.h, cgraph.h and
17322 (sreal::stream_out, sreal::stream_in): New.
17323 * sreal.h (sreal::stream_out, sreal::stream_in): Declare.
17325 2017-04-25 Jakub Jelinek <jakub@redhat.com>
17327 * Makefile.in (s-options): Invoke opt-gather.awk with LC_ALL=C in the
17330 2017-04-25 Uros Bizjak <ubizjak@gmail.com>
17333 * config/i386/i386.c (dimode_scalar_to_vector_candidate_p):
17335 (dimode_scalar_chain::compute_convert_gain): Ditto.
17336 (dimode_scalar_chain::make_vector_copies): Ditto.
17337 (dimode_scalar_chain::convert_reg): Ditto.
17338 (dimode_scalar_chain::convert_insn): Ditto.
17339 * config/i386/sse.md (VI24_AVX512BW_1): Remove mode iterator.
17340 (VI248_AVX512BW_1): New mode iterator.
17341 (<mask_codefor>ashr<mode>3<mask_name>): Merge insn pattern with
17342 <mask_codefor>ashrv2di3<mask_name> insn using VI248_AVX512BW_1
17345 2017-04-25 Martin Sebor <msebor@redhat.com>
17347 PR tree-optimization/80497
17348 * gimple-ssa-sprintf.c (get_int_range): Avoid assuming all integer
17349 constants are representable in HOST_WIDE_INT.
17350 (parse_directive): Ditto.
17352 2017-04-25 Martin Sebor <msebor@redhat.com>
17355 * dominance.c (dom_info::m_n_basic_blocks): Change type to unsigned.
17356 (new_zero_array): Adjust signature.
17357 (dom_info::dom_init): Used unsigned rather that size_t.
17358 (dom_info::dom_info): Same.
17360 2017-04-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17361 Jakub Jelinek <jakub@redhat.com>
17364 * config/arm/arm.c: Include gimple.h.
17365 (aapcs_layout_arg): Emit -Wpsabi note if arm_needs_doubleword_align
17366 returns negative, increment ncrn only if it returned positive.
17367 (arm_needs_doubleword_align): Return int instead of bool,
17368 ignore DECL_ALIGN of non-FIELD_DECL TYPE_FIELDS chain
17369 members, but if there is any such non-FIELD_DECL
17370 > PARM_BOUNDARY aligned decl, return -1 instead of false.
17371 (arm_function_arg): Emit -Wpsabi note if arm_needs_doubleword_align
17372 returns negative, increment nregs only if it returned positive.
17373 (arm_setup_incoming_varargs): Likewise.
17374 (arm_function_arg_boundary): Emit -Wpsabi note if
17375 arm_needs_doubleword_align returns negative, return
17376 DOUBLEWORD_ALIGNMENT only if it returned positive.
17378 2017-04-25 Marek Polacek <polacek@redhat.com>
17381 * fold-const.c (fold_binary_loc) <case BIT_IOR_EXPR>: Convert arg0's
17382 first argument to type.
17384 2017-04-25 Bill Seurer <seurer@linux.vnet.ibm.com>
17387 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Change
17388 type checks to test for compatibility instead of equality.
17390 2017-04-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17391 Jakub Jelinek <jakub@redhat.com>
17394 * config/aarch64/aarch64.c (struct aarch64_fn_arg_alignment): New
17396 (aarch64_function_arg_alignment): Return aarch64_fn_arg_alignment
17397 struct. Ignore DECL_ALIGN of decls other than FIELD_DECL for
17398 the alignment computation, but return their maximum in warn_alignment.
17399 (aarch64_layout_arg): Adjust aarch64_function_arg_alignment caller.
17400 Emit a -Wpsabi note if warn_alignment is 16 bytes, but alignment
17402 (aarch64_function_arg_boundary): Likewise. Simplify using MIN/MAX.
17403 (aarch64_gimplify_va_arg_expr): Adjust aarch64_function_arg_alignment
17406 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17408 * config/arc/simdext.md (dmpyh): Fix typo.
17410 2017-04-25 Richard Biener <rguenther@suse.de>
17412 PR tree-optimization/80492
17413 * alias.c (compare_base_decls): Handle registers with asm
17414 specification conservatively.
17415 * tree-ssa-alias.c (decl_refs_may_alias_p): Handle
17416 compare_base_decls returning dont-know properly.
17418 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17420 * config/arc/arc.c (LEGITIMATE_OFFSET_ADDRESS_P): Delete macro.
17421 (legitimate_offset_address_p): New function.
17422 (arc_legitimate_address_p): Use above function.
17424 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17426 * config/arc/arc.c (arc_output_mi_thunk): Emit PIC calls.
17428 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17430 * config/arc/arc.c (arc_conditional_register_usage): Use ACCL,
17431 ACCH registers whenever they are available.
17433 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17435 * config/arc/arc.c (arc_conditional_register_usage): Make D0, D1
17436 double regs fix when not used.
17438 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17440 * config/arc/arc.h (REGNO_OK_FOR_BASE_P): Consider also extension
17442 (REG_OK_FOR_INDEX_P_NONSTRICT): Likewise.
17443 (REG_OK_FOR_BASE_P_NONSTRICT): Likewise.
17445 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17447 * config/arc/arc.c (arc_output_addsi): Check for h-register class
17448 when emitting short ADD instructions.
17450 2017-04-25 Claudiu Zissulescu <claziss@synopsys.com>
17452 * config/arc/arc.md (cmpsi_cc_insn_mixed): Use 'h' register
17454 (cmpsi_cc_c_insn): Likewise.
17455 (cbranchsi4_scratch): Compute proper instruction length using
17456 compact_hreg_operand.
17457 * config/arc/predicates.md (compact_hreg_operand): New predicate.
17459 2017-04-25 Richard Biener <rguenther@suse.de>
17461 PR middle-end/80509
17462 * passes.c (pass_manager::pass_manager): Initialize
17463 m_name_to_pass_map.
17465 2017-04-25 Richard Biener <rguenther@suse.de>
17467 PR tree-optimization/79201
17468 * tree-ssa-sink.c (statement_sink_location): Handle calls.
17470 2017-04-25 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
17473 * config/s390/vector.md: Split MEM->GPR vector moves for
17474 non-s_operand addresses.
17476 2017-04-25 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
17479 * config/s390/predicates.md (reload_const_wide_int_operand): New
17481 * config/s390/s390.md ("movti"): Remove d/P alternative.
17482 ("movti_bigconst"): New pattern definition.
17484 2017-04-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
17487 * s390-protos.h (s390_expand_cs_hqi): Removed.
17488 (s390_expand_cs, s390_expand_atomic_exchange_tdsi): New prototypes.
17489 * config/s390/s390.c (s390_emit_compare_and_swap): Handle all integer
17490 modes as well as CCZ1mode and CCZmode.
17491 (s390_expand_atomic_exchange_tdsi, s390_expand_atomic): Adapt to new
17492 signature of s390_emit_compare_and_swap.
17493 (s390_expand_cs_hqi): Likewise, make static.
17494 (s390_expand_cs_tdsi): Generate an explicit compare before trying
17495 compare-and-swap, in some cases.
17496 (s390_expand_cs): Wrapper function.
17497 (s390_expand_atomic_exchange_tdsi): New backend specific expander for
17499 (s390_match_ccmode_set): Allow CCZmode <-> CCZ1 mode.
17500 * config/s390/s390.md ("atomic_compare_and_swap<mode>"): Merge the
17501 patterns for small and large integers. Forbid symref memory operands.
17502 Move expander to s390.c. Require cc register.
17503 ("atomic_compare_and_swap<DGPR:mode><CCZZ1:mode>_internal")
17504 ("*atomic_compare_and_swap<TDI:mode><CCZZ1:mode>_1")
17505 ("*atomic_compare_and_swapdi<CCZZ1:mode>_2")
17506 ("*atomic_compare_and_swapsi<CCZZ1:mode>_3"): Use s_operand to forbid
17507 symref memory operands. Remove CC mode and call s390_match_ccmode
17509 ("atomic_exchange<mode>"): Allow and implement all integer modes.
17511 2017-04-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
17513 * config/s390/s390.md (define_peephole2): New peephole to help
17514 combining the load-and-test pattern with volatile memory.
17516 2017-04-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
17518 * config/s390/s390.md ("cstorecc4"): Use load-on-condition and deal
17519 with CCZmode for TARGET_Z196.
17521 2017-04-25 Jakub Jelinek <jakub@redhat.com>
17523 PR rtl-optimization/80501
17524 * combine.c (make_compound_operation_int): Set subreg_code to SET
17525 even for AND with mask of the sign bit of mode.
17527 PR rtl-optimization/80500
17528 * loop-unroll.c (combine_var_copies_in_loop_exit): Call copy_rtx on
17529 sum's initial value.
17531 2017-04-25 Julian Brown <julian@codesourcery.com>
17532 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
17534 * config/aarch64/thunderx2t99.md (thunderx2t99_crc): New Reservation.
17536 2017-04-25 Marc Glisse <marc.glisse@inria.fr>
17538 * fold-const.c (tree_single_nonzero_warnv_p): Handle SSA_NAME.
17540 2017-04-25 Julian Brown <julian@codesourcery.com>
17541 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
17543 * config/aarch64/thunderx2t99.md (thunderx2t99_aes): New Reservation.
17544 (thunderx2t99_sha): New Reservation.
17546 2017-04-25 Julian Brown <julian@codesourcery.com>
17547 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
17549 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set<mode>): Fix
17550 type for 1-element load.
17552 2017-04-24 Marc Glisse <marc.glisse@inria.fr>
17554 * match.pd (X/[ex]C CMP Y/[ex]C): New transformation.
17556 2017-04-24 Martin Jambor <mjambor@suse.cz>
17558 PR tree-optimization/80293
17559 * tree-sra.c (scalarizable_type_p): New parameter const_decl, make
17560 char arrays not totally scalarizable if it is false.
17561 (analyze_all_variable_accesses): Pass correct value in the new
17562 parameter. Add a statistics counter.
17564 2017-04-24 Jan Hubicka <hubicka@ucw.cz>
17566 PR middle-end/79931
17567 * ipa-devirt.c (dump_possible_polymorphic_call_targets): Fix ICE.
17569 2017-04-24 Richard Biener <rguenther@suse.de>
17571 PR tree-optimization/80494
17572 * tree-scalar-evolution.c (analyze_scalar_evolution_1): Bail
17573 out for complex types.
17575 2017-04-24 Richard Biener <rguenther@suse.de>
17577 * tree-ssa-sccvn.h (run_scc_vn): Adjust prototype.
17578 * tree-ssa-sccvn.c (print_scc): Print SCC size.
17579 (extract_and_process_scc_for_name): Never fail but drop SCC to varying.
17580 (DFS): Adjust and never fail.
17581 (sccvn_dom_walker::fail): Remove.
17582 (sccvn_dom_walker::before_dom_children): Adjust.
17583 (run_scc_vn): Likewise and never fail.
17584 * tree-ssa-pre.c (pass_pre::execute): Adjust.
17585 (pass_fre::execute): Likewise.
17587 2017-04-24 Richard Biener <rguenther@suse.de>
17589 PR tree-optimization/79725
17590 * tree-ssa-sink.c (statement_sink_location): Return whether
17591 failure reason was zero uses. Move that check later.
17592 (sink_code_in_bb): Deal with zero uses by removing the stmt
17595 2017-04-24 Richard Biener <rguenther@suse.de>
17598 * tree-ssa-uninit.c (warn_uninitialized_vars): Handle some
17599 pointer-based references.
17601 2017-04-24 Richard Biener <rguenther@suse.de>
17604 * pass_manager.h (pass_manager::operator new): Remove.
17605 (pass_manager::operator delete): Likewise.
17606 * passes.c (pass_manager::operator new): Remove.
17607 (pass_manager::operator delete): Likewise.
17608 (pass_manager::pass_manager): Zero individual pass members.
17610 2017-04-23 Uros Bizjak <ubizjak@gmail.com>
17613 * config/i386/i386.c (dimode_scalar_to_vector_candidate_p)
17614 <case ASHIFT, case LSHIFTRT>: Also consider variable shifts.
17615 Check "XEXP (src, 1)" operand here.
17616 <case PLUS, case MINUS, case IOR, case XOR, case AND>:
17617 Check "XEXP (src, 1)" operand here.
17618 (dimode_scalar_chain::make_vector_copies): Detect count register
17619 of a shift instruction. Zero extend count register from QImode
17620 to DImode to satisfy vector shift pattern count operand predicate.
17621 Substitute vector shift count operand with a DImode copy.
17622 (dimode_scalar_chain::convert_reg): Ditto, zero-extend from
17625 2017-04-21 Uros Bizjak <ubizjak@gmail.com>
17627 * config/i386/i386.md (*extzvqi_mem_rex64): Move above *extzv<mode>.
17628 Remove UNSPEC_NOREX_MEM tag. Update corresponding peephole2 pattern.
17629 (*insvqi_1_mem_rex64): Move above insv<mode>_1. Remove
17630 UNSPEC_NOREX_MEM tag. Update corresponding peephole2 pattern.
17631 (UNSPEC_NOREX_MEM): Remove definition.
17633 2017-04-21 Richard Biener <rguenther@suse.de>
17635 PR tree-optimization/79547
17636 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17637 Handle strlen, strcmp, strncmp, strcasecmp, strncasecmp, memcmp,
17638 bcmp, strspn, strcspn, __builtin_object_size and __builtin_constant_p
17639 without any constraints.
17641 2017-04-21 Richard Biener <rguenther@suse.de>
17643 PR tree-optimization/78847
17644 * fold-const.c (split_tree): Handle POINTER_PLUS_EXPR.
17646 2017-04-21 Richard Biener <rguenther@suse.de>
17648 * tree.h (build_qualified_type): Annotate with CXX_MEM_STAT_INFO.
17649 (build_distinct_type_copy): Likewise.
17650 (build_variant_type_copy): Likewise.
17651 * tree.c (build_qualified_type): Pass down mem-stat info.
17652 (build_distinct_type_copy): Likewise.
17653 (build_variant_type_copy): Likewise.
17655 2017-04-21 Richard Biener <rguenther@suse.de>
17657 PR tree-optimization/80237
17658 * tree-ssa-pre.c (find_leader_in_sets): Add third set argument,
17660 (phi_translate_1): Also allow a leader in AVAIL_OUT of pred
17661 for a simplified result.
17663 2016-04-21 Richard Biener <rguenther@suse.de>
17665 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Do not require
17666 sth as strict as a simple_iv but a chrec without symbols and an
17667 operand defined in the loop we are peeling (and not some subloop).
17668 (propagate_constants_for_unrolling): Propagate all constants.
17670 2017-04-20 Uros Bizjak <ubizjak@gmail.com>
17673 * config/i386/i386.c (print_reg): Remove assert for disalowed
17674 regno values, call output_operand_lossage instead.
17676 2017-04-20 Uros Bizjak <ubizjak@gmail.com>
17679 * config/i386/constraints.md (Yc): New register constraint.
17680 * config/i386/i386.md (*float<SWI48:mode><MODEF:mode>2_mixed):
17681 Use Yc constraint for alternative 2 of operand 0. Remove
17682 preferred_for_speed attribute.
17684 2017-04-20 Alexander Monakov <amonakov@ispras.ru>
17686 * omp-low.c (lower_lastprivate_clauses): Correct handling of linear and
17687 lastprivate clauses in SIMT case.
17689 2017-04-20 Volker Reichelt <v.reichelt@netcologne.de>
17691 * doc/invoke.texi (-Wextra-semi): Document new warning option.
17693 2017-04-20 Richard Biener <rguenther@suse.de>
17695 PR tree-optimization/57796
17696 * tree-vect-stmts.c (vect_model_store_cost): Cost scatters
17697 as N scalar stores.
17698 (vect_model_load_cost): Cost gathers as N scalar loads.
17700 2017-04-20 Richard Biener <rguenther@suse.de>
17702 * ggc-page.c (ggc_allocated_p): Rename to ...
17703 (safe_lookup_page_table_entry): ... this and return the lookup
17705 (gt_ggc_m_S): Use safe_lookup_page_table_entry.
17707 2017-04-20 Richard Biener <rguenther@suse.de>
17709 PR tree-optimization/80453
17710 * tree-ssa-sccvn.h (struct vn_phi_s): Add cclhs and ccrhs members.
17711 * tree-ssa-sccvn.c (cond_stmts_equal_p): Use recorded lhs and rhs
17712 from the conditions.
17713 (vn_phi_eq): Pass them down.
17714 (vn_phi_lookup): Record them.
17715 (vn_phi_insert): Likewise.
17717 2017-04-20 Matthew Fortune <matthew.fortune@imgtec.com>
17719 * config/mips/mips.c (mips_expand_vec_perm_const): Re-fix
17720 uninitialized variable warning to avoid buffer overrun.
17722 2017-04-20 Alexander Monakov <amonakov@ispras.ru>
17725 * doc/invoke.texi (-Wmissing-field-initializers): Mention that warning
17726 is suppressed for '{ 0 }' in C.
17728 2017-04-20 Jakub Jelinek <jakub@redhat.com>
17730 * BASE-VER: Set to 8.0.0.
17732 2017-04-20 Thomas Preud'homme <thomas.preudhomme@arm.com>
17734 * config/arm/arm.c (arm_elf_asm_cdtor): Create non-default
17735 priority .init_array and .fini_array section with SECTION_NOTYPE
17738 2017-04-20 Jakub Jelinek <jakub@redhat.com>
17740 PR middle-end/80423
17741 * tree.h (build_array_type): Add typeless_storage default argument.
17742 * tree.c (type_cache_hasher::equal): Also compare
17743 TYPE_TYPELESS_STORAGE flag for ARRAY_TYPEs.
17744 (build_array_type): Add typeless_storage argument, set
17745 TYPE_TYPELESS_STORAGE to it, if shared also hash it, and pass to
17747 (build_nonshared_array_type): Adjust build_array_type_1 caller.
17748 (build_array_type): Likewise. Add typeless_storage argument.
17750 2017-04-19 Eric Botcazou <ebotcazou@adacore.com>
17751 Jakub Jelinek <jakub@redhat.com>
17753 PR tree-optimization/80426
17754 * tree-vrp.c (extract_range_from_binary_expr_1): For an additive
17755 operation on symbolic operands, also compute the overflow for the
17756 invariant part when the operation degenerates into a negation.
17758 2017-04-19 Jakub Jelinek <jakub@redhat.com>
17761 * dwarf2out.c (modified_type_die, gen_type_die_with_usage):
17762 Check for t with zero TYPE_QUALS_NO_ADDR_SPACE.
17765 * tree-ssa-loop-manip.c (find_uses_to_rename_def): Ignore debug uses.
17767 2017-04-19 Georg-Johann Lay <avr@gjlay.de>
17770 * config/avr/avr.c (tree.h): Include it.
17771 (cgraph.h): Include it.
17772 (avr_encode_section_info): Don't warn for uninitialized progmem
17773 variable if it's just an alias.
17775 2017-04-19 Richard Biener <rguenther@suse.de>
17778 * auto-profile.c (afdo_vpt_for_early_inline): Update SSA
17779 when needed by AutoPGO.
17781 2017-04-19 Paulo J. Matos <paulo@matos-sorge.com>
17784 * doc/lto.texi: Remove an extra 'that'.
17786 2017-04-19 Segher Boessenkool <segher@kernel.crashing.org>
17788 PR rtl-optimization/80429
17789 * ira.c (split_live_ranges_for_shrink_wrap): Don't split regs that
17790 are only used in debug insns.
17792 2017-04-19 Eric Botcazou <ebotcazou@adacore.com>
17793 Vladimir Makarov <vmakarov@redhat.com>
17795 * config/sparc/predicates.md (input_operand): Add comment. Return
17796 true for any memory operand when LRA is in progress.
17797 * config/sparc/sparc.c (sparc_expand_move): Minor formatting fix.
17799 2017-04-18 Jeff Law <law@redhat.com>
17802 * mips.md ({return,simple_return}_internal): Do not overwrite
17805 2017-04-18 Jakub Jelinek <jakub@redhat.com>
17807 PR tree-optimization/80443
17808 * tree-vrp.c (intersect_ranges): For signed 1-bit precision type,
17809 instead of adding 1, subtract -1 and similarly instead of subtracting
17812 2017-04-18 Richard Sandiford <richard.sandiford@arm.com>
17814 PR rtl-optimization/80357
17815 * haifa-sched.c (tmp_bitmap): New variable.
17816 (model_recompute): Handle duplicate use records.
17817 (alloc_global_sched_pressure_data): Initialize tmp_bitmap.
17818 (free_global_sched_pressure_data): Free it.
17820 2017-04-18 Bernd Edlinger <bernd.edlinger@hotmail.de>
17823 2017-02-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
17824 * Makefile.in (BUILD_SYSTEM_HEADER_DIR): New make variabe.
17825 (LIMITS_H_TEST, if_multiarch, stmp-fixinc): Use BUILD_SYSTEM_HEADER_DIR
17826 instead of SYSTEM_HEADER_DIR.
17828 2017-04-18 Jeff Law <law@redhat.com>
17830 PR middle-end/80422
17831 * cfgcleanup.c (try_crossjump_to_edge): Verify SRC1 and SRC2 have
17832 predecessors after walking up the insn chain.
17834 2017-04-18 Jakub Jelinek <jakub@redhat.com>
17837 * dwarf2out.c (modified_type_die): Try harder not to emit internal
17838 sizetype type into debug info.
17840 2017-04-18 Michael Meissner <meissner@linux.vnet.ibm.com>
17843 * config/rs6000/rs6000.c (rs6000_expand_vector_extract): Eliminate
17844 unneeded test for TARGET_UPPER_REGS_SF.
17845 * config/rs6000/vsx.md (vsx_extract_v4sf_var): Likewise.
17847 2017-04-18 Jakub Jelinek <jakub@redhat.com>
17850 * sancov.c (sancov_pass): Use gsi_start_nondebug_after_labels_bb
17851 instead of gsi_after_labels.
17853 2017-04-18 Jeff Law <law@redhat.com>
17855 * regcprop.c (maybe_mode_change): Avoid creating copies of the
17859 2017-04-13 Jeff Law <law@redhat.com>
17860 * config/mips.mips.md (zero_extendsidi2): Do not allow SP to appear
17861 in operands[1] if it is a MEM and TARGET_MIPS16 is active.
17863 2017-04-18 Georg-Johann Lay <avr@gjlay.de>
17866 * config/avr/avr.c (intl.h): Include it.
17867 (avr_pgm_check_var_decl) [reason]: Wrap diagnostic snippets into _().
17869 2017-04-18 Martin Liska <mliska@suse.cz>
17871 PR gcov-profile/78783
17872 * gcov-tool.c (gcov_output_files): Validate that destination
17873 file is either removed by the tool or by a user.
17875 2017-04-14 Andrew Burgess <andrew.burgess@embecosm.com>
17876 Guy Benyei <guybe@mellanox.com>
17878 * config/arc/arc.c (arc_reorg): Move loop_end_id into a more local
17879 block, and do not negate it, the stored id is already negative.
17881 2017-04-14 Andrew Burgess <andrew.burgess@embecosm.com>
17883 * config/arc/arc.md (doloop_begin_i): Use @pcl assembler syntax.
17885 2017-04-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17888 * config/rs6000/rs6000-cpus.def (OTHER_P9_VECTOR_MASKS): Define
17889 masks of options that should be turned off if the VSX vector
17890 options are turned off.
17891 (OTHER_P8_VECTOR_MASKS): Likewise.
17892 (OTHER_VSX_VECTOR_MASKS): Likewise.
17893 * config/rs6000/rs6000.c (rs6000_option_override_internal): Call
17894 rs6000_disable_incompatible_switches to validate no type switches
17896 (rs6000_incompatible_switch): New function to disallow turning on
17897 other vector options if -mno-vsx, -mno-power8-vector, or
17898 -mno-power9-vector are specified.
17900 2017-04-14 Claudiu Zissulescu <claziss@synopsys.com>
17902 * config/arc/arc.h (CRT_CALL_STATIC_FUNCTION): Use long calls.
17904 2017-04-14 Claudiu Zissulescu <claziss@synopsys.com>
17906 * config/arc/arc-protos.h (arc_decl_pretend_args): Remove.
17907 * config/arc/arc.c (arc_decl_pretend_args): Likewise.
17908 * config/arc/arc.h (CFA_FRAME_BASE_OFFSET): Likewise.
17909 (ARG_POINTER_CFA_OFFSET): Likewise.
17911 2017-04-14 Claudiu Zissulescu <claziss@synopsys.com>
17913 * config/arc/arc.c (arc_mode_dependent_address_p): Relax
17914 conditions to take advantage of various optimizations.
17916 2017-04-13 Jeff Law <law@redhat.com>
17918 * config/mips.mips.md (zero_extendsidi2): Do not allow SP to appear
17919 in operands[1] if it is a MEM and TARGET_MIPS16 is active.
17920 (zero_extendsidi2_dext): Likewise.
17922 2017-04-13 Jakub Jelinek <jakub@redhat.com>
17925 * fold-const.c (fold_ternary_loc): Revert
17926 use op0 instead of fold_convert_loc (loc, type, arg0) part of
17929 2017-04-13 Vladimir Makarov <vmakarov@redhat.com>
17931 PR rtl-optimization/80343
17932 * lra-remat.c (update_scratch_ops): Assign original hard reg to
17933 new scratch pseudo.
17935 2017-04-13 Denis Khalikov <d.khalikov@partner.samsung.com>
17938 * ubsan.c (ubsan_expand_bounds_ifn): Pass original index
17939 to ubsan_encode_value.
17941 2017-04-13 Jeff Law <law@redhat.com>
17943 * reload1.c (eliminate_regs_1): Call gen_rtx_raw_SUBREG for SUBREGs
17944 appearing in DEBUG_INSNs.
17946 2017-04-13 Martin Liska <mliska@suse.cz>
17948 PR gcov-profile/80413
17949 * gcov-io.c (gcov_write_string): Copy to buffer just when
17950 allocated size is greater than zero.
17952 2017-04-13 Jakub Jelinek <jakub@redhat.com>
17955 * dwarf2out.c (decls_for_scope): Ignore declarations of
17956 current_function_decl in BLOCK_NONLOCALIZED_VARS.
17958 2017-04-12 Jan Hubicka <hubicka@ucw.cz>
17961 * ipa-visibility.c (non_local_p): Fix typos.
17962 (localize_node): When localizing symbol in same comdat group,
17963 dissolve the group only when we know external symbols are going
17965 (function_and_variable_visibility): Do not localize DECL_EXTERNAL.
17967 2017-04-12 Jakub Jelinek <jakub@redhat.com>
17969 PR tree-optimization/79390
17970 * optabs.c (emit_conditional_move): If the preferred op2/op3 operand
17971 order does not result in usable sequence, retry with reversed operand
17977 * fold-const.c (fold_ternary_loc): Use op1 instead of arg1 as argument
17978 to fold_build2_loc. Convert TREE_OPERAND (tem, 0) to type. Use
17979 op0 instead of fold_convert_loc (loc, type, arg0).
17981 2017-04-12 Jeff Law <law@redhat.com>
17983 * genattrtab.c (write_eligible_delay): Verify DELAY_INSN still
17984 has a delay slot in the generated code.
17986 * config/cris/cris.md (cris_preferred_reload_class): Return
17987 GENNONACR_REGS rather than GENERAL_REGS.
17989 2017-04-12 Jakub Jelinek <jakub@redhat.com>
17992 * expr.c <CASE_CONVERT>: For EXPAND_INITIALIZER determine SIGN_EXTEND
17993 vs. ZERO_EXTEND based on signedness of treeop0's type rather than
17994 signedness of the result type.
17996 2017-04-12 Richard Biener <rguenther@suse.de>
17997 Jeff Law <law@redhat.com>
17999 PR tree-optimization/80359
18000 * tree-ssa-dse.c (maybe_trim_partially_dead_store): Do not
18001 trim stores to TARGET_MEM_REFs.
18003 2017-04-12 Richard Biener <rguenther@suse.de>
18005 PR tree-optimization/79390
18006 * gimple-ssa-split-paths.c (is_feasible_trace): Restrict
18007 threading case even more.
18009 2017-04-12 Segher Boessenkool <segher@kernel.crashing.org>
18012 * config/rs6000/sync.md (atomic_load<mode>, atomic_store<mode): Test
18013 for quad_address_p for TImode, instead of just not indexed_address.
18015 2017-04-12 Richard Biener <rguenther@suse.de>
18016 Bernd Edlinger <bernd.edlinger@hotmail.de>
18018 PR middle-end/79671
18019 * alias.c (component_uses_parent_alias_set_from): Handle
18020 TYPE_TYPELESS_STORAGE.
18021 (get_alias_set): Likewise.
18022 * tree-core.h (tree_type_common): Add typeless_storage flag.
18023 * tree.h (TYPE_TYPELESS_STORAGE): New macro.
18024 * stor-layout.c (place_union_field): Set TYPE_TYPELESS_STORAGE
18025 for types containing members with TYPE_TYPELESS_STORAGE.
18026 (place_field): Likewise.
18027 (layout_type): Likewise for ARRAY_TYPE.
18028 * lto-streamer-out.c (hash_tree): Hash TYPE_TYPELESS_STORAGE.
18029 * tree-streamer-in.c (unpack_ts_type_common_value_fields): Stream
18030 TYPE_TYPELESS_STORAGE.
18031 * tree-streamer-out.c (pack_ts_type_common_value_fields): Likewise.
18033 2017-04-12 Jakub Jelinek <jakub@redhat.com>
18036 * fold-const.c (fold_binary_loc) <case BIT_IOR_EXPR>: Convert arg0's
18037 first argument to type.
18039 2017-04-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18043 * config/rs6000/rs6000.c (rs6000_expand_unop_builtin): Return
18044 CONST0_RTX (mode) rather than const0_rtx where appropriate.
18045 (rs6000_expand_binop_builtin): Likewise.
18046 (rs6000_expand_ternop_builtin): Likewise; also add missing
18047 vsx_xxpermdi_* variants; also fix typo (arg1 => arg2) for
18048 vshasigma built-ins.
18049 * doc/extend.texi: Document that vec_xxpermdi's third argument
18050 must be a constant.
18052 2017-04-11 Uros Bizjak <ubizjak@gmail.com>
18054 * config/i386/i386.c (dimode_scalar_chain::compute_convert_gain):
18055 Use shift_const cost parameter when calculating gain of STV shifts.
18057 2017-04-11 Vladimir Makarov <vmakarov@redhat.com>
18059 PR rtl-optimization/70478
18060 * lra-constraints.c (process_alt_operands): Check memory for
18061 disfavoring memory insn operand.
18063 2017-04-11 Jakub Jelinek <jakub@redhat.com>
18065 PR middle-end/80100
18066 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Perform
18067 left shift in unsigned HOST_WIDE_INT type.
18069 PR rtl-optimization/80385
18070 * simplify-rtx.c (simplify_unary_operation_1): Don't transform
18071 (not (neg X)) into (plus X -1) for complex or non-integral modes.
18074 * omp-low.c (scan_omp_task): Don't optimize away empty tasks
18075 if they have any depend clauses.
18077 2017-04-11 Martin Liska <mliska@suse.cz>
18080 * cgraph.c (cgraph_node::dump): Dump calls_comdat_local.
18081 * ipa-split.c (split_function): Create a local comdat symbol
18082 if caller is in a comdat group.
18084 2017-04-11 Martin Liska <mliska@suse.cz>
18087 * ipa-cp.c (determine_versionability): Handle calls_comdat_local
18090 2017-04-11 Martin Sebor <msebor@redhat.com>
18092 PR middle-end/80364
18093 * gimple-ssa-sprintf.c (get_int_range): Remove second argument and
18094 always use the int type. Use INTEGRAL_TYPE_P() rather than testing
18096 (directive::set_width, directive::set_precision, format_character):
18098 (parse_directive): Use INTEGRAL_TYPE_P() rather than testing for
18101 2017-04-11 Richard Earnshaw <rearnsha@arm.com>
18104 * config/arm/arm.c (arm_configure_build_target): When -mcpu and -arch
18105 conflict, set target->arch_name instead of target->cpu_name.
18107 2017-04-11 Richard Biener <rguenther@suse.de>
18109 PR tree-optimization/80374
18110 * tree-ssa-dom.c (derive_equivalences_from_bit_ior): Use
18111 build_zero_cst, remove fold_convertible_p check again.
18113 2017-04-11 Martin Liska <mliska@suse.cz>
18116 * ubsan.c (instrument_object_size): Do not instrument register
18119 2017-04-11 Jakub Jelinek <jakub@redhat.com>
18122 * config/i386/i386-builtin-types.def
18123 (V16HI_FTYPE_V16HI_INT_V16HI_UHI_COUNT,
18124 V16HI_FTYPE_V16HI_V8HI_V16HI_UHI_COUNT,
18125 V16SI_FTYPE_V16SI_INT_V16SI_UHI_COUNT,
18126 V16SI_FTYPE_V16SI_V4SI_V16SI_UHI_COUNT,
18127 V2DI_FTYPE_V2DI_INT_V2DI_UQI_COUNT,
18128 V2DI_FTYPE_V2DI_V2DI_V2DI_UQI_COUNT,
18129 V32HI_FTYPE_V32HI_INT_V32HI_USI_COUNT,
18130 V32HI_FTYPE_V32HI_V8HI_V32HI_USI_COUNT,
18131 V4DI_FTYPE_V4DI_INT_V4DI_UQI_COUNT,
18132 V4DI_FTYPE_V4DI_V2DI_V4DI_UQI_COUNT,
18133 V4SI_FTYPE_V4SI_INT_V4SI_UQI_COUNT,
18134 V4SI_FTYPE_V4SI_V4SI_V4SI_UQI_COUNT,
18135 V8DI_FTYPE_V8DI_INT_V8DI_UQI_COUNT,
18136 V8DI_FTYPE_V8DI_V2DI_V8DI_UQI_COUNT,
18137 V8HI_FTYPE_V8HI_INT_V8HI_UQI_COUNT,
18138 V8HI_FTYPE_V8HI_V8HI_V8HI_UQI_COUNT,
18139 V8SI_FTYPE_V8SI_INT_V8SI_UQI_COUNT,
18140 V8SI_FTYPE_V8SI_V4SI_V8SI_UQI_COUNT): New function type aliases.
18141 * config/i386/i386-builtin.def (__builtin_ia32_pslld512_mask,
18142 __builtin_ia32_pslldi512_mask, __builtin_ia32_psllq512_mask,
18143 __builtin_ia32_psllqi512_mask, __builtin_ia32_psrad512_mask,
18144 __builtin_ia32_psradi512_mask, __builtin_ia32_psraq512_mask,
18145 __builtin_ia32_psraqi512_mask, __builtin_ia32_psrld512_mask,
18146 __builtin_ia32_psrldi512_mask, __builtin_ia32_psrlq512_mask,
18147 __builtin_ia32_psrlqi512_mask, __builtin_ia32_psllwi128_mask,
18148 __builtin_ia32_pslldi128_mask, __builtin_ia32_psllqi128_mask,
18149 __builtin_ia32_psllw128_mask, __builtin_ia32_pslld128_mask,
18150 __builtin_ia32_psllq128_mask, __builtin_ia32_psllwi256_mask,
18151 __builtin_ia32_psllw256_mask, __builtin_ia32_pslldi256_mask,
18152 __builtin_ia32_pslld256_mask, __builtin_ia32_psllqi256_mask,
18153 __builtin_ia32_psllq256_mask, __builtin_ia32_psradi128_mask,
18154 __builtin_ia32_psrad128_mask, __builtin_ia32_psradi256_mask,
18155 __builtin_ia32_psrad256_mask, __builtin_ia32_psraqi128_mask,
18156 __builtin_ia32_psraq128_mask, __builtin_ia32_psraqi256_mask,
18157 __builtin_ia32_psraq256_mask, __builtin_ia32_psrldi128_mask,
18158 __builtin_ia32_psrld128_mask, __builtin_ia32_psrldi256_mask,
18159 __builtin_ia32_psrld256_mask, __builtin_ia32_psrlqi128_mask,
18160 __builtin_ia32_psrlq128_mask, __builtin_ia32_psrlqi256_mask,
18161 __builtin_ia32_psrlq256_mask, __builtin_ia32_psrawi256_mask,
18162 __builtin_ia32_psraw256_mask, __builtin_ia32_psrawi128_mask,
18163 __builtin_ia32_psraw128_mask, __builtin_ia32_psrlwi256_mask,
18164 __builtin_ia32_psrlw256_mask, __builtin_ia32_psrlwi128_mask,
18165 __builtin_ia32_psrlw128_mask, __builtin_ia32_psllwi512_mask,
18166 __builtin_ia32_psllw512_mask, __builtin_ia32_psrawi512_mask,
18167 __builtin_ia32_psraw512_mask, __builtin_ia32_psrlwi512_mask,
18168 __builtin_ia32_psrlw512_mask): Use _COUNT suffixed function type
18170 * config/i386/i386.c (ix86_expand_args_builtin): Rename last_arg_count
18171 flag to second_arg_count, handle 4 argument function type _COUNT
18172 aliases, handle second_arg_count on second argument rather than last.
18174 2017-04-10 Jeff Law <law@redhat.com>
18176 PR tree-optimization/80374
18177 * tree-ssa-dom.c (derive_equivalences_from_bit_ior): Do not try to
18178 record anything if we can not convert integer_zero_node to the
18181 2017-04-10 Kelvin Nilsen <kelvin@gcc.gnu.org>
18184 * config/rs6000/rs6000.c (rs6000_option_override_internal):
18185 Enhance special handling given to the TARGET_P9_MINMAX option in
18186 relation to certain other options.
18188 2017-04-10 Bin Cheng <bin.cheng@arm.com>
18190 PR tree-optimization/80153
18191 * tree-ssa-loop-ivopts.c (add_iv_candidate_for_use): Check and
18192 remove POINTER_PLUS_EXPR's base part directly, rather than through
18195 2017-04-10 Richard Biener <rguenther@suse.de>
18196 Bin Cheng <bin.cheng@arm.com>
18198 PR tree-optimization/80153
18199 * tree-affine.c (aff_combination_to_tree): Get base pointer from
18200 the first element of pointer type aff_tree. Build result expr in
18202 (add_elt_to_tree): Convert to type unconditionally. Remove other
18203 fold_convert calls.
18204 * tree-ssa-loop-ivopts.c (alloc_iv): Pass in consistent types.
18205 (rewrite_use_nonlinear_expr): Check invariant using iv information.
18207 2017-04-10 Richard Biener <rguenther@suse.de>
18209 * tree-ssa-structalias.c (find_func_aliases): Properly handle
18212 2017-04-10 Vladimir Makarov <vmakarov@redhat.com>
18214 PR rtl-optimization/70478
18215 * lra-constraints.c (curr_small_class_check): New.
18216 (update_and_check_small_class_inputs): New.
18217 (process_alt_operands): Update curr_small_class_check. Disfavor
18218 alternative insn memory operands. Check available regs for small
18221 2017-03-31 Matthew Fortune <matthew.fortune@imgtec.com>
18224 * config/mips/mips.opt (-mvirt): Update description.
18225 * doc/invoke.texi (-mvirt): Likewise.
18227 2017-04-10 Richard Biener <rguenther@suse.de>
18229 PR middle-end/80362
18230 * fold-const.c (fold_binary_loc): Look at unstripped ops when
18231 looking for NEGATE_EXPR in -A / -B to A / B folding.
18233 2017-04-10 Martin Liska <mliska@suse.cz>
18235 PR gcov-profile/80224
18236 * gcov.c (print_usage): Fix usage string.
18237 (get_gcov_intermediate_filename): Remove.
18238 (output_gcov_file): Use both for normal and intermediate format.
18239 (generate_results): Do not initialize special file for
18240 intermediate format.
18242 2017-04-10 Richard Biener <rguenther@suse.de>
18244 PR tree-optimization/80304
18245 * tree-ssa-loop-im.c (ref_indep_loop_p_1): Also recurse
18248 2017-04-10 Nathan Sidwell <nathan@acm.org>
18251 * config/rs6000/rs6000.c (rs6000_vector_type): New.
18252 (rs6000_init_builtins): Use it.
18254 2016-04-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18256 * config/arm/arm.md (<mrc>): Add mode to SET source.
18257 (<mrrc>): Likewise.
18259 2017-04-10 Richard Biener <rguenther@suse.de>
18261 PR middle-end/80344
18262 * gimplify.c (is_gimple_mem_rhs_or_call): Allow CLOBBERs.
18264 2017-04-10 Jakub Jelinek <jakub@redhat.com>
18267 * config/i386/avx512fintrin.h (_mm512_reduce_add_epi32,
18268 _mm512_reduce_mul_epi32, _mm512_reduce_and_epi32,
18269 _mm512_reduce_or_epi32, _mm512_mask_reduce_add_epi32,
18270 _mm512_mask_reduce_mul_epi32, _mm512_mask_reduce_and_epi32,
18271 _mm512_mask_reduce_or_epi32, _mm512_reduce_min_epi32,
18272 _mm512_reduce_max_epi32, _mm512_reduce_min_epu32,
18273 _mm512_reduce_max_epu32, _mm512_mask_reduce_min_epi32,
18274 _mm512_mask_reduce_max_epi32, _mm512_mask_reduce_min_epu32,
18275 _mm512_mask_reduce_max_epu32, _mm512_reduce_add_ps,
18276 _mm512_reduce_mul_ps, _mm512_mask_reduce_add_ps,
18277 _mm512_mask_reduce_mul_ps, _mm512_reduce_min_ps, _mm512_reduce_max_ps,
18278 _mm512_mask_reduce_min_ps, _mm512_mask_reduce_max_ps,
18279 _mm512_reduce_add_epi64, _mm512_reduce_mul_epi64,
18280 _mm512_reduce_and_epi64, _mm512_reduce_or_epi64,
18281 _mm512_mask_reduce_add_epi64, _mm512_mask_reduce_mul_epi64,
18282 _mm512_mask_reduce_and_epi64, _mm512_mask_reduce_or_epi64,
18283 _mm512_reduce_min_epi64, _mm512_reduce_max_epi64,
18284 _mm512_mask_reduce_min_epi64, _mm512_mask_reduce_max_epi64,
18285 _mm512_reduce_min_epu64, _mm512_reduce_max_epu64,
18286 _mm512_mask_reduce_min_epu64, _mm512_mask_reduce_max_epu64,
18287 _mm512_reduce_add_pd, _mm512_reduce_mul_pd, _mm512_mask_reduce_add_pd,
18288 _mm512_mask_reduce_mul_pd, _mm512_reduce_min_pd, _mm512_reduce_max_pd,
18289 _mm512_mask_reduce_min_pd, _mm512_mask_reduce_max_pd): New intrinsics.
18291 2017-04-08 Vladimir Makarov <vmakarov@redhat.com>
18293 PR rtl-optimization/70478
18294 * lra-constraints.c: Reverse the last patch.
18296 2017-04-08 Andreas Tobler <andreast@gcc.gnu.org>
18298 * config/aarch64/aarch64-freebsd.h: Define MCOUNT_NAME.
18299 Add comment for WCHAR_T.
18301 2017-04-08 Martin Liska <mliska@suse.cz>
18304 2017-04-07 Martin Liska <mliska@suse.cz>
18307 * ipa-split.c (split_function): Add function part to a same comdat
18310 2017-04-08 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
18313 * config/rs6000/rs6000.c (expand_block_compare): Fix boundary check.
18315 2017-04-07 Pat Haugen <pthaugen@us.ibm.com>
18317 * rs6000/rs6000.c (vec_load_pendulum): Rename...
18318 (vec_pairing): ...to this.
18319 (power9_sched_reorder2): Rewrite code for pairing vector/vecload insns.
18320 (rs6000_sched_init): Adjust for name change.
18321 (struct rs6000_sched_context): Likewise.
18322 (rs6000_init_sched_context): Likewise.
18323 (rs6000_set_sched_context): Likewise.
18325 2017-04-07 Jakub Jelinek <jakub@redhat.com>
18331 * config/i386/avxintrin.h (_mm256_cvtsd_f64, _mm256_cvtss_f32): New
18333 * config/i386/avx512fintrin.h (_mm512_int2mask, _mm512_mask2int,
18334 _mm512_abs_ps, _mm512_mask_abs_ps, _mm512_abs_pd, _mm512_mask_abs_pd,
18335 _mm512_cvtsd_f64, _mm512_cvtss_f32): Likewise.
18337 2017-04-07 Andreas Tobler <andreast@gcc.gnu.org>
18339 * config/aarch64/aarch64-freebsd.h: Define WCHAR_TYPE.
18341 2017-04-07 Vladimir Makarov <vmakarov@redhat.com>
18343 PR rtl-optimization/70703
18344 * ira-color.c (update_conflict_hard_regno_costs): Use
18345 int64_t instead of HOST_WIDE_INT.
18347 2017-04-07 Vladimir Makarov <vmakarov@redhat.com>
18349 PR rtl-optimization/70478
18350 * lra-constraints.c (process_alt_operands): Disfavor alternative
18351 insn memory operands.
18353 2017-04-07 Jeff Law <law@redhat.com>
18355 * config/iq2000/iq2000.c (final_prescan_insn): Do not separate a
18356 CALL and NOTE_INSN_CALL_ARG_LOCATION.
18358 2017-04-07 Martin Liska <mliska@suse.cz>
18361 * config/aarch64/aarch64.c (aarch64_process_target_attr):
18362 Show error message instead of an ICE.
18364 2017-04-07 Martin Liska <mliska@suse.cz>
18367 * ipa-split.c (split_function): Add function part to a same comdat
18370 2017-04-07 Richard Biener <rguenther@suse.de>
18372 PR middle-end/80341
18373 * tree.c (get_unwidened): Also handle ! for_type case for
18375 * convert.c (do_narrow): Split out from ...
18376 (convert_to_integer_1): ... here. Do not pass final truncation
18377 type to get_unwidened for TRUNC_DIV_EXPR.
18379 2017-04-07 Richard Biener <rguenther@suse.de>
18381 * tree-affine.c (wide_int_ext_for_comb): Take type rather
18383 (aff_combination_const): Adjust.
18384 (aff_combination_scale): Likewise.
18385 (aff_combination_add_elt): Likewise.
18386 (aff_combination_add_cst): Likewise.
18387 (aff_combination_convert): Likewise.
18388 (add_elt_to_tree): Likewise. Remove unused argument.
18389 (aff_combination_to_tree): Adjust calls to add_elt_to_tree.
18391 2017-04-07 Sebastian Huber <sebastian.huber@embedded-brains.de>
18393 * config/arm/arm.h (ARM_DEFAULT_SHORT_ENUMS): Provide default
18395 * config/arm/arm.c (arm_default_short_enums): Use
18396 ARM_DEFAULT_SHORT_ENUMS.
18397 * config/arm/rtems.h (ARM_DEFAULT_SHORT_ENUMS): Define.
18399 2017-04-06 Jakub Jelinek <jakub@redhat.com>
18402 * dwarf2out.c (gen_member_die): Handle C++17 inline static data
18403 members with redundant out-of-class redeclaration.
18405 2017-04-06 Uros Bizjak <ubizjak@gmail.com>
18408 * config/i386/sse.md (*vec_extractv4si_0_zext_sse4): New pattern.
18409 * config/i386/i386.md (*zero_extendsidi2):
18410 Add (?*x,*x) and (?*v,*v) alternatives.
18412 2017-04-06 Uros Bizjak <ubizjak@gmail.com>
18415 * config/i386/i386.c (ix86_expand_builtin)
18416 <case IX86_BUILTIN_K{,OR}TEST{C,Z}{8,16,32,64}>: Determine insn operand
18417 mode from insn data. Convert operands to insn operand mode.
18418 Copy operands that don't satisfy insn predicate to a register.
18420 2017-04-06 Sam Thursfield <sam.thursfield@codethink.co.uk>
18422 * config/rs6000/x-aix: Increase memory limit for genautomata on AIX.
18425 2017-04-06 Richard Biener <rguenther@suse.de>
18427 PR tree-optimization/80334
18428 * tree-ssa-loop-ivopts.c (rewrite_use_address): Properly
18429 preserve alignment of accesses.
18431 2017-04-06 Richard Biener <rguenther@suse.de>
18433 PR tree-optimization/80262
18434 * tree-sra.c (build_ref_for_offset): Preserve address-space
18436 * tree-ssa-sccvn.c (vn_reference_maybe_forwprop_address):
18437 Drop useless address-space information on MEM_REF offsets.
18439 2017-04-05 Andreas Schwab <schwab@linux-m68k.org>
18441 * builtins.def (BUILT_IN_UPDATE_SETJMP_BUF): Fix type.
18443 2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
18445 PR rtl-optimization/70703
18446 * ira-color.c (update_conflict_hard_regno_costs): Use
18447 HOST_WIDE_INT instead of long.
18449 2017-04-05 Uros Bizjak <ubizjak@gmail.com>
18452 * config/i386/mmintrin.h: Add -msse target option when __SSE__ is
18453 not defined for x86_64 target. Add -mmmx target option when __SSE2__
18455 * config/i386/mm3dnow.h: Add -msse target when __SSE__ is not defined
18456 for x86_64 target. Handle -m3dnowa option.
18458 2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
18460 PR rtl-optimization/70703
18461 * ira-color.c (update_costs_from_allocno): Use the smallest mode.
18462 (update_conflict_hard_regno_costs): Use long instead of unsigned
18463 arithmetic for cost calculation.
18465 2017-04-05 Jakub Jelinek <jakub@redhat.com>
18466 Bernd Edlinger <bernd.edlinger@hotmail.de>
18469 * asan.c (asan_store_shadow_bytes): Fix location of last_chunk_value
18472 2017-04-05 Eric Botcazou <ebotcazou@adacore.com>
18475 * config/aarch64/aarch64.c (aarch64_emit_probe_stack_range): Replace
18476 ptr_mode with Pmode throughout.
18477 * config/aarch64/aarch64.md (probe_stack_range_<PTR:mode): Rename
18478 into probe_stack_range and use DImode.
18480 2017-04-05 Dominik Vogt <vogt@linux.vnet.ibm.com>
18483 * config/s390/s390.c (s390_register_info_gprtofpr): Return if
18484 call_eh_return is true.
18486 2017-04-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
18488 * config/s390/s390-c.c (s390_resolve_overloaded_builtin):
18489 Initialize last_match_fntype_index.
18491 2017-04-05 Jakub Jelinek <jakub@redhat.com>
18494 * tree-nvr.c: Include internal-fn.h.
18495 (pass_return_slot::execute): Ignore internal calls without
18498 2017-04-04 Jakub Jelinek <jakub@redhat.com>
18499 Richard Biener <rguenther@suse.de>
18502 * genmatch.c (capture::gen_transform): For GENERIC unshare_expr
18503 captures used multiple times, except for the last use.
18504 * generic-match-head.c: Include gimplify.h.
18506 2017-04-04 Jakub Jelinek <jakub@redhat.com>
18508 PR tree-optimization/79390
18509 * target.h (struct noce_if_info): Declare.
18510 * targhooks.h (default_noce_conversion_profitable_p): Declare.
18511 * target.def (noce_conversion_profitable_p): New target hook.
18512 * ifcvt.h (struct noce_if_info): New type, moved from ...
18513 * ifcvt.c (struct noce_if_info): ... here.
18514 (noce_conversion_profitable_p): Renamed to ...
18515 (default_noce_conversion_profitable_p): ... this. No longer
18517 (noce_try_store_flag_constants, noce_try_addcc,
18518 noce_try_store_flag_mask, noce_try_cmove, noce_try_cmove_arith,
18519 noce_convert_multiple_sets): Use targetm.noce_conversion_profitable_p
18520 instead of noce_conversion_profitable_p.
18521 * config/i386/i386.c: Include ifcvt.h.
18522 (ix86_option_override_internal): Don't override
18523 PARAM_MAX_RTL_IF_CONVERSION_INSNS default.
18524 (ix86_noce_conversion_profitable_p): New function.
18525 (TARGET_NOCE_CONVERSION_PROFITABLE_P): Redefine.
18526 * config/i386/x86-tune.def (X86_TUNE_ONE_IF_CONV_INSN): Adjust comment.
18527 * doc/tm.texi.in (TARGET_NOCE_CONVERSION_PROFITABLE_P): Add.
18528 * doc/tm.texi: Regenerated.
18530 2017-04-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18532 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Grammar
18535 2017-04-04 Thomas Preud'homme <thomas.preudhomme@arm.com>
18538 * config/arm/arm.c (thumb1_rtx_costs): Give a cost of 32
18539 instructions for small multiply cores.
18541 2017-04-04 Jeff Law <law@redhat.com>
18543 * config/mips/mips.c (mips_multi_add): Zero initialize the newly
18545 (mips_expand_vec_perm_const): Initialize elements in orig_perm
18546 that are not set by the loop over the elements.
18548 2017-04-04 Jakub Jelinek <jakub@redhat.com>
18551 * config/i386/i386.c (ix86_expand_args_builtin): If op has scalar
18552 int mode, convert_modes it to mode as unsigned, otherwise use
18553 lowpart_subreg to mode rather than SImode.
18554 * config/i386/sse.md (<mask_codefor>ashr<mode>3<mask_name>,
18555 ashr<mode>3, ashr<mode>3<mask_name>, <shift_insn><mode>3<mask_name>):
18556 Use DImode instead of SImode for the shift count operand.
18557 * config/i386/mmx.md (mmx_ashr<mode>3, mmx_<shift_insn><mode>3):
18560 2017-04-04 Richard Biener <rguenther@suse.de>
18562 PR middle-end/80281
18563 * match.pd (A + (-B) -> A - B): Make sure to preserve unsigned
18564 arithmetic done for the negate or the plus. Simplify.
18565 (A - (-B) -> A + B): Likewise.
18566 * fold-const.c (split_tree): Make sure to not negate pointers.
18568 2017-04-04 Segher Boessenkool <segher@kernel.crashing.org>
18570 PR rtl-optimization/60818
18571 * simplify-rtx.c (simplify_binary_operation_1): Do not replace
18572 a compare of comparisons with the thing compared if this results
18573 in a different machine mode.
18575 2017-04-03 Jonathan Wakely <jwakely@redhat.com>
18577 * alias.c (base_alias_check): Fix typo in comment.
18578 * cgraph.h (class ipa_polymorphic_call_context): Likewise.
18579 * cgraphunit.c (symbol_table::compile): Likewise.
18580 * collect2.c (maybe_run_lto_and_relink): Likewise.
18581 * config/arm/arm.c (arm_thumb1_mi_thunk): Likewise.
18582 * config/avr/avr-arch.h (avr_arch_info_t): Likewise.
18583 * config/avr/avr.c (avr_map_op_t): Likewise.
18584 * config/cr16/cr16.h (DATA_ALIGNMENT): Likewise.
18585 * config/epiphany/epiphany.c (TARGET_ARG_PARTIAL_BYTES): Likewise.
18586 * config/epiphany/epiphany.md (movcc): Likewise.
18587 * config/i386/i386.c (legitimize_pe_coff_extern_decl): Likewise.
18588 * config/m68k/m68k.c (struct _sched_ib, m68k_sched_variable_issue):
18590 * config/mips/mips.c (mips_save_restore_reg): Likewise.
18591 * config/rx/rx.c (rx_is_restricted_memory_address): Likewise.
18592 * config/s390/s390.c (Z10_EARLYLOAD_DISTANCE): Likewise.
18593 * config/sh/sh.c (sh_rtx_costs): Likewise.
18594 * fold-const.c (fold_truth_andor): Likewise.
18595 * genautomata.c (collapse_flag): Likewise.
18596 * gengtype.h (struct type::u::s): Likewise.
18597 * gensupport.c (has_subst_attribute, add_mnemonic_string): Likewise.
18598 * input.c (FORMAT_AMOUNT): Likewise.
18599 * ipa-cp.c (class ipcp_lattice, agg_replacements_to_vector)
18600 (known_aggs_to_agg_replacement_list): Likewise.
18601 * ipa-inline-analysis.c: Likewise.
18602 * ipa-inline.h (estimate_edge_time, estimate_edge_hints): Likewise.
18603 * ipa-polymorphic-call.c
18604 (ipa_polymorphic_call_context::restrict_to_inner_class): Likewise.
18605 * loop-unroll.c (analyze_insn_to_expand_var): Likewise.
18606 * lra.c (lra_optional_reload_pseudos, lra_subreg_reload_pseudos):
18608 * modulo-sched.c (apply_reg_moves): Likewise.
18609 * omp-expand.c (build_omp_regions_1): Likewise.
18610 * trans-mem.c (struct tm_wrapper_hasher): Likewise.
18611 * tree-ssa-loop-ivopts.c (may_eliminate_iv): Likewise.
18612 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Likewise.
18613 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Likewise.
18614 * value-prof.c: Likewise.
18615 * var-tracking.c (val_reset): Likewise.
18617 2017-04-03 Richard Biener <rguenther@suse.de>
18619 PR tree-optimization/80275
18620 * fold-const.c (split_address_to_core_and_offset): Handle
18623 2017-04-03 Eric Botcazou <ebotcazou@adacore.com>
18625 * tree-nested.c (get_descriptor_type): Make sure that the alignment of
18626 descriptors is at least equal to that of functions.
18628 2017-04-02 Uros Bizjak <ubizjak@gmail.com>
18630 * config/i386/sse.md (movdi_to_sse): Add missing DONE.
18632 2017-04-02 Uros Bizjak <ubizjak@gmail.com>
18635 * config/i386/sse.md (mov<IMOD4:mode>): Remove insn pattern.
18636 (mov<IMOD4:mode>): New expander.
18637 (*mov<IMOD4:mode>_internal): New insn and split pattern.
18639 2017-03-31 Segher Boessenkool <segher@kernel.crashing.org>
18641 PR rtl-optimization/79405
18642 * fwprop.c (propagations_left): New variable.
18643 (forward_propagate_into): Decrement it.
18644 (fwprop_init): Initialize it.
18645 (fw_prop): If the variable has reached zero, stop propagating.
18646 (fwprop_addr): Ditto.
18648 2017-03-31 Jakub Jelinek <jakub@redhat.com>
18651 * dwarf2out.c (decls_for_scope): If BLOCK_NONLOCALIZED_VAR is
18652 a FUNCTION_DECL, pass it as decl instead of origin to
18655 2017-03-31 Alexander Monakov <amonakov@ispras.ru>
18657 * config/nvptx/nvptx.c (nvptx_output_softstack_switch): Correct format
18660 2017-03-31 Pat Haugen <pthaugen@us.ibm.com>
18663 * config/rs6000/rs6000.md (extendhi<mode>2): Add test for
18664 TARGET_VSX_SMALL_INTEGER.
18666 2017-03-31 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18668 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Add
18669 reference to the OpenPOWER 64-Bit ELF V2 ABI Specification.
18671 2017-03-31 Matthew Fortune <matthew.fortune@imgtec.com>
18673 * config/mips/mips-msa.md (msa_vec_extract_<msafmt_f>): Update
18674 extraction from odd-numbered MSA register.
18676 2017-03-31 Jakub Jelinek <jakub@redhat.com>
18678 PR middle-end/80173
18679 * expmed.c (store_bit_field_1): Don't attempt to create
18680 a word subreg out of hard registers wider than word if they
18681 have HARD_REGNO_NREGS of 1 for their mode.
18683 PR middle-end/80163
18684 * varasm.c (initializer_constant_valid_p_1): Disallow sign-extending
18685 conversions to integer types wider than word and pointer.
18688 * cselib.h (rtx_equal_for_cselib_1): Add depth argument.
18689 (rtx_equal_for_cselib_p): Pass 0 to it.
18690 * cselib.c (cselib_hasher::equal): Likewise.
18691 (rtx_equal_for_cselib_1): Add depth argument. If depth
18692 is 128, don't look up VALUE locs and punt. Increment
18693 depth in recursive calls when walking VALUE locs.
18695 2017-03-31 Bernd Edlinger <bernd.edlinger@hotmail.de>
18697 * gcov.c (md5sum_to_hex): Fix output of MD5 hex bytes.
18698 (make_gcov_file_name): Use the canonical path name for generating
18700 (read_line): Fix handling of files with ascii null bytes.
18702 2017-03-30 Matthew Fortune <matthew.fortune@imgtec.com>
18704 * config/mips/mips.c (mips_expand_vector_init): Create a const_vector
18705 to initialise a vector register instead
18706 of using a const_int.
18708 2017-03-30 Jakub Jelinek <jakub@redhat.com>
18710 PR translation/80189
18711 * gimplify.c (omp_default_clause): Use %qs instead of %s in
18712 diagnostic messages.
18714 2017-03-30 Peter Bergner <bergner@vnet.ibm.com>
18717 * config/rs6000/dfp.md (dfp_dxex_<mode>): Update mode of operand 0.
18718 (dfp_diex_<mode>): Update mode of operand 1.
18719 * doc/extend.texi (dxex, dxexq): Document change to return type.
18720 (diex, diexq): Document change to argument type.
18722 2017-03-30 Martin Jambor <mjambor@suse.cz>
18725 * cgraph.h (cgraph_build_function_type_skip_args): Declare.
18726 * cgraph.c (redirect_call_stmt_to_callee): Set gimple fntype so that
18727 it reflects the signature changes performed at the callee side.
18728 * cgraphclones.c (build_function_type_skip_args): Make public, renamed
18729 to cgraph_build_function_type_skip_args.
18730 (build_function_decl_skip_args): Adjust call to the above function.
18732 2017-03-30 Jakub Jelinek <jakub@redhat.com>
18735 * config/i386/sse.md
18736 (<extract_type>_vextract<shuffletype><extract_suf>_mask): Use
18737 register as dest whenever it is a MEM not rtx_equal_p to the
18738 corresponding dup operand, and when forcing into reg move the
18739 reg into the memory afterwards.
18740 (<extract_type_2>_vextract<shuffletype><extract_suf_2>_mask):
18741 Likewise. Use <ssehalfvecmode> instead of <ssequartermode>
18742 for the force_reg mode.
18743 (avx512vl_vextractf128<mode>): Use register as dest either
18744 always when a MEM, or when it is a MEM not rtx_equal_p to the
18745 corresponding dup operand, or even not when it is a CONST_VECTOR
18746 depending on the mode and lo vs. hi.
18747 (avx512dq_vextract<shuffletype>64x2_1_maskm): Remove extraneous
18749 (avx512f_vextract<shuffletype>32x4_1_maskm): Likewise.
18750 (<mask_codefor>avx512dq_vextract<shuffletype>64x2_1<mask_name>):
18751 Likewise. Require that operands[2] is even.
18752 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>):
18753 Remove extraneous parens. Require that operands[2] is a multiple
18755 (vec_extract_lo_<mode><mask_name>): Don't bother testing if
18756 operands[0] is a MEM if <mask_applied>, the predicates/constraints
18757 disallow memory then.
18759 2017-03-30 Richard Biener <rguenther@suse.de>
18761 PR tree-optimization/77498
18762 * tree-ssa-pre.c (phi_translate_1): Do not allow simplifications
18763 to non-constants over backedges.
18765 2017-03-29 Segher Boessenkool <segher@kernel.crashing.org>
18767 PR rtl-optimization/80233
18768 * combine.c (combine_instructions): Only take NONDEBUG_INSN_P insns
18769 as last_combined_insn. Do not test for BARRIER_P separately.
18771 2017-03-29 Andreas Schwab <schwab@suse.de>
18774 * calls.c (prepare_call_address): Convert funexp to Pmode before
18775 copying to temp reg.
18777 2017-03-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18779 PR tree-optimization/80158
18780 * gimple-ssa-strength-reduction.c (replace_mult_candidate):
18781 Handle possible future case of more than one alternate
18783 (replace_rhs_if_not_dup): Likewise.
18784 (replace_one_candidate): Likewise.
18786 2017-03-28 Vladimir Makarov <vmakarov@redhat.com>
18788 PR rtl-optimization/80193
18789 * ira.c (ira): Do not check allocation for LRA.
18791 2017-03-28 Alexander Monakov <amonakov@ispras.ru>
18793 * config/nvptx/nvptx-protos.h (nvptx_output_simt_enter): Declare.
18794 (nvptx_output_simt_exit): Declare.
18795 * config/nvptx/nvptx.c (nvptx_init_unisimt_predicate): Use
18796 cfun->machine->unisimt_location. Handle NULL unisimt_predicate.
18797 (init_softstack_frame): Move initialization of crtl->is_leaf to...
18798 (nvptx_declare_function_name): ...here. Emit declaration of local
18799 memory space buffer for omp_simt_enter insn.
18800 (nvptx_output_unisimt_switch): New.
18801 (nvptx_output_softstack_switch): New.
18802 (nvptx_output_simt_enter): New.
18803 (nvptx_output_simt_exit): New.
18804 * config/nvptx/nvptx.h (struct machine_function): New fields
18805 has_simtreg, unisimt_location, simt_stack_size, simt_stack_align.
18806 * config/nvptx/nvptx.md (UNSPECV_SIMT_ENTER): New unspec.
18807 (UNSPECV_SIMT_EXIT): Ditto.
18808 (omp_simt_enter_insn): New insn.
18809 (omp_simt_enter): New expansion.
18810 (omp_simt_exit): New insn.
18811 * config/nvptx/nvptx.opt (msoft-stack-reserve-local): New option.
18813 * internal-fn.c (expand_GOMP_SIMT_ENTER): New.
18814 (expand_GOMP_SIMT_ENTER_ALLOC): New.
18815 (expand_GOMP_SIMT_EXIT): New.
18816 * internal-fn.def (GOMP_SIMT_ENTER): New internal function.
18817 (GOMP_SIMT_ENTER_ALLOC): Ditto.
18818 (GOMP_SIMT_EXIT): Ditto.
18819 * target-insns.def (omp_simt_enter): New insn.
18820 (omp_simt_exit): Ditto.
18821 * omp-low.c (struct omplow_simd_context): New fields simt_eargs,
18823 (lower_rec_simd_input_clauses): Implement SIMT privatization.
18824 (lower_rec_input_clauses): Likewise.
18825 (lower_lastprivate_clauses): Handle SIMT privatization.
18827 * omp-offload.c: Include langhooks.h, tree-nested.h, stor-layout.h.
18828 (ompdevlow_adjust_simt_enter): New.
18829 (find_simtpriv_var_op): New.
18830 (execute_omp_device_lower): Handle IFN_GOMP_SIMT_ENTER,
18831 IFN_GOMP_SIMT_ENTER_ALLOC, IFN_GOMP_SIMT_EXIT.
18833 * tree-inline.h (struct copy_body_data): New field dst_simt_vars.
18834 * tree-inline.c (expand_call_inline): Handle SIMT privatization.
18835 (copy_decl_for_dup_finish): Ditto.
18837 * tree-ssa.c (execute_update_addresses_taken): Handle GOMP_SIMT_ENTER.
18839 2017-03-28 Uros Bizjak <ubizjak@gmail.com>
18842 * config/i386/i386.c (ix86_option_override_internal): Always
18843 allow -mpreferred-stack-boundary=3 for 64-bit targets.
18845 2017-03-28 Bin Cheng <bin.cheng@arm.com>
18847 * tree-vect-loop.c (optimize_mask_stores): Add bb to the right loop.
18849 2017-03-28 Bin Cheng <bin.cheng@arm.com>
18851 * tree-vect-loop-manip.c (slpeel_add_loop_guard): New param and
18852 mark new edge's irreducible flag accordign to it.
18853 (vect_do_peeling): Check loop preheader edge's irreducible flag
18854 and pass it to function slpeel_add_loop_guard.
18856 2017-03-28 Richard Sandiford <richard.sandiford@arm.com>
18858 PR tree-optimization/80218
18859 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
18860 Update block frequencies and counts.
18862 2017-03-28 Richard Biener <rguenther@suse.de>
18864 PR tree-optimization/78644
18865 * tree-ssa-ccp.c (evaluate_stmt): When we may not use the value
18866 of a simplification result we may not use it at all.
18868 2017-03-28 Richard Biener <rguenther@suse.de>
18871 * tree-inline.c (copy_phis_for_bb): Do not create PHI node
18872 without arguments, generate default definition of a SSA name.
18874 2017-03-28 Richard Biener <rguenther@suse.de>
18876 PR middle-end/80222
18877 * gimple-fold.c (gimple_fold_indirect_ref): Do not touch
18878 TYPE_REF_CAN_ALIAS_ALL references.
18879 * fold-const.c (fold_indirect_ref_1): Likewise.
18881 2017-03-28 Martin Liska <mliska@suse.cz>
18884 * cgraphunit.c (cgraph_node::expand_thunk): Mark argument of a
18885 thunk call as DECL_GIMPLE_REG_P when vector or complex type.
18887 2017-03-28 Claudiu Zissulescu <claziss@synopsys.com>
18888 Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
18890 * config/arc/arc.h (CPP_SPEC): Add subtarget_cpp_spec.
18891 (EXTRA_SPECS): Define.
18892 (SUBTARGET_EXTRA_SPECS): Likewise.
18893 (SUBTARGET_CPP_SPEC): Likewise.
18894 * config/arc/elf.h (EXTRA_SPECS): Renamed to
18895 SUBTARGET_EXTRA_SPECS.
18896 * config/arc/linux.h (SUBTARGET_CPP_SPEC): Define.
18898 2017-03-28 Claudiu Zissulescu <claziss@synopsys.com>
18900 * config/arc/simdext.md (vst64_insn): Update pattern.
18901 (vld32wh_insn): Likewise.
18902 (vld32wl_insn): Likewise.
18903 (vld64_insn): Likewise.
18904 (vld32_insn): Likewise.
18906 2017-03-28 Marek Polacek <polacek@redhat.com>
18909 * fold-const.c (fold_comparison): Use protected_set_expr_location
18910 instead of SET_EXPR_LOCATION.
18912 2017-03-28 Markus Trippelsdorf <markus@trippelsdorf.de>
18914 * tree.c (add_expr): Avoid name lookup warning.
18916 2017-03-27 Jeff Law <law@redhat.com>
18918 PR tree-optimization/80216
18919 * tree-ssa-dom.c (derive_equivalences_from_bit_ior): Fix typo in
18920 function name. Limit recursion depth.
18921 (record_temporary_equivalences): Corresponding changes.
18923 2017-03-27 Jonathan Wakely <jwakely@redhat.com>
18925 * doc/invoke.texi (-Wno-narrowing): Reorder so default behavior is
18928 2017-03-27 Jakub Jelinek <jakub@redhat.com>
18931 * reg-notes.def (REG_CFA_NOTE): Define. Use it for CFA related
18933 * cfgcleanup.c (reg_note_cfa_p): New array.
18934 (insns_have_identical_cfa_notes): New function.
18935 (old_insns_match_p): Don't cross-jump in between /f
18936 and non-/f instructions. If both i1 and i2 are frame related,
18937 verify all CFA notes, their order and content.
18939 2017-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
18942 * config/rs6000/rs6000.md (bswaphi2_extenddi): Combine bswap
18943 HImode and SImode with zero extend to DImode to one insn.
18944 (bswap<mode>2_extenddi): Likewise.
18945 (bswapsi2_extenddi): Likewise.
18946 (bswaphi2_extendsi): Likewise.
18947 (bswaphi2): Combine bswap HImode and SImode into one insn.
18948 Separate memory insns from swapping register.
18949 (bswapsi2): Likewise.
18950 (bswap<mode>2): Likewise.
18951 (bswaphi2_internal): Delete, no longer used.
18952 (bswapsi2_internal): Likewise.
18953 (bswap<mode>2_load): Split bswap HImode/SImode into separate load,
18954 store, and gpr<-gpr swap insns.
18955 (bswap<mode>2_store): Likewise.
18956 (bswaphi2_reg): Register only splitter, combine with the splitter.
18957 (bswaphi2 splitter): Likewise.
18958 (bswapsi2_reg): Likewise.
18959 (bswapsi2 splitter): Likewise.
18960 (bswapdi2): If we have the LDBRX and STDBRX instructions, split
18961 the insns into load, store, and register/register insns.
18962 (bswapdi2_ldbrx): Likewise.
18963 (bswapdi2_load): Likewise.
18964 (bswapdi2_store): Likewise.
18965 (bswapdi2_reg): Likewise.
18967 2017-03-27 Gunther Nikl <gnikl@users.sourceforge.net>
18969 * system.h (HAVE_DESIGNATED_INITIALIZERS): Fix non C++ case.
18970 (HAVE_DESIGNATED_UNION_INITIALIZERS): Likewise.
18972 2017-03-27 Kelvin Nilsen <kelvin@gcc.gnu.org>
18975 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Edit and
18977 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
18978 special handling for target option conflicts between dform
18979 options (-mpower9-dform, -mpower9-dform-vector,
18980 -mpower9-dform-scalar) and -mno-direct-move.
18982 2017-03-27 Richard Biener <rguenther@suse.de>
18984 PR tree-optimization/80181
18985 * tree-ssa-ccp.c (likely_value): UNDEFINED ^ X is UNDEFINED.
18987 2017-03-27 Claudiu Zissulescu <claziss@synopsys.com>
18989 * config/arc/predicates.md (move_double_src_operand): Replace the
18990 call to move_double_src_operand with a call to address_operand.
18992 2017-03-27 Claudiu Zissulescu <claziss@synopsys.com>
18994 * config/arc/elf.h (ARGET_ARC_TP_REGNO_DEFAULT): Define.
18995 * config/arc/linux.h (ARGET_ARC_TP_REGNO_DEFAULT): Likewise.
18996 * config/arc/arc.opt (mtp-regno): Use ARGET_ARC_TP_REGNO_DEFAULT.
18998 2017-03-27 Claudiu Zissulescu <claziss@synopsys.com>
19000 * config/arc/predicates.md (long_immediate_loadstore_operand):
19001 Consider scaled addresses cases.
19003 2017-03-27 Claudiu Zissulescu <claziss@synopsys.com>
19005 * config/arc/arc.c (arc_epilogue_uses): BLINK should be also
19006 restored when in interrupt.
19007 * config/arc/arc.md (simple_return): ARCv2 rtie instruction
19008 doesn't have delay slot.
19010 2017-03-27 Richard Biener <rguenther@suse.de>
19013 * tree-ssa-structalias.c (associate_varinfo_to_alias): Skip
19014 inlined thunk clones.
19016 2017-03-27 Jakub Jelinek <jakub@redhat.com>
19019 * asan.c (instrument_derefs): Copy over last operand from
19020 original COMPONENT_REF to the new COMPONENT_REF with
19021 DECL_BIT_FIELD_REPRESENTATIVE.
19022 * ubsan.c (instrument_object_size): Likewise.
19024 2017-03-27 Richard Biener <rguenther@suse.de>
19026 PR tree-optimization/80170
19027 * tree-vect-data-refs.c (vect_compute_data_ref_alignment): Make
19028 sure DR/SCEV didnt fold in constants we do not see when looking
19029 at the reference base alignment.
19031 2017-03-27 Richard Biener <rguenther@suse.de>
19033 PR middle-end/80171
19034 * gimple-fold.c (fold_ctor_reference): Properly guard against
19035 NULL return value from canonicalize_constructor_val.
19037 2017-03-25 Uros Bizjak <ubizjak@gmail.com>
19040 * config/i386/i386.c (ix86_expand_builtin)
19041 <IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Do not expand arg0 between
19042 flags reg setting and flags reg using instructions.
19043 <IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Ditto. Use non-flags reg
19044 clobbering instructions to zero extend op2.
19046 2017-03-25 Gerald Pfeifer <gerald@pfeifer.com>
19048 * doc/install.texi (Configuration) <--with-aix-soname>:
19049 Update link to AIX ld.
19051 2017-03-25 Bernd Schmidt <bschmidt@redhat.com>
19053 PR rtl-optimization/80160
19054 PR rtl-optimization/80159
19055 * lra-assigns.c (must_not_spill_p): Tighten new test to also take
19056 reg_alternate_class into account.
19058 2017-03-24 Vladimir Makarov <vmakarov@redhat.com>
19061 * lra-assigns.c (assign_by_spills): Add spilled non-reload pseudos
19062 to consider in curr_insn_transform.
19064 2017-03-24 Jakub Jelinek <jakub@redhat.com>
19066 * genrecog.c (validate_pattern): Add VEC_SELECT validation.
19067 * genmodes.c (emit_min_insn_modes_c): Call emit_mode_nunits
19068 and emit_mode_inner.
19070 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19072 * config/s390/s390-builtins.def: Add VXE builtins. Add a flags
19073 argument to the overloaded builtin variants. Use the new flag to
19074 deprecate certain builtin variants.
19075 * config/s390/s390-builtin-types.def: Add new builtin types.
19076 * config/s390/s390-builtins.h: Support new flags field for
19077 overloaded builtins.
19078 * config/s390/s390-c.c (OB_DEF_VAR): New flags field.
19079 (s390_macro_to_expand): Enable vector float data type.
19080 (s390_cpu_cpp_builtins_internal): Indicate support of the new
19081 builtins by incrementing the __VEC__ version number.
19082 (s390_expand_overloaded_builtin): Support expansion of vec_xl and
19084 (s390_resolve_overloaded_builtin): Emit error messages depending
19085 on the builtin flags.
19086 * config/s390/s390.c (s390_expand_builtin): Support additional
19087 flags argument. Change error message to match the messages
19088 emitted in s390-c.c.
19089 * config/s390/s390.md: New UNSPEC_* constants.
19090 (op_type): Add new instruction types.
19091 * config/s390/vecintrin.h: Add new builtins and test data class
19093 * config/s390/vx-builtins.md (V_HW_32_64): Add V4SF.
19094 (V_HW_4, VEC_HW, VECF_HW): New mode iterators.
19095 (VEC_INEXACT, VEC_NOINEXACT): New constants.
19096 ("vec_splats<mode>", "vec_insert<mode>", "vec_promote<mode>")
19097 ("vec_insert_and_zero<mode>", "vec_mergeh<mode>")
19098 ("vec_mergel<mode>"): V_HW -> VEC_HW.
19100 ("vlrlrv16qi", "vstrlrv16qi", "vbpermv16qi", "vec_msumv2di")
19101 ("vmslg", "*vftci<mode>_cconly", "vftci<mode>_intcconly")
19102 ("*vftci<mode>", "vftci<mode>_intcc", "vec_double_s64")
19103 ("vec_double_u64", "vfmin<mode>", "vfmax<mode>"): New definition.
19105 ("and_av2df3", "and_cv2df3", "vec_andc_av2df3")
19106 ("vec_andc_cv2df3", "xor_av2df3", "xor_cv2df3", "vec_nor_av2df3")
19107 ("vec_nor_cv2df3", "ior_av2df3", "ior_cv2df3", "vec_nabs")
19108 ("*vftcidb", "*vftcidb_cconly", "vftcidb"): Remove definition.
19110 ("vec_all_<fpcmpcc:code>v2df", "vec_any_<fpcmpcc:code>v2df")
19111 ("vec_scatter_elementv4si_DI", "vec_cmp<fpcmp:code>v2df")
19112 ("vec_di_to_df_s64", "vec_di_to_df_u64", "vec_df_to_di_u64")
19113 ("vfidb", "*vldeb", "*vledb", "*vec_cmp<insn_cmp>v2df_cconly")
19114 ("vec_cmpeqv2df_cc", "vec_cmpeqv2df_cc", "vec_cmphv2df_cc")
19115 ("vec_cmphev2df_cc", "*vec_cmpeqv2df_cc")
19116 ("*vec_cmphv2df_cc", "*vec_cmphev2df_cc"): Enable new modes as ...
19118 ("vec_all_<fpcmpcc:code><mode>", "vec_any_<fpcmpcc:code><mode>")
19119 ("vec_scatter_element<V_HW_4:mode>_DI")
19120 ("vec_cmp<fpcmp:code><mode>", "vcdgb", "vcdlgb", "vclgdb")
19121 ("vec_fpint<mode>", "vflls")
19122 ("vflrd", "*vec_cmp<insn_cmp><mode>_cconly", "vec_cmpeq<mode>_cc")
19123 ("vec_cmpeq<mode>_cc", "vec_cmph<mode>_cc", "vec_cmphe<mode>_cc")
19124 ("*vec_cmpeq<mode>_cc", "*vec_cmph<mode>_cc")
19125 ("*vec_cmphe<mode>_cc"): ... these.
19127 ("vec_ctd_s64", "vec_ctsl", "vec_ctul", "vec_st2f"): Use rounding
19128 mode constant instead of magic value.
19130 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19132 * config/s390/s390.c (s390_expand_vec_compare): Support other
19133 vector floating point modes than just V2DF.
19134 (s390_expand_vcond): Likewise.
19135 (s390_hard_regno_mode_ok): Allow SFmode values in VRs.
19136 (s390_cannot_change_mode_class): Prevent mode changes between TF
19137 and V1TF in vector registers.
19138 * config/s390/s390.md (DF, SF): New mode attributes.
19139 ("*cmp<mode>_ccs", "add<mode>3", "sub<mode>3", "mul<mode>3")
19140 ("fma<mode>4", "fms<mode>4", "div<mode>3", "*neg<mode>2"): Add
19141 SFmode support for VRs.
19142 * config/s390/vector.md (V_HW, V_HW2, VT_HW, ti*, nonvec): Add new
19144 (VFT, VF_HW): New mode iterators.
19145 (vw, sdx): New mode attributes.
19146 ("addv2df3", "subv2df3", "mulv2df3", "divv2df3", "sqrtv2df2")
19147 ("fmav2df4","fmsv2df4", "negv2df2", "absv2df2", "*negabsv2df2")
19148 ("smaxv2df3", "sminv2df3", "*vec_cmp<VFCMP_HW_OP:code>v2df_nocc")
19149 ("vec_cmpuneqv2df", "vec_cmpltgtv2df", "vec_orderedv2df")
19150 ("vec_unorderedv2df"): Adjust the v2df only patterns to support
19151 also the new vector floating point modes. Renaming to ...
19153 ("add<mode>3", "sub<mode>3", "mul<mode>3", "div<mode>3")
19154 ("sqrt<mode>2", "fma<mode>4", "fms<mode>4", "neg<mode>2")
19155 ("abs<mode>2", "negabs<mode>2", "smax<mode>3")
19156 ("smin<mode>3", "*vec_cmp<VFCMP_HW_OP:code><mode>_nocc")
19157 ("vec_cmpuneq<mode>", "vec_cmpltgt<mode>", "vec_ordered<mode>")
19158 ("vec_unordered<mode>"): ... these.
19160 ("neg_fma<mode>4", "neg_fms<mode>4", "*smax<mode>3_vxe")
19161 ("*smin<mode>3_vxe", "*sminv2df3_vx", "*vec_extendv4sf")
19162 ("*vec_extendv2df"): New insn definitions.
19164 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19166 * config/s390/s390.md ("*adddi3_sign", "*subdi3_sign", "mulditi3")
19167 ("mulditi3_2", "*muldi3_sign"): New patterns.
19168 ("muldi3", "*muldi3", "mulsi3", "*mulsi3"): Add an expander and
19169 rename the pattern definition.
19171 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19173 * config/s390/s390.md ("indirect_jump"): Turn insn definition into
19175 ("*indirect_jump", "*indirect2_jump"): New pattern definitions.
19177 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19179 * config/s390/s390.c (s390_expand_vec_init): Use vllezl
19180 instruction if possible.
19181 * config/s390/vector.md (vec_halfnumelts): New mode
19183 ("*vec_vllezlf<mode>"): New pattern.
19185 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19187 * config/s390/vector.md ("popcountv16qi2", "popcountv8hi2")
19188 ("popcountv4si2", "popcountv2di2"): Rename to ...
19189 ("popcount<mode>2", "popcountv8hi2_vx", "popcountv4si2_vx")
19190 ("popcountv2di2_vx"): ... these and add !TARGET_VXE to the
19192 ("popcount<mode>2_vxe"): New pattern.
19194 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19196 * common/config/s390/s390-common.c (processor_flags_table): Add
19198 * config.gcc: Add arch12.
19199 * config/s390/driver-native.c (s390_host_detect_local_cpu):
19200 Default to arch12 for unknown CPU model numbers.
19201 * config/s390/s390-builtins.def: Add B_VXE builtin flag.
19202 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Adjust
19203 PROCESSOR_max sanity check.
19204 * config/s390/s390-opts.h (enum processor_type): Add
19206 * config/s390/s390.c (processor_table): Add arch12.
19207 (s390_expand_builtin): Add check for B_VXE flag.
19208 (s390_issue_rate): Add PROCESSOR_ARCH12.
19209 (s390_get_sched_attrmask): Likewise.
19210 (s390_get_unit_mask): Likewise.
19211 (s390_sched_score): Enable z13 scheduling for arch12.
19212 (s390_sched_reorder): Likewise.
19213 (s390_sched_variable_issue): Likewise.
19214 * config/s390/s390.h (enum processor_flags): Add PF_ARCH12 and
19216 (s390_tune_attr): Use z13 scheduling also for arch12.
19217 (TARGET_CPU_ARCH12, TARGET_CPU_ARCH12_P, TARGET_CPU_VXE)
19218 (TARGET_CPU_VXE_P, TARGET_ARCH12, TARGET_ARCH12_P, TARGET_VXE)
19219 (TARGET_VXE_P): New macros.
19220 * config/s390/s390.md: Add arch12 to cpu attribute. Add arch12
19221 and vxe to cpu_facility. Add arch12 and vxe to enabled attribute.
19222 * config/s390/s390.opt: Add arch12 as processor_type.
19224 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19226 * config/s390/s390.md
19227 ("fixuns_truncdddi2", "fixuns_trunctddi2")
19228 ("fixuns_trunc<BFP:mode><GPR:mode>2"): Merge into ...
19229 ("fixuns_trunc<FP:mode><GPR:mode>2"): New expander.
19231 ("fixuns_trunc<BFP:mode><GPR:mode>2", "fixuns_trunc<mode>si2"):
19232 Rename expanders to ...
19234 ("fixuns_trunc<BFP:mode><GPR:mode>2_emu")
19235 ("fixuns_truncdddi2_emu"): ... these.
19237 ("fixuns_trunc<mode>si2_emu"): New expander.
19239 ("*fixuns_truncdfdi2_z13"): Rename to ...
19240 ("*fixuns_truncdfdi2_vx"): ... this.
19242 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19244 * config/s390/2964.md: Remove the single element vector compare
19245 instructions which are no longer used.
19246 * config/s390/s390.c (s390_select_ccmode): Remove handling of
19248 (s390_canonicalize_comparison): Remove handling of DFmode
19250 (s390_expand_vec_compare_scalar): Remove function.
19251 (s390_emit_compare): Don't call s390_expand_vec_compare_scalar.
19252 * config/s390/s390.md ("*vec_cmp<insn_cmp>df_cconly"): Remove
19254 ("*cmp<mode>_ccs"): Add wfcdb instruction.
19256 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19258 * config/s390/s390.md ("mov<mode>_64dfp" DD_DF): Use vleig for loading a
19260 ("*mov<mode>_64" DD_DF): Remove the vector instructions. These
19261 will anyway by matched by mov<mode>_64dfp.
19263 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19265 * config/s390/s390.md ("mov<mode>" SD_SF): Change vleg/vsteg to
19266 vlef/vstef. Add missing operand to vleif.
19268 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19270 * config/s390/s390.c (s390_expand_vec_init): Enable vector load
19271 pair for all vector types with 64 bit elements.
19272 * config/s390/vx-builtins.md (V_HW_64): Move mode iterator to ...
19273 * config/s390/vector.md (V_HW_64): ... here.
19274 (V_128_NOSINGLE): New mode iterator.
19275 ("vec_init<V_HW:mode>"): Use V_128 as mode iterator.
19276 ("*vec_splat<mode>"): Use V_128_NOSINGLE mode iterator.
19277 ("*vec_tf_to_v1tf", "*vec_ti_to_v1ti"): New pattern definitions.
19278 ("*vec_load_pairv2di"): Change to ...
19279 ("*vec_load_pair<mode>"): ... this one.
19281 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19283 * config/s390/constraints.md: Add comments.
19284 (jKK): Reject element sizes > 8 bytes.
19285 * config/s390/s390.c (s390_split_ok_p): Enable splitting also for
19287 * config/s390/s390.md: Add the s_operand checks formerly in
19288 s390_split_ok_p to various splitters where they are still
19290 * config/s390/vector.md ("mov<mode>" V_128): Add GPR alternatives
19291 for 128 bit vectors. Plus two splitters.
19293 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19295 * config/s390/s390.md: Rename the cpu facilty vec to vx throughout
19298 2017-03-24 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19301 * config/s390/s390-c.c (s390_adjust_builtin_arglist): Issue an
19302 error if the boundary argument is not constant.
19304 2017-03-24 Jakub Jelinek <jakub@redhat.com>
19306 PR rtl-optimization/80112
19307 * loop-doloop.c (doloop_condition_get): Don't check condition
19308 if cmp isn't SET with IF_THEN_ELSE src.
19310 2017-03-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19312 PR tree-optimization/80158
19313 * gimple-ssa-strength-reduction.c (replace_mult_candidate): When
19314 replacing a candidate statement, also replace it for the
19315 candidate's alternate interpretation.
19316 (replace_rhs_if_not_dup): Likewise.
19317 (replace_one_candidate): Likewise.
19319 2017-03-24 Richard Biener <rguenther@suse.de>
19321 PR tree-optimization/80167
19322 * graphite-isl-ast-to-gimple.c
19323 (translate_isl_ast_to_gimple::is_valid_rename): Handle default-defs
19325 (translate_isl_ast_to_gimple::get_rename): Likewise.
19327 2017-03-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
19329 * config/rs6000/rs6000.c (rs6000_option_override_internal): Change
19330 handling of certain combinations of target options, including the
19331 combinations -mpower8-vector vs. -mno-vsx, -mpower9-vector vs.
19332 -mno-power8-vector, and -mpower9_dform vs. -mno-power9-vector.
19334 2017-03-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19337 * config/arm/arm.md (*load_multiple): Add reload_completed to
19338 matching condition.
19340 2017-03-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19341 Richard Biener <rguenth@suse.de>
19343 PR tree-optimization/79908
19344 PR tree-optimization/80136
19345 * tree-stdarg.c (expand_ifn_va_arg_1): For a VA_ARG whose LHS has
19346 been cast away, gimplify_and_add suffices.
19348 2017-03-23 Markus Trippelsdorf <markus@trippelsdorf.de>
19350 * tree-vrp.c (identify_jump_threads): Delete avail_exprs.
19352 2017-03-23 Richard Biener <rguenther@suse.de>
19354 PR tree-optimization/80032
19355 * gimplify.c (gimple_push_cleanup): Forced unconditional
19356 cleanups still have to go to the conditional_cleanups
19359 2017-03-22 Jakub Jelinek <jakub@redhat.com>
19361 PR tree-optimization/80072
19362 * tree-ssa-reassoc.c (struct operand_entry): Change id field type
19364 (next_operand_entry_id): Change type to unsigned int.
19365 (sort_by_operand_rank): Make sure to return the right return value
19366 even if unsigned fields are bigger than INT_MAX.
19367 (struct oecount): Change cnt and id type to unsigned int.
19368 (oecount_hasher::equal): Formatting fix.
19369 (oecount_cmp): Make sure to return the right return value
19370 even if unsigned fields are bigger than INT_MAX.
19371 (undistribute_ops_list): Change next_oecount_id type to unsigned int.
19374 * gimplify.c (gimplify_modify_expr_rhs) <case COND_EXPR>: Clear
19375 TREE_READONLY on result if writing it more than once.
19378 * doc/invoke.texi (-fsanitize=thread): Document that with
19379 -fnon-call-exceptions atomics are not able to throw
19383 * tsan.c: Include tree-eh.h.
19384 (instrument_builtin_call): Call maybe_clean_eh_stmt or
19385 maybe_clean_or_replace_eh_stmt where needed.
19386 (instrument_memory_accesses): Add cfg_changed argument.
19387 Call gimple_purge_dead_eh_edges on each block and set *cfg_changed
19388 if it returned true.
19389 (tsan_pass): Adjust caller. Return TODO_cleanup_cfg if cfg_changed.
19391 PR rtl-optimization/63191
19392 * config/i386/i386.c (ix86_delegitimize_address): Turn into small
19393 wrapper function, moved the whole old content into ...
19394 (ix86_delegitimize_address_1): ... this. New inline function.
19395 (ix86_find_base_term): Use ix86_delegitimize_address_1 with
19396 true as last argument instead of ix86_delegitimize_address.
19398 2017-03-22 Wilco Dijkstra <wdijkstr@arm.com>
19400 * config/aarch64/aarch64.c (generic_branch_cost): Copy
19401 cortexa57_branch_cost.
19403 2017-03-22 Wilco Dijkstra <wdijkstr@arm.com>
19405 * config/aarch64/aarch64.c (generic_tunings): Add AES fusion.
19407 2017-03-21 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
19410 * doc/md.texi (Constraints): Document wA constraint.
19411 * config/rs6000/constraints.md (wA): New.
19412 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Add wA reg_class.
19413 (rs6000_init_hard_regno_mode_ok): Init wA constraint.
19414 * config/rs6000/rs6000.h (RS6000_CONSTRAINT_wA): New.
19415 * config/rs6000/vsx.md (vsx_splat_<mode>): Use wA constraint.
19417 2017-03-22 Cesar Philippidis <cesar@codesourcery.com>
19420 * gimplify.c (is_oacc_declared): New function.
19421 (oacc_default_clause): Use it to set default flags for acc declared
19422 variables inside parallel regions.
19423 (gimplify_scan_omp_clauses): Strip firstprivate pointers for acc
19424 declared variables.
19425 (gimplify_oacc_declare): Gimplify the declare clauses. Add the
19426 declare attribute to any decl as necessary.
19428 2017-03-22 Thomas Preud'homme <thomas.preudhomme@arm.com>
19431 * config/arm/arm-isa.h (isa_bit_lpae): New feature bit.
19432 (ISA_ARMv7ve): Add isa_bit_lpae to the definition.
19433 * config/arm/arm-protos.h (arm_arch7ve): Rename into ...
19434 (arm_arch_lpae): This.
19435 * config/arm/arm.c (arm_arch7ve): Rename into ...
19436 (arm_arch_lpae): This. Define it in term of isa_bit_lpae.
19437 * config/arm/arm.h (TARGET_HAVE_LPAE): Redefine in term of
19440 2017-03-22 Martin Liska <mliska@suse.cz>
19443 * config/rs6000/rs6000.c (rs6000_inner_target_options): Show
19444 error message instead of an ICE.
19446 2017-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19448 * doc/extend.texi (6.11 Additional Floating Types): Revise.
19450 2017-03-21 Kelvin Nilsen <kelvin@gcc.gnu.org>
19452 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Add
19454 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
19457 2017-03-21 Martin Sebor <msebor@redhat.com>
19459 * doc/extend.texi: Use "cannot" instead of "can't."
19460 * doc/hostconfig.texi: Same.
19461 * doc/install.texi: Same.
19462 * doc/invoke.texi: Same.
19463 * doc/loop.texi: Same.
19464 * doc/md.texi: Same.
19465 * doc/objc.texi: Same.
19466 * doc/rtl.texi: Same.
19467 * doc/tm.texi: Same.
19468 * doc/tm.texi.in: Same.
19469 * doc/trouble.texi: Same.
19471 2017-03-21 Alexandre Oliva <aoliva@redhat.com>
19474 * dwarf2out.c (struct checksum_attributes): Add at_alignment.
19475 (collect_checksum_attributes): Set it.
19476 (die_checksum_ordered): Use it.
19478 2017-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19480 PR tree-optimization/79908
19481 * tree-stdarg.c (expand_ifn_va_arg_1): Revert the following
19482 change: For a VA_ARG whose LHS has been cast away, use
19483 force_gimple_operand to construct the side effects.
19485 2017-03-21 David Malcolm <dmalcolm@redhat.com>
19487 PR translation/80001
19488 * omp-offload.c (oacc_loop_fixed_partitions): Make diagnostics
19489 more amenable to translation.
19490 (oacc_loop_auto_partitions): Likewise.
19492 2017-03-21 Marek Polacek <polacek@redhat.com>
19493 Martin Sebor <msebor@redhat.com>
19495 PR tree-optimization/80109
19496 * gimple-ssa-warn-alloca.c (alloca_call_type): Only call get_range_info
19497 on INTEGRAL_TYPE_P.
19499 2017-03-21 Jakub Jelinek <jakub@redhat.com>
19500 Segher Boessenkool <segher@kernel.crashing.org>
19503 * combine.c (can_combine_p): Revert the 2017-03-20 change, only
19504 check reg_used_between_p between insn and one of succ or succ2
19505 depending on if succ is artificial insn not inserted into insn
19508 2017-03-21 Martin Liska <mliska@suse.cz>
19510 PR gcov-profile/80081
19511 * Makefile.in: Add gcov-dump and fix installation of gcov-tool.
19512 * doc/gcc.texi: Include gcov-dump stuff.
19513 * doc/gcov-dump.texi: New file.
19515 2017-03-21 Toma Tabacu <toma.tabacu@imgtec.com>
19517 PR rtl-optimization/79150
19518 * config/mips/mips.c (mips_block_move_loop): Emit a NOP after the
19519 conditional jump, if the jump is the last insn of the loop.
19521 2017-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19522 Richard Biener <rguenth@suse.de>
19524 PR tree-optimization/79908
19525 * tree-stdarg.c (expand_ifn_va_arg_1): For a VA_ARG whose LHS has
19526 been cast away, use force_gimple_operand to construct the side
19529 2017-03-21 Martin Liska <mliska@suse.cz>
19531 PR libfortran/79956
19532 * simplify-rtx.c (simplify_immed_subreg): Initialize a variable
19535 2017-03-21 Brad Spengler <spender@grsecurity.net>
19538 * plugin.c (htab_hash_plugin): New function.
19539 (add_new_plugin): Use it and adjust.
19540 (parse_plugin_arg_opt): Adjust.
19541 (init_one_plugin): Likewise.
19543 2017-03-21 Richard Biener <rguenther@suse.de>
19545 PR tree-optimization/80032
19546 * gimplify.c (gimple_push_cleanup): Add force_uncond parameter,
19547 if set force the cleanup to happen unconditionally.
19548 (gimplify_target_expr): Push inserted clobbers with force_uncond
19549 to avoid them being removed by control-dependent DCE.
19551 2017-03-21 Richard Biener <rguenther@suse.de>
19553 PR tree-optimization/80122
19554 * tree-inline.c (copy_bb): Do not expans va-arg packs or
19555 va_arg_pack_len when the inlined call stmt requires pack
19557 * tree-inline.h (struct copy_body_data): Make call_stmt a gcall *.
19559 2017-03-21 Jakub Jelinek <jakub@redhat.com>
19562 * tsan.c (instrument_builtin_call): If the memory model argument
19563 is not a constant, assume it is valid.
19566 * fold-const.c (round_up_loc): Negate divisor in unsigned type to
19569 2017-03-20 Segher Boessenkool <segher@kernel.crashing.org>
19571 PR rtl-optimization/79910
19572 * combine.c (can_combine_p): Do not allow combining an I0 or I1
19573 if its dest is used by an insn before I2 (other than the combined
19574 insns themselves, which are properly handled already).
19576 2017-03-20 Segher Boessenkool <segher@kernel.crashing.org>
19579 2017-03-17 Bernd Schmidt <bschmidt@redhat.com>
19581 * combine.c (record_used_regs): New static function.
19582 (try_combine): Handle situations where there is an additional
19583 instruction between I2 and I3 which needs to have a LOG_LINK
19587 2017-03-17 Jim Wilson <jim.wilson@linaro.org>
19589 * combine.c (try_combine): Delete redundant i1 test. Call
19590 prev_nonnote_nondebug_insn instead of prev_nonnote_insn.
19592 2017-03-20 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
19595 * config/rs6000/rs6000.md (*movsi_internal1): Fix constraints for
19596 alternatives 13/14.
19598 2017-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19600 PR tree-optimization/80054
19601 * gimple-ssa-strength-reduction.c (all_phi_incrs_profitable): Fail
19602 the optimization if a PHI or any of its arguments is not dominated
19603 by the candidate's basis. Use gphi* rather than gimple* as
19605 (replace_profitable_candidates): Clean up a gimple* variable that
19606 should be a gphi* variable.
19608 2017-03-20 Martin Sebor <msebor@redhat.com>
19611 * doc/extend.texi (attribute constructor): Document present limitation.
19613 2017-03-20 Kelvin Nilsen <kelvin@gcc.gnu.org>
19616 * config/rs6000/altivec.h (vec_all_ne): Under __cplusplus__ and
19617 __POWER9_VECTOR__ #ifdef control, change template definition to
19618 use Power9-specific built-in function.
19619 (vec_any_eq): Likewise.
19620 * config/rs6000/vector.md (vector_ae_v2di_p): Change the flag used
19621 to control outcomes from this test.
19622 (vector_ae_<mode>p): For VEC_F modes, likewise.
19624 2017-03-20 Ian Lance Taylor <iant@google.com>
19626 * config/i386/i386.c (ix86_function_regparm): Save an extra
19627 register for -fsplit-stack with DECL_STATIC_CHAIN.
19629 2017-03-17 Palmer Dabbelt <palmer@dabbelt.com>
19632 * config/riscv/riscv.c (riscv_preferred_reload_class): Remove.
19633 (TARGET_PREFERRED_RELOAD_CLASS): Likewise.
19635 2017-03-17 Palmer Dabbelt <palmer@dabbelt.com>
19637 * config/riscv/riscv.c (riscv_print_operand): Use "fence
19639 * config/riscv/sync.mc (mem_thread_fence_1): Use "fence
19642 2017-03-20 Marek Polacek <polacek@redhat.com>
19645 * asan.c (DEF_SANITIZER_BUILTIN): Use do { } while (0).
19647 2017-03-20 Richard Biener <rguenther@suse.de>
19649 PR tree-optimization/80113
19650 * graphite-isl-ast-to-gimple.c (copy_loop_phi_nodes): Do not
19651 allocate extra SSA name for PHI def.
19652 (add_close_phis_to_outer_loops): Likewise.
19653 (add_close_phis_to_merge_points): Likewise.
19654 (copy_loop_close_phi_args): Likewise.
19655 (copy_cond_phi_nodes): Likewise.
19657 2017-03-20 Martin Liska <mliska@suse.cz>
19659 PR middle-end/79753
19660 * tree-chkp.c (chkp_build_returned_bound): Do not build
19661 returned bounds for a LHS that's not a BOUNDED_P type.
19663 2017-03-20 Martin Liska <mliska@suse.cz>
19667 * tree-chkp.c (chkp_find_bounds_1): Handle REAL_CST,
19668 COMPLEX_CST and VECTOR_CST.
19670 2017-03-20 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
19673 * config/s390/s390.md ("cmp<mode>_ccs_0"): Add a clobber of the
19674 target operand. A new splitter adds the clobber statement in case
19675 the target operand is dead anyway.
19677 2017-03-19 Gerald Pfeifer <gerald@pfeifer.com>
19679 * doc/install.texi (Specific) <sparc-*-linux*>: No longer refer
19680 to age-old versions of binutils and glibc.
19682 2017-03-18 Segher Boessenkool <segher@kernel.crashing.org>
19684 * doc/contrib.texi (Contributors): Remove duplicate entry for myself.
19686 2017-03-18 Gerald Pfeifer <gerald@pfeifer.com>
19688 * doc/contrib.texi (Contributors): Add Segher Boessenkool.
19690 2017-03-18 Gerald Pfeifer <gerald@pfeifer.com>
19692 * doc/install.texi (Specific) <arm-*-eabi>: Remove old
19693 requirement for binutils 2.13.
19695 2017-03-17 Jim Wilson <jim.wilson@linaro.org>
19697 * combine.c (try_combine): Delete redundant i1 test. Call
19698 prev_nonnote_nondebug_insn instead of prev_nonnote_insn.
19700 2017-03-17 Palmer Dabbelt <palmer@dabbelt.com
19702 * doc/install.texi (Specific) <riscv32-*-elf>: Add riscv32-*-elf,
19703 riscv32-*-linux, riscv64-*-elf, riscv64-*-linux to the table of
19705 <riscv64-*-elf>: Re-arrange section
19706 <riscv32-*-elf>: Add a note about requiring binutils 2.28.
19707 <riscv32-*-linux>: Likewise.
19708 <riscv64-*-elf>: Likewise
19709 <riscv64-*-linux>: Likewise.
19711 2017-03-17 Richard Earnshaw <rearnsha@arm.com>
19714 * aarch64.opt(verbose-cost-dump): Fix typo.
19716 2017-03-17 Pat Haugen <pthaugen@us.ibm.com>
19719 * config/rs6000/rs6000.md (copysign<mode>3_fcpsgn): Test
19720 for VECTOR_UNIT_VSX_P (<MODE>mode) too.
19722 2017-03-17 Bernd Schmidt <bschmidt@redhat.com>
19724 * reload.c (find_reloads): When reloading a nonoffsettable address,
19725 use RELOAD_OTHER for it and its address reloads.
19727 PR rtl-optimization/79910
19728 * combine.c (record_used_regs): New static function.
19729 (try_combine): Handle situations where there is an additional
19730 instruction between I2 and I3 which needs to have a LOG_LINK
19733 2017-03-17 Jeff Law <law@redhat.com>
19735 PR tree-optimization/71437
19736 * tree-vrp.c (simplify_stmt_for_jump_threading): Lookup the
19737 conditional in the hash table first.
19738 (vrp_dom_walker::before_dom_children): Extract condition from
19739 ASSERT_EXPR. Record condition, its inverion and any implied
19740 conditions as well.
19742 2017-03-17 Marek Polacek <polacek@redhat.com>
19743 Markus Trippelsdorf <markus@trippelsdorf.de>
19745 PR tree-optimization/80079
19746 * gimple-ssa-store-merging.c (class pass_store_merging): Initialize
19749 2017-03-17 Richard Biener <rguenther@suse.de>
19751 PR middle-end/80075
19752 * tree-eh.c (stmt_could_throw_1_p): Only handle gimple assigns.
19753 Properly verify the LHS before the RHS possibly claims to be
19755 (stmt_could_throw_p): Hande gimple conds fully here. Clobbers
19758 2017-03-17 Martin Jambor <mjambor@suse.cz>
19760 * doc/invoke.texi (Option Options): Include -fipa-vrp in the list.
19761 (List of -O2 options): Likewise.
19762 (-fipa-bit-cp): Replace "ipa" with "interprocedural."
19765 2017-03-17 Tom de Vries <tom@codesourcery.com>
19767 * gcov-dump.c (print_usage): Print bug_report_url.
19769 2017-03-17 Richard Biener <rguenther@suse.de>
19771 PR middle-end/80050
19772 * genmatch.c (parser::next): Remove pointless check for CPP_EOF.
19773 (parser::peek): Likewise.
19775 2017-03-17 Richard Biener <rguenther@suse.de>
19777 PR tree-optimization/80048
19778 * sese.c (free_sese_info): Properly release rename_map and
19779 copied_bb_map elements.
19781 2017-03-16 Alexandre Oliva <aoliva@redhat.com>
19783 * gimple-ssa-store-merging.c (struct imm_store_chain_info):
19784 Add linked-list forward and backlinks. Insert on
19785 construction, remove on destruction.
19786 (class pass_store_merging): Add m_stores_head field.
19787 (pass_store_merging::terminate_and_process_all_chains):
19788 Iterate over m_stores_head list.
19789 (pass_store_merging::terminate_all_aliasing_chains):
19791 (pass_store_merging::execute): Check for debug stmts first.
19792 Push new chains onto the m_stores_head stack.
19794 2017-03-16 Michael Meissner <meissner@linux.vnet.ibm.com>
19797 * config/rs6000/vsx.md (vsx_splat_<mode>, VSX_D iterator): Allow a
19798 SPLAT operation on ISA 2.07 64-bit systems that have direct move,
19799 but no MTVSRDD support, by doing MTVSRD and XXPERMDI.
19801 2017-03-16 Jeff Law <law@redhat.com>
19803 PR tree-optimization/71437
19804 * tree-ssa-dom.c (dom_opt_dom_walker): Remove thread_across_edge
19805 member function. Implementation moved into after_dom_children
19806 member function and into the threader's thread_outgoing_edges
19808 (dom_opt_dom_walker::after_dom_children): Simplify by moving
19809 some code into new thread_outgoing_edges.
19810 * tree-ssa-threadedge.c (thread_across_edge): Make static and simplify
19811 definition. Simplify marker handling (do it here). Assume we always
19812 have the available expression and the const/copies tables.
19813 (thread_outgoing_edges): New function extracted from tree-ssa-dom.c
19815 * tree-ssa-threadedge.h (thread_outgoing_edges): Declare.
19816 * tree-vrp.c (equiv_stack): No longer file scoped.
19817 (vrp_dom_walker): New class.
19818 (vrp_dom_walker::before_dom_children): New member function.
19819 (vrp_dom_walker::after_dom_children): Likewise.
19820 (identify_jump_threads): Setup domwalker. Use it rather than
19821 walking edges in a random order by hand. Simplify setup/finalization.
19822 (finalize_jump_threads): Remove.
19823 (vrp_finalize): Do not call identify_jump_threads here.
19824 (execute_vrp): Do it here instead and call thread_through_all_blocks
19827 PR tree-optimization/71437
19828 * tree-ssa-dom.c (pfn_simplify): Add basic_block argument. All
19830 (simplify_stmt_for_jump_threading): Add basic_block argument. All
19832 (lhs_of_dominating_assert): Moved from here into tree-vrp.c.
19833 (dom_opt_dom_walker::thread_across_edge): Remove
19834 handle_dominating_asserts argument. All callers changed.
19835 (record_temporary_equivalences_from_stmts_at_dest): Corresponding
19836 changes. Remove calls to lhs_of_dominating_assert. Other
19837 uses of handle_dominating_asserts turn into unconditional code
19838 (simplify_control_stmt_condition_1): Likewise.
19839 (simplify_control_stmt_condition): Likewise.
19840 (thread_through_normal_block, thread_across_edge): Likewise.
19841 * tree-ssa-threadedge.h (thread_across_edge): Corresponding changes.
19842 * tree-vrp.c (lhs_of_dominating_assert): Move here. Return original
19843 object if it is not an SSA_NAME.
19844 (simplify_stmt_for_jump_threading): Call lhs_of_dominating_assert
19845 before calling into the VRP specific simplifiers.
19846 (identify_jump_threads): Remove handle_dominating_asserts
19849 2017-03-16 Jakub Jelinek <jakub@redhat.com>
19852 * tree-diagnostic.c (default_tree_printer): No longer static.
19853 * tree-diagnostic.h (default_tree_printer): New prototype.
19855 2017-03-16 Tamar Christina <tamar.christina@arm.com>
19857 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>)
19858 Change ins into fmov.
19860 2017-03-16 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19862 * config/aarch64/iterators.md (h_con): Return "x" for V4HF and V8HF.
19863 * config/aarch64/aarch64-simd.md (*aarch64_fma4_elt_from_dup<mode>):
19864 Use h_con constraint for operand 1.
19865 (*aarch64_fnma4_elt_from_dup<mode>): Likewise.
19866 (*aarch64_mulx_elt_from_dup<mode>): Likewise for operand 2.
19868 2017-03-15 Jeff Law <law@redhat.com>
19870 PR tree-optimization/71437
19871 * tree-ssa-dom.c (derive_equivalences_from_bit_ior): New function.
19872 (record_temporary_equivalences): Use it.
19874 PR tree-optimization/71437
19875 * tree-ssa-dom.c (struct cond_equivalence): Moved from here into
19876 tree-ssa-scopedtables.
19877 (lookup_avail_expr, build_and_record_new_cond): Likewise.
19878 (record_conditions, record_cond, vuse_eq): Likewise.
19879 (record_edge_info): Adjust to API tweak of record_conditions.
19880 (simplify_stmt_for_jump_threading): Similarly for lookup_avail_expr.
19881 (record_temporary_equivalences, optimize_stmt): Likewise.
19882 (eliminate_redundant_computations): Likewise.
19883 (record_equivalences_from_stmt): Likewise.
19884 * tree-ssa-scopedtables.c: Include options.h and params.h.
19885 (vuse_eq): New function, moved from tree-ssa-dom.c
19886 (build_and_record_new_cond): Likewise.
19887 (record_conditions): Likewise. Accept vector of conditions rather
19888 than edge_equivalence structure for first argument.
19889 for the first argument.
19890 (avail_exprs_stack::lookup_avail_expr): New member function, moved
19891 from tree-ssa-dom.c.
19892 (avail_exprs_stack::record_cond): Likewise.
19893 * tree-ssa-scopedtables.h (struct cond_equivalence): Moved here
19894 from tree-ssa-dom.c.
19895 (avail_exprs_stack): Add new member functions lookup_avail_expr
19897 (record_conditions): Declare.
19899 2017-03-15 Vladimir Makarov <vmakarov@redhat.com>
19902 * lra-constraints.c (process_alt_operands): Increase reject for
19903 reloading an input/output operand.
19905 2017-03-15 Michael Meissner <meissner@linux.vnet.ibm.com>
19908 * config/rs6000/rs6000.md (float<QHI:mode><IEEE128:mode>2): Define
19909 insns to convert from signed/unsigned char/short to IEEE 128-bit
19911 (floatuns<QHI:mode><IEEE128:mode>2): Likewise.
19913 2017-03-15 Uros Bizjak <ubizjak@gmail.com>
19916 * config/i386/i386.c (ix86_vector_duplicate_value): Create
19917 subreg of inner mode for values already in registers.
19919 2017-03-15 Bernd Schmidt <bschmidt@redhat.com>
19921 * config/c6x/c6x.c (hwloop_optimize): Handle case where the old
19922 iteration reg is used after the loop.
19924 2017-03-14 Martin Sebor <msebor@redhat.com>
19926 PR tree-optimization/79800
19927 * gimple-ssa-sprintf.c (format_floating: Add argument. Handle
19928 precision in negative-positive range.
19929 (format_floating): Call non-const overload with adjusted precision.
19931 2017-03-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19934 * config/rs6000/rs6000.h (TARGET_FRSQRTES): Add check for
19937 2017-03-14 Martin Sebor <msebor@redhat.com>
19939 PR middle-end/80020
19940 * builtin-attrs.def (ATTR_ALLOC_SIZE_2_NOTHROW_LIST): New macro.
19941 * builtins.def (aligned_alloc): Use it.
19944 * Makefile.in (GTFILES): Add calls.c.
19945 * calls.c: Include "gt-calls.h".
19947 2017-03-14 Bernd Schmidt <bschmidt@redhat.com>
19949 PR rtl-optimization/79728
19950 * regs.h (struct target_regs): New field
19951 x_contains_allocatable_regs_of_mode.
19952 (contains_allocatable_regs_of_mode): New macro.
19953 * reginfo.c (init_reg_sets_1): Initialize it, and change
19954 contains_reg_of_mode so it includes global regs as well.
19955 * reload.c (push_reload): Use contains_allocatable_regs_of_mode
19956 rather than contains_regs_of_mode.
19958 2017-03-14 Martin Liska <mliska@suse.cz>
19960 * doc/invoke.texi: Document options that can't be combined with
19961 -fcheck-pointer-bounds.
19963 2017-03-14 Martin Liska <mliska@suse.cz>
19965 PR middle-end/79831
19966 * doc/invoke.texi (-Wchkp): Document the option.
19968 2017-03-14 Martin Liska <mliska@suse.cz>
19970 * Makefile.in: Install gcov-dump.
19972 2017-03-14 Martin Liska <mliska@suse.cz>
19974 * multiple_target.c (expand_target_clones): Bail out for
19975 an invalid attribute.
19977 2017-03-14 Richard Biener <rguenther@suse.de>
19979 * alias.c (struct alias_set_entry): Pack properly.
19980 * cfgloop.h (struct loop): Likewise.
19981 * cse.c (struct set): Likewise.
19982 * ipa-utils.c (struct searchc_env): Likewise.
19983 * loop-invariant.c (struct invariant): Likewise.
19984 * lra-remat.c (struct cand): Likewise.
19985 * recog.c (struct change_t): Likewise.
19986 * rtl.h (struct address_info): Likewise.
19987 * symbol-summary.h (function_summary): Likewise.
19988 * tree-loop-distribution.c (struct partition): Likewise.
19989 * tree-object-size.c (struct object_size_info): Likewise.
19990 * tree-ssa-loop-ivopts.c (struct cost_pair): Likewise.
19991 * tree-ssa-threadupdate.c (struct ssa_local_info_t): Likewise.
19992 * tree-vect-data-refs.c (struct _vect_peel_info): Likewise.
19993 * tree-vect-slp.c (struct _slp_oprnd_info): Likewise.
19994 * tree-vect-stmts.c (struct simd_call_arg_info): Likewise.
19995 * tree-vectorizer.h (struct _loop_vec_info): Likewise.
19996 (struct _stmt_vec_info): Likewise.
19998 2017-03-14 Martin Liska <mliska@suse.cz>
20001 * multiple_target.c (create_dispatcher_calls): Check that
20002 a target can create a function dispatcher.
20004 2017-03-14 Martin Liska <mliska@suse.cz>
20007 * multiple_target.c (expand_target_clones): Drop local.local
20008 flag for default implementation.
20010 2017-03-14 Richard Biener <rguenther@suse.de>
20012 PR tree-optimization/80030
20013 * tree-vect-stmts.c (vectorizable_store): Plug memleak.
20015 2017-03-13 Kito Cheng <kito.cheng@gmail.com>
20017 * config/riscv/riscv.c (riscv_emit_float_compare>: Use
20018 gcc_fallthrough() instead of __attribute__((fallthrough));
20020 2017-03-13 Gerald Pfeifer <gerald@pfeifer.com>
20022 * doc/gcc.texi: Remove "up" link to (DIR).
20023 * doc/gccint.texi: Ditto.
20025 2017-03-13 Gerald Pfeifer <gerald@pfeifer.com>
20027 * doc/install.texi (Specific) <avr>: Remove reference to
20030 2017-03-13 Jeff Law <law@redhat.com>
20032 * config/riscv/riscv.c (riscv_emit_float_compare): Use fallthru
20033 attribute rather than comments.
20035 * config/pdp11/pdp11.md (movmemhi): Adjust operand numbers to
20036 match_scratch operand is highest.
20038 2017-03-13 Martin Liska <mliska@suse.cz>
20040 PR middle-end/78339
20041 * ipa-pure-const.c (warn_function_noreturn): If the declarations
20042 is a CHKP clone, use original declaration.
20044 2017-03-13 Claudiu Zissulescu <claziss@synopsys.com>
20046 * config/arc/arc.c (arc_init): Use multiplier whenever we have it.
20047 (arc_conditional_register_usage): Use a different allocation order
20048 when optimizing for size.
20049 * common/config/arc/arc-common.c (arc_option_optimization_table):
20050 Section anchors default on when optimizing for size.
20052 2017-03-13 Claudiu Zissulescu <claziss@synopsys.com>
20054 * config/arc/arc.md (*tst_bitfield_tst): Fix pattern.
20056 2017-03-13 Claudiu Zissulescu <claziss@synopsys.com>
20058 * config/arc/arc.c (arc_output_addsi): Emit code density adds.
20059 * config/arc/arc.md (cpu_facility): Add cd variant.
20060 (*movqi_insn): Add code density variant.
20061 (*movhi_insn): Likewise.
20062 (*movqi_insn): Likewise.
20063 (*addsi3_mixed): Likewise.
20064 (subsi3_insn): Likewise.
20066 2017-03-13 Claudiu Zissulescu <claziss@synopsys.com>
20068 * config/arc/arc.md (movsi_cond_exec): Update constraint.
20070 2017-03-13 Claudiu Zissulescu <claziss@synopsys.com>
20072 * config/arc/arc.c (arc_legitimize_pic_address): Handle PIC
20073 expressions with MINUS and UNARY ops.
20075 2017-03-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20078 * config/arm/neon.md (vec_sel_widen_ssum_lo<VQI:mode><VW:mode>3):
20080 (vec_sel_widen_ssum_lo<mode><V_half>3): ... This. Avoid mismatch
20081 between vec_select and vector argument.
20082 (vec_sel_widen_ssum_hi<VQI:mode><VW:mode>3): Rename to...
20083 (vec_sel_widen_ssum_hi<mode><V_half>3): ... This. Likewise.
20084 (vec_sel_widen_usum_lo<VQI:mode><VW:mode>3): Rename to...
20085 (vec_sel_widen_usum_lo<mode><V_half>3): ... This.
20086 (vec_sel_widen_usum_hi<VQI:mode><VW:mode>3): Rename to...
20087 (vec_sel_widen_usum_hi<mode><V_half>3): ... This.
20089 2017-03-13 Richard Biener <rguenther@suse.de>
20092 * params.def (vect-max-peeling-for-alignment): Fix typo.
20094 2017-03-12 Gerald Pfeifer <gerald@pfeifer.com>
20096 * doc/install.texi (Specific) <mips-*-*>: Remove description of
20097 issue that only occurred with binutils below 2.18.
20099 2017-03-12 Gerald Pfeifer <gerald@pfeifer.com>
20101 * doc/install.texi (Specific) <cris-axis-elf>: No longer
20102 refer to binutils 2.11/2.12 minimum.
20104 2017-03-12 Gerald Pfeifer <gerald@pfeifer.com>
20106 * doc/install.texi (Specific) <powerpc-*-*>: Remove link to
20107 ftp.kernel.org and simplify binutils requirement.
20109 2017-03-11 Gerald Pfeifer <gerald@pfeifer.com>
20111 * doc/invoke.texi (Warning Options): Fix spelling of link-time
20113 (Optimize Options): Ditto. Also remove redundancy.
20115 2017-03-10 David Malcolm <dmalcolm@redhat.com>
20117 PR translation/79848
20118 * ipa-devirt.c (warn_types_mismatch): Simplify uses of "%<%s%>" to
20120 * ipa-pure-const.c (suggest_attribute): Likewise. Convert _
20121 to G_ to avoid double translation.
20123 2017-03-10 David Malcolm <dmalcolm@redhat.com>
20125 PR translation/79923
20126 * auto-profile.c (get_combined_location): Convert leading
20127 character of diagnostics to lower case and remove trailing period.
20128 (read_profile): Likewise for various diagnostics.
20129 * config/arm/arm.c (arm_option_override): Remove trailing period
20130 from various diagnostics.
20131 * config/msp430/msp430.c (msp430_expand_delay_cycles): Likewise.
20132 (msp430_expand_delay_cycles): Likewise.
20134 2017-03-10 David Malcolm <dmalcolm@redhat.com>
20137 * config/aarch64/aarch64.c (aarch64_validate_mcpu): Quote the
20138 full command-line argument, rather than just "str".
20139 (aarch64_validate_march): Likewise.
20140 (aarch64_validate_mtune): Likewise.
20142 2017-03-10 Bernd Schmidt <bschmidt@redhat.com>
20144 PR rtl-optimization/78911
20145 * lra-assigns.c (must_not_spill_p): New function.
20146 (spill_for): Use it.
20148 2017-03-10 Jakub Jelinek <jakub@redhat.com>
20150 PR tree-optimization/79981
20151 * tree-vrp.c (extract_range_basic): Handle IMAGPART_EXPR of
20152 ATOMIC_COMPARE_EXCHANGE ifn result.
20153 (stmt_interesting_for_vrp, vrp_visit_stmt): Handle
20154 IFN_ATOMIC_COMPARE_EXCHANGE.
20156 2017-03-10 David Malcolm <dmalcolm@redhat.com>
20159 * opts.c (parse_sanitizer_options): Add missing question mark to
20160 "did you mean" message.
20162 2017-03-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20164 * config/rs6000/rs6000-builtin.def (VMULEUB_UNS): Remove orphaned
20166 (VMULEUH_UNS): Likewise.
20167 (VMULOUB_UNS): Likewise.
20168 (VMULOUH_UNS): Likewise.
20169 * config/rs6000/rs6000.c (builtin_function_type): Remove
20170 references to ALTIVEC_BUILTIN_VMUL[EO]U[BH]_UNS.
20172 2017-03-10 David Malcolm <dmalcolm@redhat.com>
20175 * read-rtl-function.c (function_reader::read_rtx_operand): Update
20176 x with result of extra_parsing_for_operand_code_0.
20177 (function_reader::extra_parsing_for_operand_code_0): Convert
20178 return type from void to rtx, returning x. When reading
20179 SYMBOL_REF with SYMBOL_FLAG_HAS_BLOCK_INFO, reallocate x to the
20180 larger size containing struct block_symbol.
20182 2017-03-10 Segher Boessenkool <segher@kernel.crashing.org>
20184 * config/rs6000/rs6000.c (rs6000_option_override_internal): Disallow
20185 -mfloat128-hardware without -m64.
20187 2017-03-10 Will Schmidt <will_schmidt@vnet.ibm.com>
20190 * config/rs6000/rs6000.c (builtin_function_type): Add VMUL*U[HB]
20191 entries to the case statement that marks unsigned arguments to
20192 overloaded functions.
20194 2017-03-10 Kelvin Nilsen <kelvin@gcc.gnu.org>
20196 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix
20197 two typographic errors in the handling of TARGET_UPPER_REGS_DI.
20199 2017-03-10 Pat Haugen <pthaugen@us.ibm.com>
20202 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Test
20203 TARGET_UPPER_REGS_DI when setting 'wi' constraint regclass.
20205 2017-03-10 Martin Liska <mliska@suse.cz>
20209 * toplev.c (process_options): Enable MPX with LSAN and UBSAN.
20210 * tree-chkp.c (chkp_walk_pointer_assignments): Verify that
20213 2017-03-10 Olivier Hainque <hainque@adacore.com>
20215 * tree-switch-conversion (array_value_type): Start by resetting
20216 candidate type to it's main variant.
20218 2017-03-10 Jakub Jelinek <jakub@redhat.com>
20220 PR rtl-optimization/79909
20221 * combine.c (try_combine): Use simplify_replace_rtx on individual
20222 CALL_INSN_FUNCTION_USAGE elements instead of replace_rtx on copy_rtx
20223 of the whole CALL_INSN_FUNCTION_USAGE.
20225 PR tree-optimization/79972
20226 * gimple-ssa-warn-alloca.c (alloca_call_type): Only call
20227 get_range_info on SSA_NAMEs. Formatting fixes.
20229 2017-03-10 Richard Biener <rguenther@suse.de>
20230 Jakub Jelinek <jakub@redhat.com>
20232 PR tree-optimization/77975
20233 * tree-ssa-loop-niter.c (get_base_for): Allow phi argument from latch
20234 edge to be constant.
20235 (get_val_for): For constant x return it. Formatting fix.
20236 (loop_niter_by_eval): Avoid pointless looping if the next iteration
20237 would use the same bases as the current one.
20239 2017-03-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20241 * config/rs6000/rs6000.c (rs6000_gen_le_vsx_permute): Use rotate
20242 instead of vec_select for V1TImode.
20243 * conifg/rs6000/vsx.md (VSX_LE): Remove mode iterator that is no
20245 (VSX_LE_128): Add V1TI to this mode iterator.
20246 (*vsx_le_perm_load_<mode>): Change to use VSX_D mode iterator.
20247 (*vsx_le_perm_store_<mode>): Likewise.
20248 (pre-reload splitter for VSX stores): Likewise.
20249 (post-reload splitter for VSX stores): Likewise.
20250 (*vsx_xxpermdi2_le_<mode>): Likewise.
20251 (*vsx_lxvd2x2_le_<mode>): Likewise.
20252 (*vsx_stxvd2x2_le_<mode>): Likewise.
20254 2017-03-09 Michael Eager <eager@eagercon.com>
20256 Correct failures with --enable-checking=yes,rtl.
20258 * config/microblaze/microblaze.c (microblaze_expand_shift):
20259 Replace GET_CODE test with CONST_INT_P and INTVAL test with
20260 test for const0_rtx.
20261 * config/microblaze/microblaze.md (ashlsi3_byone, ashrsi3_byone,
20262 lshrsi3_byone): Replace INTVAL with test for const1_rtx.
20264 2017-03-09 Richard Biener <rguenther@suse.de>
20266 PR tree-optimization/79977
20267 * graphite-scop-detection.c (scop_detection::merge_sese):
20268 Handle the case of extra exits to blocks dominating the entry.
20270 2017-03-09 Toma Tabacu <toma.tabacu@imgtec.com>
20272 * doc/sourcebuild.texi (Effective-Target Keywords, Other attributes):
20275 2017-03-09 Vladimir Makarov <vmakarov@redhat.com>
20277 PR rtl-optimization/79949
20278 * lra-constraints.c (process_alt_operands): Check memory when
20279 trying to predict a cycle. Print about the overall increase.
20281 2017-03-09 Richard Biener <rguenther@suse.de>
20283 PR middle-end/79971
20284 * gimple-expr.c (useless_type_conversion_p): Preserve
20285 TYPE_SATURATING for fixed-point types.
20287 2017-03-09 Richard Biener <rguenther@suse.de>
20290 * ipa-prop.c (ipa_modify_formal_parameters): Avoid changing
20291 alignment of BLKmode params.
20293 2017-03-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20296 * config/aarch64/iterators.md (VALL_F16_NO_V2Q): New mode iterator.
20297 (VALL_NO_V2Q): Likewise.
20299 * config/aarch64/aarch64-simd.md
20300 (aarch64_dup_lane_<vswap_width_name><mode>): Use VALL_F16_NO_V2Q
20302 (*aarch64_simd_vec_copy_lane_<vswap_width_name><mode>): Use
20303 VALL_NO_V2Q mode iterator.
20304 (*aarch64_vgetfmulx<mode>): Use VDQF iterator.
20306 2017-03-09 Martin Liska <mliska@suse.cz>
20308 PR tree-optimization/79631
20309 * tree-chkp-opt.c (chkp_is_constant_addr): Call
20310 tree_int_cst_sign_bit just for INTEGER constants.
20312 2017-03-09 Martin Liska <mliska@suse.cz>
20316 * toplev.c (process_options): Disable -fcheck-pointer-bounds with
20319 2017-03-09 Marek Polacek <polacek@redhat.com>
20322 * tree.c (inchash::add_expr): Handle TREE_VEC.
20324 2017-03-09 Martin Liska <mliska@suse.cz>
20327 (chkp_narrow_size_and_offset): New function.
20328 (chkp_parse_array_and_component_ref): Support BIT_FIELD_REF.
20329 (void chkp_parse_bit_field_ref): New function.
20330 (chkp_make_addressed_object_bounds): Add case for BIT_FIELD_REF.
20331 (chkp_process_stmt): Use chkp_parse_bit_field_ref.
20333 2017-03-09 Martin Liska <mliska@suse.cz>
20336 * tree-chkp.c (chkp_get_bound_for_parm): Get bounds for a param.
20337 (chkp_find_bounds_1): Remove gcc_unreachable.
20339 2017-03-09 Jakub Jelinek <jakub@redhat.com>
20342 * asan.c (get_mem_refs_of_builtin_call): For BUILT_IN_ATOMIC* and
20343 BUILT_IN_SYNC*, determine the access type from the size suffix and
20344 always build a MEM_REF with that type. Handle forgotten
20345 BUILT_IN_SYNC_FETCH_AND_NAND_16 and BUILT_IN_SYNC_NAND_AND_FETCH_16.
20348 * config/i386/avx512vlintrin.h (_mm256_cmpge_epi32_mask,
20349 _mm256_cmpge_epi64_mask, _mm256_cmpge_epu32_mask,
20350 _mm256_cmpge_epu64_mask, _mm256_cmple_epi32_mask,
20351 _mm256_cmple_epi64_mask, _mm256_cmple_epu32_mask,
20352 _mm256_cmple_epu64_mask, _mm256_cmplt_epi32_mask,
20353 _mm256_cmplt_epi64_mask, _mm256_cmplt_epu32_mask,
20354 _mm256_cmplt_epu64_mask, _mm256_cmpneq_epi32_mask,
20355 _mm256_cmpneq_epi64_mask, _mm256_cmpneq_epu32_mask,
20356 _mm256_cmpneq_epu64_mask, _mm256_mask_cmpge_epi32_mask,
20357 _mm256_mask_cmpge_epi64_mask, _mm256_mask_cmpge_epu32_mask,
20358 _mm256_mask_cmpge_epu64_mask, _mm256_mask_cmple_epi32_mask,
20359 _mm256_mask_cmple_epi64_mask, _mm256_mask_cmple_epu32_mask,
20360 _mm256_mask_cmple_epu64_mask, _mm256_mask_cmplt_epi32_mask,
20361 _mm256_mask_cmplt_epi64_mask, _mm256_mask_cmplt_epu32_mask,
20362 _mm256_mask_cmplt_epu64_mask, _mm256_mask_cmpneq_epi32_mask,
20363 _mm256_mask_cmpneq_epi64_mask, _mm256_mask_cmpneq_epu32_mask,
20364 _mm256_mask_cmpneq_epu64_mask, _mm_cmpge_epi32_mask,
20365 _mm_cmpge_epi64_mask, _mm_cmpge_epu32_mask, _mm_cmpge_epu64_mask,
20366 _mm_cmple_epi32_mask, _mm_cmple_epi64_mask, _mm_cmple_epu32_mask,
20367 _mm_cmple_epu64_mask, _mm_cmplt_epi32_mask, _mm_cmplt_epi64_mask,
20368 _mm_cmplt_epu32_mask, _mm_cmplt_epu64_mask, _mm_cmpneq_epi32_mask,
20369 _mm_cmpneq_epi64_mask, _mm_cmpneq_epu32_mask, _mm_cmpneq_epu64_mask,
20370 _mm_mask_cmpge_epi32_mask, _mm_mask_cmpge_epi64_mask,
20371 _mm_mask_cmpge_epu32_mask, _mm_mask_cmpge_epu64_mask,
20372 _mm_mask_cmple_epi32_mask, _mm_mask_cmple_epi64_mask,
20373 _mm_mask_cmple_epu32_mask, _mm_mask_cmple_epu64_mask,
20374 _mm_mask_cmplt_epi32_mask, _mm_mask_cmplt_epi64_mask,
20375 _mm_mask_cmplt_epu32_mask, _mm_mask_cmplt_epu64_mask,
20376 _mm_mask_cmpneq_epi32_mask, _mm_mask_cmpneq_epi64_mask,
20377 _mm_mask_cmpneq_epu32_mask, _mm_mask_cmpneq_epu64_mask): Move
20378 definitions outside of __OPTIMIZE__ guarded section.
20381 * config/i386/avx512bwintrin.h (_mm512_packs_epi32,
20382 _mm512_maskz_packs_epi32, _mm512_mask_packs_epi32,
20383 _mm512_packus_epi32, _mm512_maskz_packus_epi32,
20384 _mm512_mask_packus_epi32): Move definitions outside of __OPTIMIZE__
20387 2017-03-09 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
20389 * config/s390/vx-builtins.md ("vfee<mode>", "vfeez<mode>")
20390 ("vfenez<mode>"): Add missing constraints.
20392 2017-03-08 Martin Sebor <msebor@redhat.com>
20395 * config/nds32/nds32.c (nds32_option_override):
20396 Fix misspelled diagnostic.
20398 2017-03-08 Jakub Jelinek <jakub@redhat.com>
20401 * gimplify.c (gimplify_omp_for): Replace index var in outer
20402 taskloop statement with an artificial variable and add
20403 OMP_CLAUSE_PRIVATE clause for it.
20405 2017-03-08 Richard Biener <rguenther@suse.de>
20407 PR tree-optimization/79955
20408 * tree-ssa-uninit.c (warn_uninitialized_vars): Do not warn
20409 for accesses that are completely outside of the variable.
20411 2017-03-08 Andrew Haley <aph@redhat.com>
20413 PR tree-optimization/79943
20414 * tree-ssa-loop-split.c (compute_new_first_bound): When
20415 calculating the new upper bound, (END-BEG) should be added, not
20418 2017-03-08 Jakub Jelinek <jakub@redhat.com>
20420 * config/avr/avr.md (setmemhi): Make sure match_dup
20421 operand number comes before match_scratch.
20423 2017-03-08 Richard Biener <rguenther@suse.de>
20425 PR tree-optimization/79920
20426 * tree-vect-slp.c (vect_create_mask_and_perm): Remove and inline
20427 with ncopies == 1 to ...
20428 (vect_transform_slp_perm_load): ... here. Properly compute
20429 all element loads by iterating VF times over the group. Do
20430 not handle ncopies (computed in a broken way) in
20431 vect_create_mask_and_perm.
20433 2017-03-08 Jakub Jelinek <jakub@redhat.com>
20436 * internal-fn.c (expand_vector_ubsan_overflow): If arg0 or arg1
20437 is a uniform vector, use uniform_vector_p return value instead of
20438 building ARRAY_REF on folded VIEW_CONVERT_EXPR to array type.
20440 2017-03-07 Marek Polacek <polacek@redhat.com>
20442 PR middle-end/79809
20443 * gimple-ssa-warn-alloca.c (pass_walloca::gate): Use HOST_WIDE_INT.
20444 (alloca_call_type): Likewise.
20446 2017-03-07 Martin Liska <mliska@suse.cz>
20448 * gcov.c (process_args): Put comment to correct location.
20450 2017-03-07 Martin Liska <mliska@suse.cz>
20452 PR middle-end/68270
20453 * tree-chkp.c (chkp_may_narrow_to_field): Add new argument ref.
20454 Use array_at_struct_end_p instead of DECL_CHAIN (field).
20455 (chkp_narrow_bounds_for_field): Likewise.
20456 (chkp_parse_array_and_component_ref): Pass one more argument to
20459 2017-03-07 Richard Biener <rguenther@suse.de>
20461 * tree-vect-loop-manip.c (slpeel_add_loop_guard): Preserve
20464 2017-03-07 Segher Boessenkool <segher@kernel.crashing.org>
20466 * config/i386/i386.c (ix86_local_alignment): Align most aggregates
20467 of 16 bytes and more to 16 bytes, not those of 16 bits and more.
20469 2017-03-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20472 * params.def (PARAM_STORE_MERGING_ALLOW_UNALIGNED): Add full stop
20473 to end of description.
20474 (PARAM_MAX_STORES_TO_MERGE): Likewise.
20476 2017-03-07 Jakub Jelinek <jakub@redhat.com>
20478 PR rtl-optimization/79901
20479 * config/i386/sse.md (*avx512bw_<code><mode>3<mask_name>): Renamed to
20481 (*avx512f_<code><mode>3<mask_name>): ... this.
20482 (<code><mode>3 with maxmin code iterator): Use VI8_AVX2_AVX512F
20483 iterator instead of VI8_AVX2_AVX512BW.
20485 PR rtl-optimization/79901
20486 * expr.c (expand_expr_real_2): For vector MIN/MAX, if there is no
20487 min/max expander, expand it using expand_vec_cond_expr.
20490 * ubsan.c (ubsan_encode_value): Call mark_addressable on the
20493 2017-03-06 Jakub Jelinek <jakub@redhat.com>
20496 * dwarf2out.h (dw_vec_const): Change array type from unsigned char *
20497 to void * for PCH reasons.
20498 * dwarf2out.c (output_loc_operands, output_die): Cast
20499 v.val_vec.array to unsigned char *.
20501 2017-03-06 John David Anglin <danglin@gcc.gnu.org>
20504 * config/pa/pa-64.h (PAD_VARARGS_DOWN): Don't pad down complex and
20507 2017-03-06 Vladimir Makarov <vmakarov@redhat.com>
20509 PR rtl-optimization/79571
20510 * lra-constraints.c (process_alt_operands): Calculate static
20511 reject and subtract it from overall when only addresses will be
20514 2017-03-06 Julia Koval <julia.koval@intel.com>
20517 * config/i386/i386.c (ix86_minimum_incoming_stack_boundary): Set
20518 incoming stack boundary to 128 for 64-bit targets.
20520 2017-03-06 Richard Biener <rguenther@suse.de>
20522 PR tree-optimization/79894
20523 * tree-vectorizer.c (vectorize_loops): Set loop_vectorized_call
20524 to NULL after folding it.
20526 2017-03-06 Richard Biener <rguenther@suse.de>
20528 PR tree-optimization/79824
20529 * tree-vect-stmts.c (get_group_load_store_type): Fix alignment
20530 check disabling peeling for gaps.
20532 2017-03-06 Toma Tabacu <toma.tabacu@imgtec.com>
20534 * doc/sourcebuild.texi (Effective-Target Keywords, Environment
20535 attributes): Document gettimeofday.
20537 2017-03-06 Robin Dapp <rdapp@linux.vnet.ibm.com>
20539 * config/s390/s390.c (s390_option_override_internal): Set
20540 PARAM_MIN_VECT_LOOP_BOUND
20542 2017-03-06 Robin Dapp <rdapp@linux.vnet.ibm.com>
20544 * config/s390/s390.c (s390_asm_output_function_label): Use nopr %r0.
20545 * config/s390/s390.md: Likewise.
20547 2017-03-06 Jakub Jelinek <jakub@redhat.com>
20550 * config/i386/sse.md (VI8F_256_512): Remove mode iterator.
20551 (<avx2_avx512>_perm<mode>): Rename to ...
20552 (avx2_perm<mode>): ... this. Use VI8F_256 iterator instead
20554 (<avx512>_perm<mode>_mask): Rename to ...
20555 (avx512vl_perm<mode>_mask): ... this. Use VI8F_256 iterator instead
20557 (<avx2_avx512>_perm<mode>_1<mask_name>): Rename to ...
20558 (avx2_perm<mode>_1<mask_name): ... this. Use VI8F_256 iterator
20559 instead of VI8F_256_512.
20560 (avx512f_perm<mode>): New define_expand.
20561 (avx512f_perm<mode>_mask): Likewise.
20562 (avx512f_perm<mode>_1<mask_name>): New define_insn.
20563 (<avx512>_vec_dup<mode>_1): Fix up vec_select mode.
20565 2017-03-06 Prachi Godbole <prachi.godbole@imgtec.com>
20567 * config/mips/mips-msa.md (msa_fmax_a_<msafmt>, msa_fmin_a_<msafmt>,
20568 msa_max_a_<msafmt>, msa_min_a_<msafmt>): Introduce mode interator for
20570 (smin<mode>3, smax<mode>3): Change operand print code from 'B' to 'E'.
20572 2017-03-06 Martin Liska <mliska@suse.cz>
20575 * asan.c (asan_expand_poison_ifn): Do not expand ASAN_POISON
20576 when having a SSA NAME w/o VAR_DECL assigned to it.
20578 2017-03-06 Prachi Godbole <prachi.godbole@imgtec.com>
20580 * config/mips/mips-msa.md (msa_dotp_<su>_d, msa_dpadd_<su>_d,
20581 msa_dpsub_<su>_d): Fix MODE for vec_select.
20583 2017-03-06 Prachi Godbole <prachi.godbole@imgtec.com>
20585 * config/mips/mips.c (mips_gen_const_int_vector): Change type of last
20587 * config/mips/mips-protos.h (mips_gen_const_int_vector): Likewise.
20589 2017-03-06 Richard Biener <rguenther@suse.de>
20591 * lto-streamer.c (lto_check_version): Use %qs in diagnostics.
20592 * plugin.c (register_plugin_info): Likewise.
20593 * tree-chkp.c (chkp_make_static_const_bounds): Likewise.
20595 2017-03-05 Jakub Jelinek <jakub@redhat.com>
20597 * config/i386/sse.md (sse_storehps, sse_storelps,
20598 avx_<castmode><avxsizesuffix>_<castmode>,
20599 avx512f_<castmode><avxsizesuffix>_<castmode>,
20600 avx512f_<castmode><avxsizesuffix>_256<castmode>): Require
20601 in condition that at least one operand is not a MEM.
20603 2017-03-03 Jakub Jelinek <jakub@redhat.com>
20605 PR middle-end/79805
20606 * internal-fn.def (ATOMIC_BIT_TEST_AND_SET, ATOMIC_BIT_TEST_AND_RESET,
20607 ATOMIC_BIT_TEST_AND_COMPLEMENT, ATOMIC_COMPARE_EXCHANGE): Remove
20609 * gimple-fold.c (fold_builtin_atomic_compare_exchange): Set
20610 gimple_call_nothrow_p flag based on whether original builtin can throw.
20611 If it can, emit following stmts on the fallthrough edge.
20612 * tree-ssa-ccp.c (optimize_atomic_bit_test_and): Similarly, except
20613 don't create new bb if inserting just debug stmts on the edge, try to
20614 insert them on the fallthru bb or just reset debug stmts.
20616 2017-03-03 Segher Boesssenkool <segher@kernel.crashing.org>
20619 * config/rs6000/rs6000.c (rs6000_final_prescan_insn): Save and
20620 restore recog_data (including the operand rtxes inside it) around
20621 the call to get_insn_template.
20623 2017-03-03 Martin Sebor <msebor@redhat.com>
20625 PR tree-optimization/79699
20626 * context.c (context::~context): Free MPFR caches to avoid
20627 a memory leak on program exit.
20629 2017-03-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20631 * config/aarch64/aarch64.c (aarch64_float_const_representable_p):
20632 Use wide_int::ulow () instead of .elt (0).
20634 2017-03-03 Uros Bizjak <ubizjak@gmail.com>
20636 * config/i386/i386.md (*pushtf): Change *roF constraint to *roC.
20637 (*pushxf): Limit oF constraint to 32bit targets and add oC
20638 constraint for 64bit targets.
20639 (pushxf splitter): Use PUSH_ROUNDING to calculate stack adjustment.
20640 (*pushdf): Change rmF constraint to rmC.
20642 2017-03-03 Martin Liska <mliska@suse.cz>
20644 * tree-ssa-loop-prefetch.c (pass_loop_prefetch::execute):
20645 Remove unused variable.
20647 2017-03-03 Jakub Jelinek <jakub@redhat.com>
20650 * config/i386/i386.c (ix86_expand_multi_arg_builtin): If target
20651 is a memory operand, increase num_memory.
20652 (ix86_expand_args_builtin): Likewise.
20654 2017-03-03 Jan Hubicka <jh@suse.cz>
20657 * ipa-devirt.c (maybe_record_node): Properly handle
20658 __cxa_pure_virtual visibility.
20660 2017-03-03 Martin Liska <mliska@suse.cz>
20662 PR tree-optimization/79803
20663 * tree-ssa-loop-prefetch.c (tree_ssa_prefetch_arrays): Remove
20665 (pass_loop_prefetch::execute): Disabled optimization if an
20666 assumption about L1 cache size is not met.
20668 2017-03-03 Martin Liska <mliska@suse.cz>
20670 PR rtl-optimization/79574
20671 * gcse.c (struct gcse_expr): Use HOST_WIDE_INT instead of int.
20672 (hash_scan_set): Likewise.
20673 (dump_hash_table): Likewise.
20674 (hoist_code): Likewise.
20676 2017-03-03 Richard Biener <rguenther@suse.de>
20678 * fixed-value.c (fixed_from_string): Restore use of elt (1)
20679 in place of uhigh ().
20680 (fixed_convert_from_real): Likewise.
20682 2017-03-03 Uros Bizjak <ubizjak@gmail.com>
20685 * config/i386/i386.md (*pushxf_rounded): Use Pmode instead of DImode.
20687 2017-03-03 Richard Biener <rguenther@suse.de>
20689 PR middle-end/79818
20690 * match.pd ( X +- C1 CMP C2 -> X CMP C2 -+ C1): Add missing
20691 TYPE_OVERFLOW_UNDEFINED check.
20693 2017-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20695 * config/rs6000/vector.md (vector_ne_<mode>_p): Correct operand
20697 (vector_ae_<mode>_p): Likewise.
20698 (vector_nez_<mode>_p): Likewise.
20699 (vector_ne_v2di_p): Likewise.
20700 (vector_ae_v2di_p): Likewise.
20701 (vector_ne_<mode>_p): Likewise.
20702 * config/rs6000/vsx.md (vsx_tsqrt<mode>2_fg): Correct operand
20704 (vsx_tsqrt<mode>2_fe): Likewise.
20706 2017-03-02 Uros Bizjak <ubizjak@gmail.com>
20709 * config/i386/i386.md (*pushxf_rounded): New insn_and_split pattern.
20711 2017-03-02 Jakub Jelinek <jakub@redhat.com>
20713 PR rtl-optimization/79780
20714 * cprop.c (one_cprop_pass): When second and further conditional trap
20715 in a single basic block is turned into an unconditional trap, turn it
20716 into a deleted note to avoid RTL verification failures.
20718 2017-03-02 Richard Biener <rguenther@suse.de>
20720 * fold-const.c (const_binop): Use ulow () instead of elt (0).
20722 2017-03-02 Richard Biener <rguenther@suse.de>
20724 PR tree-optimization/79345
20726 * tree-ssa-alias.c (walk_aliased_vdefs_1): Take a limit
20727 param and abort the walk, returning -1 if it is hit.
20728 (walk_aliased_vdefs): Take a limit param and pass it on.
20729 * tree-ssa-alias.h (walk_aliased_vdefs): Add a limit param,
20730 defaulting to 0 and return a signed int.
20731 * tree-ssa-uninit.c (struct check_defs_data): New struct.
20732 (check_defs): New helper.
20733 (warn_uninitialized_vars): Use walk_aliased_vdefs to warn
20734 about uninitialized memory.
20735 * fixed-value.c (fixed_from_string): Use ulow/uhigh to avoid
20736 bogus uninitialized warning.
20737 (fixed_convert_from_real): Likewise.
20739 2017-03-02 Bin Cheng <bin.cheng@arm.com>
20741 PR tree-optimization/66768
20742 * tree-ssa-loop-ivopts.c (find_interesting_uses_address): Skip addr
20743 iv_use if base object can't be determined.
20745 2017-03-02 Jakub Jelinek <jakub@redhat.com>
20747 PR tree-optimization/79345
20748 * gensupport.h (struct pattern_stats): Add min_scratch_opno field.
20749 * gensupport.c (get_pattern_stats_1) <case MATCH_SCRATCH>: Update it.
20750 (get_pattern_stats): Initialize it.
20751 * genemit.c (gen_expand): Verify match_scratch numbers come after
20752 match_operand/match_dup numbers.
20753 * config/i386/i386.md (<s>mul<mode>3_highpart): Swap match_dup and
20754 match_scratch numbers.
20755 * config/i386/sse.md (avx2_gathersi<mode>, avx2_gatherdi<mode>):
20757 * config/s390/s390.md (trunctdsd2): Likewise.
20759 2017-03-02 Richard Biener <rguenther@suse.de>
20761 * wide-int.h (wide_int_storage::operator=): Implement in terms
20764 2017-03-02 Richard Biener <rguenther@suse.de>
20766 PR tree-optimization/79777
20767 * tree-ssa-pre.c (eliminate_insert): Give up if we simplify
20768 the to insert expression to sth existing.
20770 2017-03-01 Martin Sebor <msebor@redhat.com>
20772 PR middle-end/79692
20773 * gimple-ssa-sprintf.c
20774 (directive::known_width_and_precision): New function.
20775 (format_integer): Use it.
20776 (get_mpfr_format_length): Consider the full range of precision
20777 when computing %g output with the # flag. Set the likely byte
20778 count to 3 rather than 1 when precision is indeterminate.
20779 (format_floating): Correct the lower bound of precision.
20781 2017-03-01 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20783 * doc/invoke.texi: Document default code model for 64-bit Linux.
20785 2017-03-01 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
20788 * config/rs6000/rs6000.md (peephole2 for udiv/umod): Should emit
20789 udiv rather than div since input pattern is unsigned.
20791 2017-03-01 Uros Bizjak <ubizjak@gmail.com>
20793 * config/i386/i386.c (print_reg): Warn for values of
20794 unsupported size in integer register.
20796 2017-03-01 Michael Meissner <meissner@linux.vnet.ibm.com>
20799 * config/rs6000/predicates.md (current_file_function_operand): Do
20800 not allow self calls to be local if the function is replaceable.
20802 2017-03-01 Kelvin Nilsen <kelvin@gcc.gnu.org>
20805 * config/rs6000/altivec.h (vec_ctz and others): Change the
20806 preprocessor macro that controls conditional compilation from
20807 _ARCH_PWR9 to __POWER9_VECTOR__.
20808 (vec_all_ne): Change parameterization of __altivec_scalar_pred
20809 macro expansion under preprocessor #ifdef __POWER9_VECTOR__
20810 control (instead of _ARCH_PWR9 control) so that template
20811 definition uses power9-specific function.
20812 (vec_any_eq): Likewise.
20813 (vec_all_ne): Change macro definition to use a power9-specific
20814 expansion under #ifdef __POWER9_VECTOR__ control (instead of
20815 _ARCH_PWR9 control).
20816 (vec_any_eq) Likewise.
20817 * config/rs6000/rs6000-builtin.def (CMPNEF): Remove BU_P9V_AV_2
20818 expansion for CMPNEF to remove support for xvcmpnesp instruction.
20819 (CMPNED): Remove BU_P9V_AV2 expansion for CMPNED to remove
20820 support for xvcmpnedp instruction.
20821 (VCMPNEB_P): Replace BU_P9V_AV_P macro expansion with BU_P9V_AV_2
20822 macro expansion so that Power9 implementation of vec_all_ne does
20823 not use the AltiVec predicate framework.
20824 (VCMPNEH_P): Likewise.
20825 (VCMPNEW_P): Likewise.
20826 (VCMPNED_P): Likewise.
20827 (VCMPNEFP_P): Likewise.
20828 (VCMPNEDP_P): Likewise.
20829 (VCMPAEB_P): Add BU_P9V_AV_2 macro expansion to change
20830 implementation of vec_any_eq to not use AltiVec predicate
20832 (VCMPAEH_P): Likewise.
20833 (VCMPAEW_P): Likewise.
20834 (VCMPAED_P): Likewise.
20835 (VCMPAEFP_P): Likewise.
20836 (VCMPAEDP_P): Likewise.
20837 (VCMPNE_P): Replace BU_P9V_OVERLOAD_P macro expansion with
20838 BU_P9V_OVERLOAD_2 so that Power9 implementation of vec_all_ne does
20839 not use the AltiVec predicate framework.
20840 (VCMPAE_P): Add BU_P9V_OVERLOAD_2 macro to change implementation
20841 of vec_any_eq to not use AltiVec predicate framework.
20842 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Add
20843 support for predefined __POWER9_VECTOR__ macro to indicate that
20844 Power9 instruction selection is enabled.
20845 (altivec_overloaded_builtins): Remove extraneous
20846 ALTIVEC_BUILTIN_VEC_CMPNE entry for overloaded
20847 function argument types RS6000_BTI_bool_V16QI and
20848 RS6000_BTI_bool_V16QI. Remove erroneous ALTIVEC_BUILTIN_VEC_CMPNE
20849 entry for overloaded function argument types RS6000_BTI_bool_V4SI
20850 andRS6000_BTI_bool_V4SI, mapping to P9V_BUILTIN_CMPNEB. Remove
20851 two entries mapping to P9V_BUITIN_CMPNED and one entry mapping to
20852 P9V_BUILTIN_CMPNEF to force use of instructions not specific to
20853 Power9 for implementations of vec_cmpne. Change the signature for
20854 all definitions of the overloaded P9V_BUILTIN_VEC_CMPNE_P function
20855 (representing vec_all_ne) to remove the previously described first
20856 argument of type RS6000_BTI_INTSI, as this was an artifact of
20857 reliance on the AltiVec predicate framework, which is no longer
20858 used in the implementation of these functions. Add
20859 P9V_BUILTIN_VEC_VCMPAE_P entries (representing the vec_anyeq
20860 function) to match all of the P9V_BUILTIN_VEC_VCMNE_P entries
20861 since, unlike the AltiVec predicate framework implementation, we
20862 do not share function descriptors between vec_alle and vec_anyeq.
20863 (altivec_resolve_overloaded_builtin): Add SFmode and DFmode to the
20864 set of modes that receive special treatment even when
20865 TARGET_P9_VECTOR is true. The special treatment emits code that
20866 does not depend on Power9 instructions.
20867 * config/rs6000/vector.md (vector_ne_<mode>_p): Change this
20868 define_expand to not rely on AltiVec predicate framework.
20869 (vector_ae_<mode>p): New define_expand to represent vec_any_eq
20871 (vector_ne_v2di_p): Change this define_expand to not rely on
20872 AltiVec predicate framework.
20873 (vector_ae_v2di_p): New define_expand to represent vec_any_eq
20875 (vector_ne_<mode>_p): Change this define_expand to not rely on
20876 AltiVec predicate framework.
20877 (vector_ae_<mode>p): New define_expand to represent vec_any_eq
20879 * config/rs6000/vsx.md (*vsx_ne_<mode>_p): For modes VSX_EXTRACT_I
20880 (V16QI, V8HI, V4SI), correct a typo in the code emitted for this
20881 define_insn pattern.
20882 (*vsx_ne_<mode>_p): For modes VSX_F (V4SF and V2DF), remove this
20883 define_insn pattern because the xvcmpne<VSs>. instruction is not
20885 (vcmpne<VSs>): Remove this define_insn because xvcmpne<VSs>
20886 instruction is not supported.
20888 2017-03-01 Jakub Jelinek <jakub@redhat.com>
20890 * config/nvptx/nvptx.c: Include intl.h.
20892 2017-03-01 Martin Jambor <mjambor@suse.cz>
20895 * ipa-prop.h (ipa_bits): Removed field known.
20896 (ipa_jump_func): Removed field vr_known. Changed fields bits and m_vr
20897 to pointers. Adjusted their comments to warn about their sharing.
20898 (ipcp_transformation_summary): Change bits to a vector of pointers.
20899 (ipa_check_create_edge_args): Moved to ipa-prop.c, declare.
20900 (ipa_get_ipa_bits_for_value): Declare.
20901 * tree-vrp.h (value_range): Mark as GTY((for_user)).
20902 * ipa-prop.c (ipa_bit_ggc_hash_traits): New.
20903 (ipa_bits_hash_table): Likewise.
20904 (ipa_vr_ggc_hash_traits): Likewise.
20905 (ipa_vr_hash_table): Likewise.
20906 (ipa_print_node_jump_functions_for_edge): Adjust for bits and m_vr
20907 being pointers and vr_known being removed.
20908 (ipa_set_jf_unknown): Likewise.
20909 (ipa_get_ipa_bits_for_value): New function.
20910 (ipa_set_jfunc_bits): Likewise.
20911 (ipa_get_value_range): New overloaded functions.
20912 (ipa_set_jfunc_vr): Likewise.
20913 (ipa_compute_jump_functions_for_edge): Use the above functions to
20914 construct bits and vr parts of jump functions.
20915 (ipa_check_create_edge_args): Move here from ipa-prop.h, also allocate
20916 ipa_bits_hash_table and ipa_vr_hash_table if they do not already
20918 (ipcp_grow_transformations_if_necessary): Also allocate
20919 ipa_bits_hash_table and ipa_vr_hash_table if they do not already
20921 (ipa_node_params_t::duplicate): Do not copy bits, just pointers to
20922 them. Fix too long lines.
20923 (ipa_write_jump_function): Adjust for bits and m_vr being pointers and
20924 vr_known being removed.
20925 (ipa_read_jump_function): Use new setter functions to construct bits
20926 and vr parts of jump functions or set them to NULL.
20927 (write_ipcp_transformation_info): Adjust for bits being pointers.
20928 (read_ipcp_transformation_info): Likewise.
20929 (ipcp_update_bits): Likewise. Fix excessively long lines a trailing
20931 Include gt-ipa-prop.h.
20932 * ipa-cp.c (propagate_bits_across_jump_function): Adjust for bits
20934 (ipcp_store_bits_results): Likewise.
20935 (propagate_vr_across_jump_function): Adjust for m_vr being a pointer.
20936 Do not write to existing jump functions but use a temporary instead.
20938 2017-03-01 Jakub Jelinek <jakub@redhat.com>
20941 * fold-const.c (make_bit_field_ref): If orig_inner is COMPONENT_REF,
20942 attempt to use its first operand as BIT_FIELD_REF base.
20944 2017-03-01 Richard Biener <rguenther@suse.de>
20946 PR middle-end/79721
20947 * tree-chrec.c (chrec_evaluate): Perform computation of Newtons
20948 interpolating formula in wrapping arithmetic.
20949 (chrec_apply): Convert chrec_evaluate return value to wanted type.
20951 2017-03-01 Jakub Jelinek <jakub@redhat.com>
20953 PR tree-optimization/79734
20954 * tree-vect-generic.c (expand_vector_condition): Optimize
20955 AVX512 vector boolean VEC_COND_EXPRs into bitwise operations.
20956 Handle VEC_COND_EXPR where comparison has different inner width from
20957 type's inner width.
20959 2017-02-28 Sandra Loosemore <sandra@codesourcery.com>
20961 * doc/invoke.texi (ARC Options): Copy-edit to fix punctuation,
20962 markup, and similar issues. Remove @opindex entries for things
20963 that aren't options. Add missing -mmpy-option entries.
20965 2017-02-28 Jakub Jelinek <jakub@redhat.com>
20967 PR tree-optimization/79737
20968 * gimple-ssa-store-merging.c (encode_tree_to_bitpos): If bitlen is
20969 a multiple of BITS_PER_UNIT and !BYTES_BIG_ENDIAN, clear
20970 tmpbuf[byte_size - 1]. Call natice_encode_expr with byte_size - 1
20971 instead of byte_size. Formatting fix.
20972 (shift_bytes_in_array_right): Formatting fix.
20974 2017-02-28 Eric Botcazou <ebotcazou@adacore.com>
20977 * config/sparc/sparc.c (sparc_frame_pointer_required): Add missing
20978 condition on optimize for the leaf function test.
20980 2017-02-28 Martin Liska <mliska@suse.cz>
20983 * read-rtl-function.c (function_reader::handle_unknown_directive):
20984 Bail out when one uses -flto.
20986 2017-02-28 Martin Liska <mliska@suse.cz>
20988 * common.opt: Replace space with tabular for options of <number>
20990 * config/i386/i386.opt: Show <number> value for
20991 -mlarge-data-threshold.
20992 * opts.c (print_filtered_help): Do not display number in hexadecimal
20995 2017-02-28 Martin Liska <mliska@suse.cz>
20997 * common.opt: Fix --help=option -Q for options which are of
21000 2017-02-28 Uros Bizjak <ubizjak@gmail.com>
21002 * config/i386/i386.c (print_reg): Error out for values
21003 of 8-bit size in invalid integer register.
21005 2017-02-28 Martin Sebor <msebor@redhat.com>
21007 PR tree-optimization/79691
21008 * passes.def (pass_all_optimizations_g): Enable pass_sprintf_length.
21010 2017-02-28 Jakub Jelinek <jakub@redhat.com>
21013 * config/i386/i386.c (ix86_print_operand) <case 'R'>: Replace
21014 gcc_unreachable with output_operand_lossage.
21016 2017-02-28 Richard Biener <rguenther@suse.de>
21018 PR tree-optimization/79740
21019 * tree-ssa-sccvn.c (vn_nary_op_insert_into): Allow redundant
21021 (visit_nary_op): Insert the nary into the hashtable if we
21022 pattern-matched sth.
21023 * tree-ssa-pre.c (eliminate_insert): Robustify.
21025 2017-02-28 Richard Biener <rguenther@suse.de>
21027 PR middle-end/79731
21028 * fold-const.c (decode_field_reference): Reject out-of-bound
21031 2017-02-28 Jakub Jelinek <jakub@redhat.com>
21033 * config/i386/i386.c: Include intl.h.
21034 (ix86_option_override_internal): Use cond ? G_("...") : G_("...")
21035 instead of just cond ? "..." : "...".
21036 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims): Likewise.
21037 * coverage.c (read_counts_file): Likewise.
21038 * omp-offload.c: Include intl.h.
21039 (oacc_loop_fixed_partitions): Use cond ? G_("...") : G_("...") instead
21040 of just cond ? "..." : "...".
21041 * gcov.c (read_count_file): Use cond ? N_("...") : N_("...") instead
21042 of just cond ? "..." : "...".
21044 2017-02-28 Richard Earnshaw <rearnsha@arm.com>
21047 * config/arm/parsecpu.awk (gen_data): Set tuning target to 'tune for'
21049 * config/arm/arm-cpus.in (cortex-m0plus.small-multiply): Correct
21050 'tune for' CPU name.
21051 * config/arm/arm-cpu-data.h: Regenerated.
21053 2017-02-28 Richard Biener <rguenther@suse.de>
21055 PR tree-optimization/79732
21056 * tree-inline.c (expand_call_inline): Do not shadow var.
21058 2017-02-28 Richard Biener <rguenther@suse.de>
21060 PR tree-optimization/79723
21061 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Preserve
21062 address-space properly.
21064 2017-02-28 Thomas Schwinge <thomas@codesourcery.com>
21066 * doc/optinfo.texi (Optimization groups): Fix option used for
21068 * doc/invoke.texi (-fopt-info): Document "omp".
21069 * dumpfile.h: Sort OPTGROUP_OMP before OPTGROUP_VEC.
21070 (OPTGROUP_ALL): Add OPTGROUP_OMP.
21071 * hsa-gen.c (pass_data_gen_hsail): Use OPTGROUP_OMP.
21072 * ipa-hsa.c (pass_data_ipa_hsa): Likewise.
21073 * omp-simd-clone.c (pass_data_omp_simd_clone): Likewise.
21075 * dumpfile.h (OPTGROUP_OPENMP): Rename to OPTGROUP_OMP. Adjust
21077 * dumpfile.c (optgroup_options): Instead of "openmp", associate
21078 OPTGROUP_OMP with "omp".
21080 2017-02-27 Pat Haugen <pthaugen@us.ibm.com>
21083 * config/rs6000/rs6000-c.c (struct altivec_builtin_types): Use VSRAD
21084 for arithmetic shift of unsigned V2DI.
21086 2017-02-27 Claudiu Zissulescu <claziss@synopsys.com>
21088 * config.gcc (arc*-): Clean up, use arc/big.h, arc/elf.h, and
21089 arc/linux.h headers.
21090 * config/arc/arc.h (TARGET_OS_CPP_BUILTINS): Remove.
21091 (LINK_SPEC): Likewise.
21092 (ARC_TLS_EXTRA_START_SPEC): Likewise.
21093 (EXTRA_SPECS): Likewise.
21094 (STARTFILE_SPEC): Likewise.
21095 (ENDFILE_SPEC): Likewise.
21096 (LIB_SPEC): Likewise.
21097 (TARGET_SDATA_DEFAULT): Likewise.
21098 (TARGET_MMEDIUM_CALLS_DEFAULT): Likewise.
21099 (MULTILIB_DEFAULTS): Likewise.
21100 (DWARF2_UNWIND_INFO): Likewise.
21101 * config/arc/big.h: New file.
21102 * config/arc/elf.h: Likewise.
21103 * config/arc/linux.h: Likewise.
21104 * config/arc/t-uClibc: Remove.
21106 2017-02-27 Bin Cheng <bin.cheng@arm.com>
21108 PR tree-optimization/77536
21109 * tree-ssa-loop-manip.c (niter_for_unrolled_loop): New function.
21110 (tree_transform_and_unroll_loop): Use above function to compute the
21111 estimated niter of unrolled loop and use it when scaling profile.
21112 Also use count info rather than frequency if it's non-zero.
21113 * tree-ssa-loop-manip.h niter_for_unrolled_loop(): New declaration.
21114 * tree-vect-loop.c (scale_profile_for_vect_loop): New function.
21115 (vect_transform_loop): Call above function.
21117 2017-02-27 Richard Biener <rguenther@suse.de>
21119 PR tree-optimization/45397
21120 * tree-ssa-pre.c (eliminate_insert): Handle BIT_AND_EXPR.
21121 * tree-ssa-sccvn.c (valueized_wider_op): New helper.
21122 (visit_nary_op): Add pattern matching for CSEing sign-changed
21123 or truncated operations with wider ones.
21125 2017-02-27 Richard Biener <rguenther@suse.de>
21127 PR tree-optimization/79690
21128 * tree-vect-stmts.c (vectorizable_store): Use vector type
21129 built from the DR with address-space.
21131 2017-02-26 Gerald Pfeifer <gerald@pfeifer.com>
21133 * doc/invoke.texi (Optimize Options): Refine the description
21134 of asan-use-after-return.
21136 2017-02-25 Alan Modra <amodra@gmail.com>
21138 PR rtl-optimization/79584
21139 * lra-constraints.c (base_to_reg): Reload ad->base, the entire
21140 base, not ad->base_term, the reg within base. Remove assertion
21141 that ad->base == ad->base_term. Replace gen_int_mode using
21142 bogus mode with const0_rtx.
21144 2017-02-25 Jakub Jelinek <jakub@redhat.com>
21146 PR middle-end/79396
21147 * tree-eh.c (operation_could_trap_p, stmt_could_throw_1_p): Handle
21148 FMA_EXPR like tcc_binary or tcc_unary.
21150 * tree-ssa-loop-niter.c (number_of_iterations_exit): Simplify warning.
21153 * dwarf2out.c (struct dw_loc_list_struct): Add noted_variable_value
21155 (size_of_loc_descr): Handle DW_OP_GNU_variable_value.
21156 (output_loc_operands): Handle DW_OP_call_ref and
21157 DW_OP_GNU_variable_value.
21158 (struct variable_value_struct): New type.
21159 (struct variable_value_hasher): Likewise.
21160 (variable_value_hash): New variable.
21161 (string_types): Remove.
21162 (copy_loc_descr): New function.
21163 (add_loc_descr_to_each): Clarify comment. Use copy_loc_descr.
21164 (prepend_loc_descr_to_each): New function.
21165 (add_loc_list): Fix comment typo. Use prepend_loc_descr_to_each
21166 instead of add_loc_descr_to_each if the first argument is single
21167 location list and the second has multiple.
21168 (resolve_args_picking_1): Handle DW_OP_GNU_variable_value.
21169 (loc_list_from_tree_1): For early_dwarf, emit DW_OP_GNU_variable_value
21170 when looking for variable value which doesn't have other location info.
21171 (loc_list_from_tree): Formatting fix.
21172 (gen_array_type_die): Simplify DW_AT_string_length handling.
21173 (adjust_string_types): Remove.
21174 (gen_subprogram_die): Don't call adjust_string_types nor test/set
21175 string_types. Call resolve_variable_values.
21176 (prune_unused_types_walk_loc_descr): Handle DW_OP_GNU_variable_value.
21177 (resolve_addr_in_expr): Likewise. Add A argument.
21178 (copy_deref_exprloc): Remove deref argument. Adjust for the
21179 original expression being DW_OP_GNU_variable_value with optionally
21180 DW_OP_stack_value after it instead of DW_OP_call4 with DW_OP_deref
21181 optionally after it.
21182 (optimize_string_length): Rework for DW_OP_GNU_variable_value.
21183 (resolve_addr): Adjust optimize_string_length and resolve_addr_in_expr
21184 callers. Set remove_AT_byte_size if removing DW_AT_string_length.
21185 (variable_value_hasher::hash, variable_value_hasher::equal): New
21187 (resolve_variable_value_in_expr, resolve_variable_value,
21188 resolve_variable_values, note_variable_value_in_expr,
21189 note_variable_value): New functions.
21190 (dwarf2out_early_finish): Call note_variable_value on all toplevel
21193 2017-02-24 Jakub Jelinek <jakub@redhat.com>
21196 * opts.h (handle_generated_option): Add GENERATED_P argument.
21197 * opts-common.c (handle_option): Adjust function comment.
21198 (handle_generated_option): Add GENERATED_P argument, pass it to
21200 (control_warning_option): Pass false to handle_generated_option
21202 * opts.c (maybe_default_option): Pass true to handle_generated_option
21204 * optc-gen.awk: Likewise.
21206 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21208 * config/sh/sh.md (tstsi_t): If operands[0] is a SUBREG instead of
21209 a REG, look at the REG it is a SUBREG of.
21210 (splitter for cmpeqsi_t): Ditto.
21212 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21214 * config/pa/pa.c (pa_combine_instructions): Do not share RTL. Make
21215 the special USEs with the pattern of the insn, not the insn itself.
21217 2017-02-24 Matthew Fortune <matthew.fortune@imgtec.com>
21220 * doc/invoke.texi: Document -mload-store-pairs.
21222 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21223 Sandra Loosemore <sandra@codesourcery.com>
21225 * config/nios2/nios2.c (nios2_simple_const_p): Returns false if the
21226 argument isn't a CONST_INT.
21227 (nios2_alternate_compare_const): Assert op is a CONST_INT.
21228 (nios2_valid_compare_const_p): Assert op is a CONST_INT.
21229 (nios2_validate_compare): Bypass alternate compare logic if *op2
21230 is not a CONST_INT.
21231 (ldstwm_operation_p): Return false if first_base is not a REG or
21232 if first_offset is not a CONST_INT.
21234 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21236 * config/cris/cris.md: Use correct operand in a define_peephole2.
21238 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21240 * config/c6x/c6x.c (predicate_insn): Do not incorrectly share RTL.
21242 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
21244 * config/arc/arc.c (arc_ccfsm_advance): Only take the PATTERN of
21245 this_insn if it is an INSN or JUMP_INSN.
21246 (force_offsettable): Look at base, not at addr.
21247 * config/arc/predicates.md (brcc_nolimm_operator): Don't call INTVAL
21248 on things that aren't necessarily CONST_INTs.
21250 2017-02-24 Uros Bizjak <ubizjak@gmail.com>
21252 * doc/invoke.texi (x86 Options, -mfpmath=sse): Mention that
21253 -mfpmath=sse is the default also for x86-32 targets with SSE2
21254 instruction set when @option{-ffast-math} is enabled
21256 2017-02-24 Jeff Law <law@redhat.com>
21258 PR rtl-optimizatoin/79286
21259 * ira.c (update_equiv_regs): Drop may_trap_p exception to
21262 2017-02-24 Richard Biener <rguenther@suse.de>
21264 PR tree-optimization/79389
21265 * gimple-ssa-split-paths.c (is_feasible_trace): Properly skip
21268 2017-02-24 Aldy Hernandez <aldyh@redhat.com>
21270 * tree-ssa-loop-niter.c (number_of_iterations_exit): Update
21271 function comment to reflect reality.
21272 (loop_exits_before_overflow): Fix typo in function description.
21274 2017-02-24 Richard Biener <rguenther@suse.de>
21276 PR tree-optimization/79389
21277 * gimple-ssa-split-paths.c (is_feasible_trace): Verify more
21278 properly that a threading opportunity exists. Detect conditional
21279 copy/constant propagation opportunities.
21281 2017-02-23 Eric Botcazou <ebotcazou@adacore.com>
21283 * config/visium/visium.md (type): Add trap.
21284 (b): New mode attribute.
21285 (*btst): Rename into...
21286 (*btst<mode>): ...this and adjust.
21287 (*cbranchsi4_btst_insn): Rename into...
21288 (*cbranch<mode>4_btst_insn): ...this and adjust.
21289 (trap): New define_insn.
21291 2017-02-23 Jakub Jelinek <jakub@redhat.com>
21293 PR tree-optimization/79389
21294 * ifcvt.c (struct noce_if_info): Add rev_cond field.
21295 (noce_reversed_cond_code): New function.
21296 (noce_emit_store_flag): Use rev_cond if non-NULL instead of
21297 reversed_comparison_code. Formatting fix.
21298 (noce_try_store_flag): Test rev_cond != NULL in addition to
21299 reversed_comparison_code.
21300 (noce_try_store_flag_constants): Likewise.
21301 (noce_try_store_flag_mask): Likewise.
21302 (noce_try_addcc): Use rev_cond if non-NULL instead of
21303 reversed_comparison_code.
21304 (noce_try_cmove_arith): Likewise. Formatting fixes.
21305 (noce_try_minmax, noce_try_abs): Clear rev_cond.
21306 (noce_find_if_block): Initialize rev_cond.
21307 (find_cond_trap): Call noce_get_condition with then_bb == trap_bb
21308 instead of false as last argument never attempt to reverse it
21311 2017-02-23 Bin Cheng <bin.cheng@arm.com>
21313 PR tree-optimization/79663
21314 * tree-predcom.c (combine_chains): Process refs in reverse order
21315 only for ZERO length chains, and add explaining comment.
21317 2017-02-23 Jeff Law <law@redhat.com>
21319 PR tree-optimization/79578
21320 * tree-ssa-dse.c (clear_bytes_written_by): Use OEP_ADDRESS_OF
21321 in call to operand_equal_p.
21323 2017-01-23 Dominique d'Humieres <dominiq@lps.ens.fr>
21326 * config/i386/cpuid.h: Fix another undefined behavior.
21328 2017-02-23 Richard Biener <rguenther@suse.de>
21330 PR tree-optimization/79683
21331 * tree-vect-stmts.c (vect_analyze_stmt): Do not overwrite
21332 vector types for data-refs.
21334 2017-02-23 Martin Liska <mliska@suse.cz>
21336 * params.def (PARAM_MIN_NONDEBUG_INSN_UID): Change default to 0.
21338 2017-02-23 Jakub Jelinek <jakub@redhat.com>
21340 PR middle-end/79665
21341 * internal-fn.c (get_range_pos_neg): Moved to ...
21342 * tree.c (get_range_pos_neg): ... here. No longer static.
21343 * tree.h (get_range_pos_neg): New prototype.
21344 * expr.c (expand_expr_real_2) <case TRUNC_DIV_EXPR>: If both arguments
21345 are known to be in between 0 and signed maximum inclusive, try to
21346 expand both unsigned and signed divmod and use the cheaper one from
21349 2017-02-22 Jeff Law <law@redhat.com>
21351 PR tree-optimization/79578
21352 * tree-ssa-dse.c (clear_bytes_written_by): Use operand_equal_p
21353 to compare base operands.
21355 2017-02-22 Segher Boessenkool <segher@kernel.crashing.org>
21358 * config/rs6000/rs6000.md (*fsel<SFDF:mode><SFDF2:mode>4): Use
21359 gpc_reg_operand instead of fpr_reg_operand.
21361 2017-02-22 Sameera Deshpande <sameera.deshpande@imgtec.com>
21363 * config/mips/mips.c (mips_return_in_memory): Force FP
21364 vector types to be returned in memory for o32 ABI.
21366 2017-02-22 Jakub Jelinek <jakub@redhat.com>
21368 * dwarf2out.c (gen_variable_die): For -gdwarf-5, use DW_TAG_variable
21369 instead of DW_TAG_member for static data member declarations and don't
21370 set no_linkage_name for static inline data members.
21371 (gen_member_die): For -gdwarf-5 don't change DW_TAG_variable
21374 2017-02-22 Martin Liska <mliska@suse.cz>
21376 * doc/invoke.texi: Replace inequality signs with square brackets
21379 2017-02-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21381 PR tree-optimization/68644
21382 * gcc.dg/tree-ssa/ivopts-lt-2.c: Skip for powerpc*-*-*.
21384 2017-02-22 Matthew Fortune <matthew.fortune@imgtec.com>
21387 * lra-constraints.c (simplify_operand_subreg): Handle
21388 WORD_REGISTER_OPERATIONS targets.
21390 2017-02-22 Jakub Jelinek <jakub@redhat.com>
21393 * reg-stack.c (emit_swap_insn): Treat (float_extend:?F (mem:?F))
21394 and (const_double:?F) like (mem:?F) for the purpose of fxch %st(1)
21395 elimination by swapping fld*.
21397 2017-02-22 Richard Biener <rguenther@suse.de>
21399 PR tree-optimization/79673
21400 * tree-ssa-pre.c (compute_avail): Use wide_int_to_tree to
21401 convert the [TARGET_]MEM_REF offset INTEGER_CST, scrapping off
21402 irrelevant address-space qualifiers and avoiding a
21403 ADDR_SPACE_CONVERT_EXPR from fold_convert.
21405 2017-02-22 Richard Biener <rguenther@suse.de>
21407 PR tree-optimization/79666
21408 * tree-vrp.c (extract_range_from_binary_expr_1): Make sure
21409 to not symbolically negate if that may introduce undefined
21412 2017-02-22 Martin Liska <mliska@suse.cz>
21415 * data-streamer-in.c (streamer_read_gcov_count): Remove assert.
21416 * data-streamer-out.c (streamer_write_gcov_count_stream):
21418 * value-prof.c (stream_out_histogram_value): Make assert more
21419 precise based on type of counter.
21421 2017-02-21 Uros Bizjak <ubizjak@gmail.com>
21424 * config/i386/i386.md (standard_x87sse_constant_load splitter):
21425 Use nonimmediate_operand instead of memory_operand for operand 1.
21426 (float-extend standard_x87sse_constant_load splitter): Ditto.
21428 2017-02-21 Jeff Law <law@redhat.com>
21430 PR tree-optimization/79621
21431 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behavior): Ignore
21432 blocks with edges to themselves.
21434 2017-02-21 Jakub Jelinek <jakub@redhat.com>
21437 * tree-chkp-opt.c (chkp_optimize_string_function_calls): Use
21438 is_gimple_call instead of comparing gimple_code with GIMPLE_CALL.
21439 Use gimple_call_builtin_p.
21442 * sel-sched.c (moveup_expr_cached): Don't call sel_bb_head
21443 on temporarily removed DEBUG_INSNs.
21445 PR tree-optimization/79649
21446 * tree-loop-distribution.c (classify_partition): Give up on
21447 non-generic address space loads/stores.
21449 2017-02-21 Aldy Hernandez <aldyh@redhat.com>
21451 * doc/loop.texi (Loop manipulation): Remove nonexistent
21452 tree_ssa_loop_version from the documentation.
21453 * cfgloopmanip.c (loop_version): Document CONDITION_BB argument.
21455 2017-02-21 Jakub Jelinek <jakub@redhat.com>
21458 * config/i386/i386.c (ix86_expand_split_stack_prologue): Call
21459 make_reg_eh_region_note_nothrow_nononlocal on call_insn.
21460 * config/rs6000/rs6000.c: Include except.h.
21461 (rs6000_expand_split_stack_prologue): Call
21462 make_reg_eh_region_note_nothrow_nononlocal on the call insn.
21464 2017-02-21 Martin Jambor <mjambor@suse.cz>
21467 * ipa-prop.c (ipa_prop_write_jump_functions): Bail out if no edges
21468 have been analyzed.
21470 2017-02-21 Martin Jambor <mjambor@suse.cz>
21472 * common.opt (-fipa-cp-alignment): Mark as ignored and preserved
21473 for backward compatibility only.
21474 * doc/invoke.texi (Option Summary): Remove all references to
21475 -fipa-cp-alignment.
21477 2017-02-21 Matthew Fortune <matthew.fortune@imgtec.com>
21481 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
21483 * lra-constraints.c (curr_insn_transform): Handle
21484 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
21486 2017-02-21 Martin Liska <mliska@suse.cz>
21488 * config/i386/i386.opt: Replace -masm-dialect with -masm.
21490 2017-02-21 Thomas Schwinge <thomas@codesourcery.com>
21492 PR translation/79638
21493 * config/nvptx/nvptx.c (ENTRY_TEMPLATE): Single out "%ntid.y".
21495 2017-02-21 Eric Botcazou <ebotcazou@adacore.com>
21498 * config/arm/arm.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define.
21499 (arm_function_ok_for_sibcall): Return false for an indirect call by
21500 descriptor if all the argument registers are used.
21501 (arm_relayout_function): Use FUNCTION_ALIGNMENT macro to adjust the
21502 alignment of the function.
21504 2017-02-21 Jakub Jelinek <jakub@redhat.com>
21506 PR tree-optimization/61441
21507 * simplify-rtx.c (simplify_const_unary_operation): For
21508 -fsignaling-nans and sNaN operand, return NULL_RTX rather than
21509 the sNaN unmodified.
21511 2017-02-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
21513 * Makefile.in (BUILD_SYSTEM_HEADER_DIR): New make variabe.
21514 (LIMITS_H_TEST, if_multiarch, stmp-fixinc): Use BUILD_SYSTEM_HEADER_DIR
21515 instead of SYSTEM_HEADER_DIR.
21517 2017-02-20 Gerald Pfeifer <gerald@pfeifer.com>
21518 Martin Liška <mliska@suse.cz>
21520 * doc/invoke.texi (use-after-scope-direct-emission-threshold):
21521 Fix typos and grammar, use active voice, and clarify.
21523 2017-02-20 Marek Polacek <polacek@redhat.com>
21525 PR middle-end/79537
21526 * gimplify.c (gimplify_expr): Handle unused *&&L;.
21529 * ubsan.c (ubsan_type_descriptor): Check if TYPE_MAX_VALUE is null.
21531 2017-02-20 Jakub Jelinek <jakub@redhat.com>
21534 * config/i386/i386.c (ix86_expand_builtin): Handle
21535 OPTION_MASK_ISA_AVX512VL and OPTION_MASK_ISA_64BIT in
21536 ix86_builtins_isa[fcode].isa as a requirement of those
21537 flags and any other flag in the bitmask.
21538 (ix86_init_mmx_sse_builtins): Use 0 instead of
21539 ~OPTION_MASK_ISA_64BIT as mask.
21540 * config/i386/i386-builtin.def (__builtin_ia32_rdtsc,
21541 __builtin_ia32_rdtscp, __builtin_ia32_pause, __builtin_ia32_bsrsi,
21542 __builtin_ia32_rdpmc, __builtin_ia32_rolqi, __builtin_ia32_rolhi,
21543 __builtin_ia32_rorqi, __builtin_ia32_rorhi): Likewise.
21545 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
21548 * lra-constraints.c (split_reg): Check requested split mode
21549 is supported by the register.
21551 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
21553 * lra-constraints.c (simplify_operand_subreg): Remove early
21556 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
21559 * lra-constraints.c (curr_insn_transform): Tighten condition
21560 for converting SUBREG reloads from OP_OUT to OP_INOUT.
21562 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
21565 * lra-constraints.c (curr_insn_transform): Handle
21566 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
21568 2017-02-19 Uros Bizjak <ubizjak@gmail.com>
21571 2016-05-30 Uros Bizjak <ubizjak@gmail.com>
21573 * config/i386/sync.md (mfence_nosse): Use "lock orl $0, -4(%esp)".
21575 2017-02-19 Jonathan Wakely <jwakely@redhat.com>
21578 * doc/invoke.texi (C++ Dialect Options) [-Wliteral-suffix]: Update
21581 2017-02-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
21583 * gimple-pretty-print.c (dump_ternary_rhs): Adjust gimple dump format
21586 2017-02-18 Jakub Jelinek <jakub@redhat.com>
21588 * final.c (last_columnnum, override_columnnum): New variables.
21589 (final_start_function): Set last_columnnum, pass it to begin_prologue
21590 hook and pass 0 to dwarf2out_begin_prologue.
21591 (final_scan_insn): Update override_columnnum. Pass last_columnnum
21592 to source_line debug hook.
21593 (notice_source_line): Compute last_columnnum and for debug_column_info
21594 return true on column changes.
21595 * debug.h (struct gcc_debug_hooks): Add column argument to
21596 source_line and begin_prologue hooks.
21597 (debug_nothing_int_charstar_int_bool): Remove prototype.
21598 (debug_nothing_int_int_charstar,
21599 debug_nothing_int_int_charstar_int_bool): New prototypes.
21600 (dwarf2out_begin_prologue): Add column argument.
21601 * debug.c (do_nothing_debug_hooks): Adjust source_line and
21602 begin_prologue hooks.
21603 (debug_nothing_int_charstar_int_bool): Remove.
21604 (debug_nothing_int_int_charstar,
21605 debug_nothing_int_int_charstar_int_bool): New functions.
21606 * dwarf2out.c (dwarf2out_begin_prologue): Add column argument, pass it
21607 through to dwarf2out_source_line.
21608 (dwarf2_lineno_debug_hooks): Adjust begin_prologue hook.
21609 (dwarf2out_source_line): Add column argument, emit it if requested.
21610 * sdbout.c (sdbout_source_line, sdbout_begin_prologue): Add column
21612 * xcoffout.h (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
21613 * xcoffout.c (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
21614 * vmsdbgout.c (vmsdbgout_begin_prologue): Add column argument, pass it
21615 through to dwarf2out_begin_prologue.
21616 (vmsdbgout_source_line): Add column argument, pass it through to
21617 dwarf2out_source_line.
21618 * dbxout.c (dbxout_begin_prologue): Add column argument, adjust
21619 dbxout_source_line caller.
21620 (dbxout_source_line): Add column argument.
21622 * common.opt (gno-column-info, gcolumn-info): New options.
21623 * dwarf2out.c (dwarf2_lineno_debug_hooks): Formatting fix.
21624 (check_die): Also test for multiple DW_AT_decl_column attributes.
21625 (add_src_coords_attributes, dwarf2out_imported_module_or_decl_1): Add
21626 DW_AT_decl_column if requested.
21627 (gen_subprogram_die): Compare and/or add also DW_AT_decl_column
21629 (gen_variable_die): Likewise.
21630 (add_call_src_coords_attributes): Add DW_AT_call_column if requested.
21631 * doc/invoke.texi (-gcolumn-info, -gno-column-info): Document.
21634 * config/i386/i386.opt (m3dnowa): Replace Undocumented with Report.
21635 * common/config/i386/i386-common.c (OPTION_MASK_ISA_3DNOW_A_SET): Define.
21636 (ix86_handle_option): Handle OPT_m3dnowa.
21637 * doc/invoke.texi (-m3dnowa): Document.
21638 * doc/extend.texi (__builtin_ia32_pmulhuw, __builtin_ia32_pf2iw): Use
21639 -m3dnowa instead of -m3dnow -march=athlon.
21642 * config/i386/i386.c (ix86_print_operand): Use output_operand_lossage
21643 instead of gcc_assert for K, r and R code checks. Formatting fixes.
21645 2017-02-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21648 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21649 support for CODE_FOR_vsx_xxpermdi_v2d[fi]_be.
21650 * config/rs6000/rs6000.md (reload_gpr_from_vsx<mode>): Call
21651 generator for vsx_xxpermdi_<mode>_be.
21652 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Remove logic to
21653 force big-endian semantics.
21654 (vsx_xxpermdi_<mode>_be): New define_expand with same
21655 implementation as previous version of vsx_xxpermdi_<mode>.
21657 2017-02-17 Jakub Jelinek <jakub@redhat.com>
21659 PR tree-optimization/79327
21660 * gimple-ssa-sprintf.c (format_integer): Remove likely_adjust
21661 variable, its initialization and use.
21663 2017-02-17 Julia Koval <julia.koval@intel.com>
21665 * common/config/i386/i386-common.c (OPTION_MASK_ISA_RDPID_SET): New.
21666 (OPTION_MASK_ISA_PKU_UNSET): New.
21667 (ix86_handle_option): Handle -mrdpid.
21668 * config/i386/cpuid.h (bit_RDPID): New.
21669 * config/i386/driver-i386.c (host_detect_local_cpu):
21670 Detect RDPID feature.
21671 * config/i386/i386-builtin.def (__builtin_ia32_rdpid): New.
21672 * config/i386/i386-c.c (ix86_target_macros_internal):
21674 * config/i386/i386.c (ix86_target_string): Add -mrdpid to isa2_opts.
21675 (ix86_valid_target_attribute_inner_p): Add "rdpid".
21676 (ix86_expand_builtin): Handle IX86_BUILTIN_RDPID.
21677 * config/i386/i386.h (TARGET_RDPID, TARGET_RDPID_P): New.
21678 * config/i386/i386.md (define_insn "rdpid"): New.
21679 * config/i386/i386.opt Add -mrdpid.
21680 * config/i386/immintrin.h (_rdpid_u32): New.
21682 2017-02-17 Vladimir Makarov <vmakarov@redhat.com>
21684 PR rtl-optimization/79541
21685 * lra-constraints.c (curr_insn_transform): Remove wrong asm insn
21686 instead of transforming it into USE.
21688 2017-02-17 Segher Boessenkool <segher@kernel.crashing.org>
21690 * config/rs6000/rs6000.md (extendsfdf2): Remove default arguments.
21691 If HONOR_SNANS (SFmode) force the input to a register.
21692 (*extendsfdf2_fpr): Add !HONOR_SNANS (SFmode) condition.
21693 (*extendsfdf2_snan): New pattern, used when using SNaNs; it generates
21694 an frsp or similar insn.
21696 2017-02-17 Martin Liska <mliska@suse.cz>
21698 PR rtl-optimization/79577
21699 * params.def (selsched-max-sched-times): Increase minimum to 1.
21701 2017-02-17 Martin Liska <mliska@suse.cz>
21703 PR rtl-optimization/79574
21704 * gcse.c (want_to_gcse_p): Prevent integer overflow.
21706 2017-02-17 Martin Liska <mliska@suse.cz>
21708 PR tree-optimization/79529
21709 * tree-ssa-loop-unswitch.c (is_maybe_undefined): Use
21710 ssa_defined_default_def_p to handle cases which are implicitly
21712 * tree-ssa.c (ssa_defined_default_def_p): New function.
21713 (ssa_undefined_value_p): Use ssa_defined_default_def_p to handle cases
21714 which are implicitly defined.
21715 * tree-ssa.h (ssa_defined_default_def_p): Declare.
21717 2017-02-17 Richard Biener <rguenther@suse.de>
21719 PR middle-end/79576
21720 * params.def (max-ssa-name-query-depth): Limit to 10.
21722 2017-02-17 Richard Biener <rguenther@suse.de>
21724 PR tree-optimization/79552
21725 * tree-ssa-structalias.c (visit_loadstore): Properly verify
21728 2017-02-17 Richard Biener <rguenther@suse.de>
21731 * genmatch.c (output_line_directive): Handle DIR_SEPARATOR_2.
21733 2017-02-17 Marek Polacek <polacek@redhat.com>
21735 PR middle-end/79536
21736 * fold-const.c (fold_negate_expr_1): Renamed from fold_negate_expr.
21737 (fold_negate_expr): New wrapper.
21739 2017-02-16 Sandra Loosemore <sandra@codesourcery.com>
21741 * doc/invoke.texi (C++ Dialect Options) [-Wno-non-template-friend]:
21742 Correct terminology and de-emphasize pre-standard behavior.
21744 2017-02-16 Alan Modra <amodra@gmail.com>
21746 PR rtl-optimization/79286
21747 * ira.c (def_dominates_uses): New function.
21748 (update_equiv_regs): Don't create an equivalence for insns that
21749 may trap where the register def does not dominate the use.
21751 2017-02-16 Vladimir Makarov <vmakarov@redhat.com>
21753 PR rtl-optimization/78127
21754 * lra.c (lra): Call lra_eliminate before finish the loop after
21757 2017-02-16 Richard Biener <rguenther@suse.de>
21759 * graphite.h: Do not include isl/isl_val_gmp.h, instead include
21761 * graphite-isl-ast-to-gimple.c (gmp_cst_to_tree): Remove.
21762 (gcc_expression_from_isl_expr_int): Use generic isl_val interface.
21763 * graphite-sese-to-poly.c: Do not include isl/isl_val_gmp.h.
21764 (isl_val_int_from_wi): New function.
21765 (extract_affine_gmp): Rename to ...
21766 (extract_affine_wi): ... this, take a widest_int.
21767 (extract_affine_int): Just wrap extract_affine_wi.
21768 (add_param_constraints): Use isl_val_int_from_wi.
21769 (add_loop_constraints): Likewise, and extract_affine_wi.
21771 2017-02-15 Jeff Law <law@redhat.com>
21773 PR middle-end/79521
21774 * ira-costs.c (scan_one_insn): Check have_regs_of_mode before calling
21775 ira_init_register_move_cost_if_necessary.
21777 2017-02-15 Martin Sebor <msebor@redhat.com>
21779 PR middle-end/32003
21780 * doc/invoke.texi (-fdump-final-insns): Replace option accidentally
21781 removed in a prior commit.
21783 2017-02-15 Bin Cheng <bin.cheng@arm.com>
21785 PR tree-optimization/79347
21786 * tree-vect-loop-manip.c (vect_do_peeling): Maintain profile
21787 counters during peeling.
21789 2017-02-15 Thomas Schwinge <thomas@codesourcery.com>
21791 * Makefile.in (site.exp): Remove "set ISLVER".
21793 2017-02-15 Jakub Jelinek <jakub@redhat.com>
21796 * real.c (real_from_integer): Call real_convert even for decimal.
21798 2017-02-15 Dominik Vogt <vogt@linux.vnet.ibm.com>
21801 * config/s390/s390.c: define TARGET_CUSTOM_FUNCTION_DESCRIPTORS.
21803 2017-02-14 Andrew Pinski <apinski@cavium.com>
21805 * config/aarch64/aarch64-cores.def (thunderx2t99): Move to under 'C"
21806 cores and change the partno/implementer to be correct.
21807 (thunderx2t99p1): New core which replaces thunderx2t99 and still has
21808 the 'B" as the implementer.
21809 * config/aarch64/aarch64-tune.md: Regenerate.
21811 2017-02-14 Carl Love <cel@us.ibm.com>
21813 * config/rs6000/rs6000.c: Add case statement entry to make the
21814 xvcvuxdsp built-in argument unsigned.
21815 * config/rs6000/vsx.md: Fix the source and return operand types so they
21816 match the instruction definitions from the ISA document. Fix typo
21817 in the instruction generation for the (define_insn "vsx_xvcvuxdsp"
21820 2017-02-14 Vladimir Makarov <vmakarov@redhat.com>
21823 * lra-int.h (struct lra_operand_data, struct lra_insn_reg): Add
21824 member early_clobber_alts.
21825 * lra-lives.c (reg_early_clobber_p): New.
21826 (process_bb_lives): Use it.
21827 * lra.c (new_insn_reg): New arg early_clobber_alts. Use it.
21828 (debug_operand_data): Initialize early_clobber_alts.
21829 (setup_operand_alternative): Set up early_clobber_alts.
21830 (collect_non_operand_hard_regs): Ditto. Pass early clobber
21831 alternatives to new_insn_reg.
21832 (add_regs_to_insn_regno_info): Add arg early_clobber_alts. Use
21834 (lra_update_insn_regno_info): Pass the new arg.
21836 2017-02-14 Jakub Jelinek <jakub@redhat.com>
21838 PR middle-end/79505
21839 * omp-offload.c (free_oacc_loop): Release loop->ifns vector.
21840 (new_oacc_loop_raw): Don't clear already cleared fields.
21843 * config/i386/avx512pfintrin.h (_mm512_prefetch_i32gather_pd,
21844 _mm512_prefetch_i32gather_ps, _mm512_prefetch_i64gather_pd,
21845 _mm512_prefetch_i64gather_ps): New inline functions and macros.
21847 2017-02-14 Uros Bizjak <ubizjak@gmail.com>
21850 * config/i386/i386.md (*movxf_internal): Add (o,rC) alternative.
21852 2017-02-14 H.J. Lu <hongjiu.lu@intel.com>
21855 * config/i386/i386.c (timode_scalar_chain::convert_insn): Insert
21856 the extra instruction to the right place to store 128-bit constant
21859 2017-02-14 Martin Sebor <msebor@redhat.com>
21861 PR middle-end/79448
21862 * gimple-ssa-sprintf.c (format_directive): Avoid issuing INT_MAX
21863 warning for strings of unknown length.
21865 2017-02-13 Segher Boessenkool <segher@kernel.crashing.org>
21867 * config.gcc (supported_defaults) [powerpc*-*-*]: Update.
21869 2017-02-14 Jeff Law <law@redhat.com>
21872 * ira-costs.c (scan_one_insn): Initialize register move costs
21873 for pseudos seen in USE/CLOBBER insns.
21875 PR tree-optimization/79095
21876 * tree-vrp.c (extract_range_from_binary_expr_1): For EXACT_DIV_EXPR,
21877 if the numerator has the range ~[0,0] make the resultant range ~[0,0].
21878 (extract_range_from_binary_expr): For MINUS_EXPR with no derived range,
21879 if the operands are known to be not equal, then the resulting range
21881 (intersect_ranges): If the new range is ~[0,0] and the old range is
21882 wide, then prefer ~[0,0].
21883 * tree-vrp.c (overflow_comparison_p_1): New function.
21884 (overflow_comparison_p): New function.
21885 * tree-vrp.c (register_edge_assert_for_2): Register additional asserts
21886 if NAME is used in an overflow test.
21887 (vrp_evaluate_conditional_warnv_with_ops): If the ops represent an
21888 overflow check that can be expressed as an equality test, then adjust
21889 ops to be that equality test.
21891 2017-02-14 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
21893 * config/s390/s390-builtin-types.def: Remove flags argument.
21894 * config/s390/s390.c (s390_init_builtins): Likewise.
21896 2017-02-14 Martin Liska <mliska@suse.cz>
21898 * tree-ssa-loop-unswitch.c (hoist_guard): Release get_loop_body
21899 vector. Fix trailing white spaces.
21901 2017-02-14 James Greenhalgh <james.greenhalgh@arm.com>
21903 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Handle
21906 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21908 PR rtl-optimization/68664
21909 * config/arm/arm.c (arm_sched_can_speculate_insn):
21910 New function. Declare prototype.
21911 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
21913 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21915 PR rtl-optimization/68664
21916 * config/aarch64/aarch64.c (aarch64_sched_can_speculate_insn):
21918 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
21920 2017-02-14 Amit Pawar <amit.pawar@amd.com>
21922 * config/i386/i386.c (znver1_cost): Fix the alignment for function and
21923 max skip bytes for function, loop and jump.
21925 2017-02-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
21927 * gimple-pretty-print.c (dump_unary_rhs): Adjust dump format for
21928 ABS_EXPR for gimple dump.
21930 2017-02-14 Jakub Jelinek <jakub@redhat.com>
21933 * config/sh/sh.c (expand_cbranchdi4): Don't clear operands[4].
21935 PR tree-optimization/79408
21936 * tree-vrp.c (simplify_div_or_mod_using_ranges): Handle also the
21937 case when on TRUNC_MOD_EXPR op0 is INTEGER_CST.
21938 (simplify_stmt_using_ranges): Call simplify_div_or_mod_using_ranges
21939 also if rhs1 is INTEGER_CST.
21941 2017-02-14 Richard Biener <rguenther@suse.de>
21943 PR middle-end/79432
21944 * tree-into-ssa.c (insert_phi_nodes): When the function can
21945 have abnormal edges rewrite SSA names with broken use-def
21946 dominance out of SSA and register them for PHI insertion.
21948 2017-02-13 Martin Sebor <msebor@redhat.com>
21950 PR middle-end/79496
21951 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call): Avoid
21952 clearing info.nowrite flag when snprintf size argument is a range.
21954 2017-02-13 Jakub Jelinek <jakub@redhat.com>
21956 * cprop.c (cprop_jump): Add missing space in string literal.
21957 * tree-ssa-structalias.c (rewrite_constraints): Likewise.
21958 (get_constraint_for_component_ref): Likewise.
21959 * df-core.c (df_worklist_dataflow_doublequeue): Likewise.
21960 * tree-outof-ssa.c (insert_partition_copy_on_edge): Likewise.
21961 * lra-constraints.c (process_alt_operands): Likewise.
21962 * ipa-inline.c (inline_small_functions): Likewise.
21963 * tree-ssa-sccvn.c (visit_reference_op_store): Likewise.
21964 * cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
21965 * trans-mem.c (diagnose_tm_1_op): Likewise.
21966 * omp-grid.c (grid_find_single_omp_among_assignments): Likewise.
21967 (grid_parallel_clauses_gridifiable): Likewise.
21969 * config/nvptx/mkoffload.c (process): Add space in between
21972 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
21973 "MOD4_SSE_REGS" and "ALL_REGS".
21975 * spellcheck.c (test_data): Add , in between "foo" and "food".
21977 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
21980 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
21981 boundary crossing check and subsequent code generation agree.
21983 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21985 * config/aarch64/aarch64.c (has_memory_op): Delete.
21986 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
21989 2017-02-13 Jakub Jelinek <jakub@redhat.com>
21991 PR rtl-optimization/79388
21992 PR rtl-optimization/79450
21993 * combine.c (distribute_notes): When removing TEM_INSN for which
21994 corresponding dest has last value recorded, invalidate that last
21997 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21999 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
22000 of explicit '@'. Add missing assembly comment marker on branch costs
22003 2017-02-13 Nathan Sidwell <nathan@acm.org>
22005 * gengtype-lex.l (<in_struct>): Add '/'.
22007 2017-02-13 Martin Liska <mliska@suse.cz>
22010 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
22012 2017-02-13 Richard Biener <rguenther@suse.de>
22014 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
22016 * configure: Re-generate.
22017 * config.in: Likewise.
22018 * graphite-dependences.c: Simplify as if
22019 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
22020 * graphite-isl-ast-to-gimple.c: Likewise.
22021 * graphite-optimize-isl.c: Likewise.
22022 * graphite-poly.c: Likewise.
22023 * graphite-sese-to-poly.c: Likewise.
22024 * graphite.h: Likewise.
22025 * toplev.c: Include isl/version.h and use isl_version () for
22026 printing the ISL version.
22027 * doc/install.texi: Update ISL requirement.
22029 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
22031 * doc/standards.texi (Standards): Update reference to
22034 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
22036 * doc/extend.texi (Named Address Spaces): sourceware.org now
22038 * doc/install.texi (Binaries): Ditto.
22041 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
22043 * doc/cpp.texi: Replace "stringify"/"stringification" with C
22044 standard terminology "stringize"/"stringizing" throughout.
22045 * doc/cppinternals.texi: Likewise.
22047 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
22049 * doc/extend.texi: Fix some spelling mistakes and typos.
22050 * doc/invoke.texi: Likewise.
22052 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
22055 * params.def (inline-min-speedup) Change from 10 to 8.
22057 2017-02-11 Jakub Jelinek <jakub@redhat.com>
22059 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
22062 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
22065 * ipa-inline-analysis.c (get_minimal_bb): New function.
22066 (record_modified): Use it.
22067 (remap_edge_change_prob): Handle also ancestor functions.
22069 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
22071 * doc/contrib.texi (Contributors): Remove broken link into
22072 the Mauve CVS repository.
22074 2017-02-11 Jakub Jelinek <jakub@redhat.com>
22076 PR middle-end/79454
22077 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
22078 result computation whenever lhs doesn't have vector mode, not
22079 just when it has BLKmode.
22081 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
22083 * doc/makefile.texi (profiledbootstrap): Refer to the
22084 installation instructions only in textual form.
22086 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
22089 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
22091 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
22093 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
22094 (Specific): Update mingw-w64 reference.
22096 (Specific): Remove broken link to Renesas RX processor.
22098 2017-02-10 Richard Biener <rguenther@suse.de>
22100 * toplev.c (process_options): Do not mention obsolete graphite
22101 options when printing sorry message about missing graphite support.
22102 Mention -floop-nest-optimize.
22104 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
22106 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
22107 (vtst_p16): Likewise.
22108 (vtstq_p8): Likewise.
22109 (vtstq_p16): Likewise.
22111 (vtstq_p64): Likewise.
22112 * config/arm/arm_neon.h (vgetq_lane_p64): New.
22113 (vset_lane_p64): New.
22114 (vsetq_lane_p64): New.
22116 2017-02-10 Jakub Jelinek <jakub@redhat.com>
22118 PR tree-optimization/79411
22119 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
22120 stmt operands are SSA_NAMEs used in abnormal phis.
22121 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
22124 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
22127 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
22130 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
22132 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
22134 2017-02-09 Jakub Jelinek <jakub@redhat.com>
22136 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
22140 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
22141 not arbitrary TREE_CONSTANT.
22144 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
22145 "omp declare target link" attribute unless is_global_var.
22146 * omp-offload.c (find_link_var_op): Likewise.
22148 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
22149 Chung-Lin Tang <cltang@codesourcery.com>
22151 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
22153 (gimplify_adjust_omp_clauses): Don't delete TILE.
22154 (gimplify_omp_for): Deal with TILE.
22155 * internal-fn.c (expand_GOACC_TILE): New function.
22156 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
22158 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
22159 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
22161 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
22162 avoid DIV for outermost collapse var.
22163 (expand_oacc_for): Insert tile element loop as needed. Adjust.
22164 Remove out of date comments, fix whitespace.
22165 * omp-general.c (omp_extract_for_data): Deal with tiling.
22166 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
22167 adjust OLF_DIM_BASE value.
22168 (struct omp_for_data): Add tiling field.
22169 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
22170 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
22171 for auto loops. Remove default auto determining, moved to
22172 oacc_loop_fixed_partitions.
22173 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
22174 stmts, add e_mask field.
22175 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
22176 (oacc_thread_numbers): Use oacc_dim_call.
22177 (oacc_xform_tile): New.
22178 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
22179 (finish_oacc_loop): Adjust for ifns vector.
22180 (oacc_loop_discover_walk): Append loop abstraction sites to list,
22181 add case for GOACC_TILE fns.
22182 (oacc_loop_xform_loop): Delete.
22183 (oacc_loop_process): Iterate over call list directly, and add
22184 handling for GOACC_TILE fns.
22185 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
22187 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
22188 vector partitioning to outer loops. Assign 2 partitions to loops
22189 when available. Add TILE handling.
22190 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
22191 (execite_oacc_device_lower): Process GOACC_TILE fns,
22192 ignore unknown specs.
22193 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
22194 * tree.c (omp_clause_num_ops): Adjust TILE ops.
22195 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
22197 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
22199 * configure.ac (ACX_BUGURL): Update.
22200 * configure: Regenerate.
22202 2017-02-09 Richard Biener <rguenther@suse.de>
22204 PR tree-optimization/69823
22205 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
22206 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
22208 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
22210 * config/arc/arc-c.def: Add __NPS400__ definition.
22211 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
22212 (TARGET_NPS400): Define.
22214 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
22216 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
22218 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
22219 pointer, arch_info.
22220 (arc_cpu_types): Fill the arch_info field with a pointer into the
22221 arc_arch_types table.
22222 (arc_selected_cpu): Declare.
22223 * config/arc/arc.c (arc_selected_cpu): Make global.
22224 (arc_selected_arch): Delete.
22225 (arc_base_cpu): Delete.
22226 (arc_override_options): Remove references to deleted variables,
22227 update access to arch information.
22228 (ARC_OPT): Update access to arch information.
22229 (ARC_OPTX): Likewise.
22230 * config/arc/arc.h (arc_base_cpu): Remove declaration.
22231 (TARGET_ARC600): Update access to arch information.
22232 (TARGET_ARC601): Likewise.
22233 (TARGET_ARC700): Likewise.
22234 (TARGET_EM): Likewise.
22235 (TARGET_HS): Likewise.
22236 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
22239 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
22242 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
22243 condition/operands for integer GE/LE/GEU/LEU operations.
22245 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
22247 PR translation/79397
22248 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
22251 2017-02-08 Martin Jambor <mjambor@suse.cz>
22254 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
22255 whether allocation happened.
22256 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
22257 nothing was allocated.
22259 2017-02-08 Jakub Jelinek <jakub@redhat.com>
22261 PR tree-optimization/79408
22262 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
22263 constant, but SSA_NAME with a known integer range, use the minimum
22264 of that range instead of op1 to determine if modulo can be replaced
22265 with its first operand.
22267 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22269 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
22271 2017-02-08 Richard Biener <rguenther@suse.de>
22273 PR tree-optimization/71824
22274 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
22275 Check all loops contained in the merged region.
22277 2017-02-07 Andrew Pinski <apinski@cavium.com>
22279 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
22281 2017-02-07 Andrew Pinski <apinski@cavium.com>
22283 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
22284 (thunderxt88): Likewise.
22285 (thunderxt81): Disable LSE and change v8.1 to v8.
22286 (thunderxt83): Likewise.
22288 2017-02-07 Jakub Jelinek <jakub@redhat.com>
22289 Richard Biener <rguenther@suse.de>
22291 PR middle-end/79399
22292 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
22293 type from int to size_t.
22294 * ira-costs.c (struct_costs_size): Change type from int to size_t.
22296 2017-02-07 Jakub Jelinek <jakub@redhat.com>
22298 PR rtl-optimization/79386
22299 * cprop.c (bypass_conditional_jumps): Initialize
22300 bypass_last_basic_block already before splitting bbs after
22301 unconditional traps...
22302 (bypass_conditional_jumps): ... rather than here.
22305 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
22306 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
22307 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
22308 fix -masm=intel patterns.
22310 2017-02-07 Richard Biener <rguenther@suse.de>
22312 PR tree-optimization/79256
22313 PR middle-end/79278
22314 * builtins.c (get_object_alignment_2): Use min_align_of_type
22315 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
22316 and ADJUST_FIELD_ALIGN.
22318 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
22320 * doc/tm.texi: Regenerate.
22321 * stor-layout.c (layout_decl): Adjust.
22322 (update_alignment_for_field): Likewise.
22323 (place_field): Likewise.
22324 (min_align_of_type): Likewise.
22325 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
22326 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
22327 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
22328 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
22329 * config/frv/frv.c (frv_adjust_field_align): Likewise.
22330 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
22331 * config/i386/i386.c (x86_field_alignment): Likewise.
22332 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
22333 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
22334 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
22335 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
22336 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
22337 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
22341 2017-01-30 Richard Biener <rguenther@suse.de>
22343 PR tree-optimization/79256
22344 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
22345 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
22348 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
22350 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
22351 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
22352 builtins to SImode and emit a zero-extend, if necessary.
22354 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
22356 * docs/invoke.texi (RISC-V Options): Alphabetize.
22358 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
22360 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
22363 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
22365 * config/riscv/riscv.c: New file.
22366 * common/config/riscv/riscv-common.c: Likewise.
22367 * config.gcc: Likewise.
22368 * config/riscv/constraints.md: Likewise.
22369 * config/riscv/elf.h: Likewise.
22370 * config/riscv/generic.md: Likewise.
22371 * config/riscv/linux.h: Likewise.
22372 * config/riscv/multilib-generator: Likewise.
22373 * config/riscv/peephole.md: Likewise.
22374 * config/riscv/pic.md: Likewise.
22375 * config/riscv/predicates.md: Likewise.
22376 * config/riscv/riscv-builtins.c: Likewise.
22377 * config/riscv/riscv-c.c: Likewise.
22378 * config/riscv/riscv-ftypes.def: Likewise.
22379 * config/riscv/riscv-modes.def: Likewise.
22380 * config/riscv/riscv-opts.h: Likewise.
22381 * config/riscv/riscv-protos.h: Likewise.
22382 * config/riscv/riscv.h: Likewise.
22383 * config/riscv/riscv.md: Likewise.
22384 * config/riscv/riscv.opt: Likewise.
22385 * config/riscv/sync.md: Likewise.
22386 * config/riscv/t-elf-multilib: Likewise.
22387 * config/riscv/t-linux: Likewise.
22388 * config/riscv/t-linux-multilib: Likewise.
22389 * config/riscv/t-riscv: Likewise.
22390 * configure.ac: Likewise.
22391 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
22392 Waterman as RISC-V maintainers.
22393 * doc/install.texi: Add RISC-V entries.
22394 * doc/invoke.texi: Add RISC-V options section.
22395 * doc/md.texi: Add RISC-V constraints section.
22396 * configure: Regenerated.
22398 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
22401 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
22402 false values to be constant vectors with all 0 or all 1 bits set.
22403 (vcondu<mode><mode>): Likewise.
22404 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
22406 (fpmask_comparison_operator): Update comment.
22407 (vecint_comparison_operator): New predicate.
22408 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
22409 vector conditionals when the true and false values are constant
22410 vectors with all 0 bits or all 1 bits set.
22412 2017-02-06 Martin Sebor <msebor@redhat.com>
22414 PR tree-optimization/79376
22415 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
22417 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
22419 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
22420 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
22421 to simplify split condition.
22423 2017-02-06 Jakub Jelinek <jakub@redhat.com>
22425 * omp-expand.c (oxpand_omp_atomic_fetch_op,
22426 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
22429 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
22431 PR rtl-optimization/68664
22432 * target.def (can_speculate_insn): New hook.
22433 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
22434 * doc/tm.texi: Regenerate.
22435 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
22436 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
22437 (rs6000_sched_can_speculate_insn): New function.
22439 2017-02-06 Jakub Jelinek <jakub@redhat.com>
22441 PR tree-optimization/79284
22442 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
22443 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
22444 vectorizable_mask_load_store, vectorizable_operation,
22445 vect_is_simple_cond, get_same_sized_vectype): Use it instead
22446 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
22447 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
22448 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
22449 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
22450 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
22451 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
22452 is_gimple_assign (stmt). Replace another such test with
22453 is_gimple_assign (stmt).
22455 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
22458 * config/avr/avr.c (rtl-iter.h): Include it.
22459 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
22460 (avr_legitimate_combined_insn): ...and implementation.
22462 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
22464 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
22465 * config/s390/s390.c (s390_const_operand_ok)
22466 (s390_canonicalize_comparison, s390_extract_part)
22467 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
22468 (s390_contiguous_bitmask_p, s390_rtx_costs)
22469 (legitimize_pic_address): Likewise.
22470 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
22471 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
22472 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
22473 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
22474 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
22476 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
22478 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
22479 REGNO($0) == REGNO($1).
22481 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
22483 * config/s390/linux.h(SIZE_TYPE): Add comment.
22485 2017-02-06 Julian Brown <julian@codesourcery.com>
22486 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
22487 Virendra Pathak <virendra.pathak@broadcom.com>
22489 * config/aarch64/aarch64-cores.def: Change the scheduler
22491 * config/aarch64/aarch64.md: Include thunderx2t99.md.
22492 * config/aarch64/thunderx2t99.md: New file.
22494 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
22496 * doc/standards.texi (Go Language): Update link to language
22499 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
22501 * tree-eh.c (lower_resx): Sanitize profile.
22502 (cleanup_empty_eh_move_lp): Likewise.
22504 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
22507 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
22509 * cfgloopmanip.h (loop_version): Update prototype.
22510 * modulo-sched.c (sms_schedule): Update call of loop_version.
22511 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
22512 * tree-parloops.c (gen_parallel_loop): Likewise.
22513 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
22514 * tree-ssa-loop-split.c (split_loop): Likewise.
22515 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
22516 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
22518 2017-02-05 Martin Liska <mliska@suse.cz>
22521 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
22523 (print_operand_address): Initialize a struct to zero.
22525 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
22527 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
22528 garbage collector only in textual form.
22530 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
22532 * doc/extend.texi (x86 specific memory model extensions for
22533 transactional memory): Simplify a phrase.
22535 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
22538 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
22539 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
22540 (atomic_storedi_1): Likewise.
22542 2017-02-04 Jakub Jelinek <jakub@redhat.com>
22544 PR tree-optimization/79338
22545 * tree-parloops.c (gather_scalar_reductions): Don't call
22546 vect_analyze_loop_form for loop->inner before destroying loop's
22549 2017-02-03 Martin Sebor <msebor@redhat.com>
22551 PR tree-optimization/79327
22552 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
22553 when precision has resulted in leading zeros.
22554 (format_integer): Adjust the likely counter to assume an unknown
22555 argument that may be zero is non-zero.
22557 2017-02-03 Jason Merrill <jason@redhat.com>
22560 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
22561 avoid copying non-taken branch.
22563 2017-02-03 Jakub Jelinek <jakub@redhat.com>
22565 PR tree-optimization/79340
22566 * tree-vect-loop.c (vectorizable_reduction): Release
22567 vec_defs elements after safe_splicing them into other vectors.
22570 PR tree-optimization/79327
22571 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
22572 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
22574 (format_integer): Use wide_int_to_tree instead of build_int_cst
22575 + to_?hwi. If argmin is NULL, just set argmin and argmax to
22576 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
22577 of shortest and longest sequence.
22579 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
22581 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
22582 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
22584 2017-02-03 Walter Lee <walt@tilera.com>
22587 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
22588 after initial stackframe link reg save.
22589 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
22591 2017-02-03 Jakub Jelinek <jakub@redhat.com>
22594 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
22595 wu for stxssp alternative.
22597 2017-02-03 Martin Sebor <msebor@redhat.com>
22599 PR tree-optimization/79352
22600 * gimple-fold.c (get_range_strlen): Add argument.
22601 (get_range_strlen): Change return type to bool.
22602 (get_maxval_strlen): Pass in a dummy argument.
22603 * gimple-fold.h (get_range_strlen): Change return type to bool.
22604 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
22605 * tree.h (array_at_struct_end_p): Add argument.
22606 * tree.c (array_at_struct_end_p): Handle it.
22608 2017-02-03 Martin Liska <mliska@suse.cz>
22611 * multiple_target.c (create_dispatcher_calls): Redirect edge
22612 from a caller of a dispatcher.
22613 (expand_target_clones): Make the clones local.
22614 (ipa_target_clone): Do both target clones and resolvers.
22615 (ipa_dispatcher_calls): Remove the pass.
22616 (pass_dispatcher_calls::gate): Likewise.
22617 (make_pass_dispatcher_calls): Likewise.
22618 * passes.def (pass_target_clone): Put as very first IPA early
22621 2017-02-03 Martin Liska <mliska@suse.cz>
22623 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
22624 in case of a function with ifunc attribute.
22626 2017-02-03 Martin Liska <mliska@suse.cz>
22628 * cgraph.c (cgraph_node::dump): Dump function version info.
22629 * symtab.c (symtab_node::dump_base): Add missing new line.
22631 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
22633 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
22634 (ifcombine_ifandif): Use it.
22636 2017-02-03 Martin Liska <mliska@suse.cz>
22638 * doc/invoke.texi: Document default value for
22639 use-after-scope-direct-emission-threshold.
22641 2017-02-03 Martin Liska <mliska@suse.cz>
22643 PR tree-optimization/79339
22644 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
22645 (format_floating): Likewise.
22647 2017-02-03 Martin Liska <mliska@suse.cz>
22650 * ipa-prop.c (ipa_node_params_t::insert): Remove current
22652 (ipa_node_params_t::remove): Likewise.
22653 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
22654 initialization from removed ipa_node_params_t::insert.
22655 (ipa_node_params::~ipa_node_params): Move from removed
22656 ipa_node_params_t::release.
22657 * symbol-summary.h (symbol_summary::m_released): New member.
22658 Do not release a summary twice. Do not allow to call finalizer
22659 for types of a summary that live in GGC memory.
22661 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
22663 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
22666 2017-02-02 Martin Sebor <msebor@redhat.com>
22668 PR middle-end/79275
22669 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
22670 (format_string): Tighten up the range of output for non-constant
22671 strings and correct the expected range for wide non-constant strings.
22673 2017-02-02 Martin Sebor <msebor@redhat.com>
22675 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
22677 PR middle-end/32003
22678 * doc/invoke.texi (-fdump-tree-): Remove pass-specific options from
22680 (-fdump-tree-@var): Add to index and document how to come up
22681 with pass-specific option and dump file names.
22682 (-fdump-passes): Clarify where to look for output.
22684 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
22686 PR middle-end/77445
22687 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
22688 statistics of the analyzed path; allow threading for speed when
22689 any of BBs along the path are optimized for speed.
22691 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
22693 PR middle-end/78468
22694 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
22695 settings of the virtual registers.
22698 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
22700 * explow.c (get_dynamic_stack_size): Take known alignment of stack
22701 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
22704 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
22706 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
22707 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
22709 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
22711 * config/s390/s390.md: Add missing comments with the expanded
22713 * config/s390/vector.md: Likewise.
22714 * config/s390/vx-builtins.md: Likewise.
22716 2017-02-02 Jakub Jelinek <jakub@redhat.com>
22719 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
22720 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
22721 conditions on a single line.
22723 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
22725 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
22726 __S390_VX__ to __VX__.
22728 2017-02-01 Andrew Pinski <apinski@cavium.com>
22730 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
22731 stmt_info to record_stmt_cost.
22732 (vect_get_known_peeling_cost): Pass stmt_info if known to
22734 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
22735 cpu_vector_cost field into
22736 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
22737 field into vec_int_stmt_cost and vec_fp_stmt_cost.
22738 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
22739 splitting of scalar_stmt_cost and vec_stmt_cost.
22740 (thunderx_vector_cost): Likewise.
22741 (cortexa57_vector_cost): LIkewise.
22742 (exynosm1_vector_cost): Likewise.
22743 (xgene1_vector_cost): Likewise.
22744 (thunderx2t99_vector_cost): Improve after the splitting of the two
22746 (aarch64_builtin_vectorization_cost): Update for the splitting of
22747 scalar_stmt_cost and vec_stmt_cost.
22749 2017-02-01 Torvald Riegel <triegel@redhat.com>
22750 Richard Henderson <rth@redhat.com>
22752 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
22753 conditional on existance of a fast atomic load.
22754 * optabs-query.c (can_atomic_load_p): New function.
22755 * optabs-query.h (can_atomic_load_p): Declare it.
22756 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
22757 no fast atomic load is available for the particular size of access.
22758 (expand_atomic_compare_and_swap): Likewise.
22759 (expand_atomic_load): Likewise.
22760 (expand_atomic_store): Likewise.
22761 (expand_atomic_fetch_op): Likewise.
22762 * testsuite/lib/target-supports.exp
22763 (check_effective_target_sync_int_128): Remove x86 because it provides
22764 no fast atomic load.
22765 (check_effective_target_sync_int_128_runtime): Likewise.
22767 2017-02-01 Richard Biener <rguenther@suse.de>
22769 * graphite.c: Include tree-vectorizer.h for find_loop_location.
22770 (graphite_transform_loops): Provide opt-info for optimized nests.
22771 * tree-parloop.c (parallelize_loops): Provide opt-info for
22772 parallelized loops.
22774 2017-02-01 Richard Biener <rguenther@suse.de>
22776 PR middle-end/79315
22777 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
22778 was not set before.
22780 2017-02-01 Richard Biener <rguenther@suse.de>
22782 PR tree-optimization/71824
22783 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
22784 Verify the loops are valid in the merged SESE region.
22785 (scop_detection::can_represent_loop_1): Check analyzing the
22786 evolution of the number of iterations in the region succeeds.
22788 2017-01-31 Ian Lance Taylor <iant@golang.org>
22790 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
22791 REG_ARGS_SIZE note to 32-bit push insns and call insn.
22793 2017-01-31 David Malcolm <dmalcolm@redhat.com>
22795 PR preprocessor/79210
22796 * input.c (get_substring_ranges_for_loc): Replace line_width
22797 assertion with error-handling.
22799 2017-01-31 Richard Biener <rguenther@suse.de>
22801 PR tree-optimization/77318
22802 * graphite-sese-to-poly.c (extract_affine): Fix assert.
22803 (create_pw_aff_from_tree): Take loop parameter.
22804 (add_condition_to_pbb): Pass loop of the condition to
22805 create_pw_aff_from_tree.
22807 2017-01-31 Jakub Jelinek <jakub@redhat.com>
22809 * config/s390/s390.c (s390_asan_shadow_offset): New function.
22810 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
22812 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
22816 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
22818 (convert_int_to_float128): Likewise.
22819 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
22820 (convert_int_to_float128): Likewise.
22821 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
22822 (UNSPEC_IEEE128_CONVERT): Likewise.
22823 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
22824 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
22825 Use local variables for IBM extended format.
22826 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
22827 (fix_trunc<mode>si2_fprs): Likewise.
22828 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
22829 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
22830 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
22831 to know that we can now have integers of all sizes in vector
22833 (fix<uns>_<mode>di2_hw): Likewise.
22834 (float<uns>_<mode>si2_hw): Likewise.
22835 (fix_<mode>si2_hw): Likewise.
22836 (fixuns_<mode>si2_hw): Likewise.
22837 (float<uns>_<mode>di2_hw): Likewise.
22838 (float_<mode>di2_hw): Likewise.
22839 (float_<mode>si2_hw): Likewise.
22840 (floatuns_<mode>di2_hw): Likewise.
22841 (floatuns_<mode>si2_hw): Likewise.
22842 (xscvqp<su>wz_<mode>): Delete, no longer used.
22843 (xscvqp<su>dz_<mode>): Likewise.
22844 (xscv<su>dqp_<mode>): Likewise.
22845 (ieee128_mfvsrd_64bit): Likewise.
22846 (ieee128_mfvsrd_32bit): Likewise.
22847 (ieee128_mfvsrwz): Likewise.
22848 (ieee128_mtvsrw): Likewise.
22849 (ieee128_mtvsrd_64bit): Likewise.
22850 (ieee128_mtvsrd_32bit): Likewise.
22852 2017-01-31 Martin Liska <mliska@suse.cz>
22855 * ipa-prop.c (ipa_free_all_node_params): Call release method
22856 instead of ~sumbol_summary to not to trigger double times
22859 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
22861 PR tree-optimization/71691
22862 * bitmap.h (class auto_bitmap): New.
22863 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
22864 is_maybe_undefined instead of ssa_undefined_value_p.
22866 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
22868 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
22869 __S390_ARCH_LEVEL__ to __ARCH__.
22871 2017-01-31 Jakub Jelinek <jakub@redhat.com>
22873 PR tree-optimization/79267
22874 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
22875 if should_remove_lhs_p is true.
22877 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
22880 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
22881 (add_alignment_attribute): New.
22882 (base_type_die): Add alignment attribute.
22883 (subrange_type_die): Likewise.
22884 (modified_type_die): Likewise.
22885 (gen_array_type_die): Likewise.
22886 (gen_descr_array_type_die: Likewise.
22887 (gen_enumeration_type_die): Likewise.
22888 (gen_subprogram_die): Likewise.
22889 (gen_variable_die): Likewise.
22890 (gen_field_die): Likewise.
22891 (gen_ptr_to_mbr_type_die): Likewise.
22892 (gen_struct_or_union_type_die): Likewise.
22893 (gen_subroutine_type_die): Likewise.
22894 (gen_typedef_die): Likewise.
22895 (base_type_cmp): Compare alignment attribute.
22897 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
22900 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
22901 (setb_unsigned) New pattern for setb with CCUNS.
22902 * config/rs6000/rs6000.c (expand_block_compare): Use a different
22903 subfc./subfe sequence to avoid overflow problems. Generate a
22904 shorter sequence with cmpld/setb for power9.
22905 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
22906 for generating subfc. instruction.
22907 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
22908 now uses this instruction.
22910 2017-01-30 Ian Lance Taylor <iant@google.com>
22913 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
22914 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
22916 2017-01-30 Martin Sebor <msebor@redhat.com>
22918 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
22919 Move constant to the right of a relational operator.
22920 (get_mpfr_format_length, format_character, format_string): Ditto.
22921 (should_warn_p, maybe_warn): Same.
22923 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
22925 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
22928 * asan.c (get_translation_unit_decl): Remove function.
22929 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
22931 2017-01-30 Martin Liska <mliska@suse.cz>
22933 PR gcov-profile/79259
22934 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
22935 -fprofile-generate.
22937 2017-01-30 Martin Liska <mliska@suse.cz>
22940 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
22941 Initialize variables with NULL value.
22943 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
22946 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
22948 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
22950 2017-01-30 Richard Biener <rguenther@suse.de>
22952 PR tree-optimization/79276
22953 * tree-vrp.c (process_assert_insertions): Properly adjust common
22954 when removing a duplicate.
22956 * gcc.dg/torture/pr79276.c: New testcase.
22958 2017-01-30 Richard Biener <rguenther@suse.de>
22960 PR tree-optimization/79256
22961 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
22962 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
22964 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
22966 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
22969 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
22970 ("*r<noxa>sbg_<mode>_sll_bitmask")
22971 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
22972 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
22973 Use contiguous_bitmask_nowrap_operand.
22975 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22978 * config/rs6000/altivec.h (vec_xl): Revise #define.
22979 (vec_xst): Likewise.
22981 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
22983 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
22985 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
22987 PR rtl-optimization/79194
22988 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
22989 traps before call to bypass_conditional_jumps.
22991 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
22993 PR tree-optimization/71374
22994 * lra-constraints.c (check_conflict_input_operands): New.
22995 (match_reload): Use it.
22997 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
23000 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
23001 account to calculate conflict_set.
23003 2017-01-27 Bin Cheng <bin.cheng@arm.com>
23005 PR rtl-optimization/78559
23006 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
23007 other_insn in combine.
23009 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
23011 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
23012 uint16_type_node for BT_UINT16.
23014 2017-01-27 David Malcolm <dmalcolm@redhat.com>
23016 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
23017 "RTL Tests" to menu.
23018 (GIMPLE Tests): New node.
23019 (RTL Tests): New node.
23021 2017-01-27 Richard Biener <rguenther@suse.de>
23023 PR tree-optimization/79245
23024 * tree-loop-distribution.c (distribute_loop): Apply cost
23025 modeling also to detected patterns.
23027 2017-01-27 Richard Biener <rguenther@suse.de>
23029 PR tree-optimization/71433
23030 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
23031 (compare_assert_loc): New function.
23032 (process_assert_insertions): Sort and optimize assert locations
23033 to remove duplicates and push down identical assertions on
23034 edges to their destination block.
23036 2017-01-27 Richard Biener <rguenther@suse.de>
23038 PR tree-optimization/79244
23039 * tree-vrp.c (remove_range_assertions): Forcefully propagate
23040 out SSA names even if abnormal.
23042 2017-01-27 Jakub Jelinek <jakub@redhat.com>
23044 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
23045 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
23046 instead of MPFR_RNDN.
23048 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
23051 * arm.c (arm_option_override): Don't call build_target_option_node
23052 until after doing all option overrides.
23053 (arm_valid_target_attribute_tree): Likewise.
23055 2017-01-27 Martin Liska <mliska@suse.cz>
23057 * doc/invoke.texi (-fprofile-arcs): Document profiling support
23058 for {cd}tors and C++ {cd}tors.
23060 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
23062 * config/s390/s390.md ("*setmem_long_and")
23063 ("*setmem_long_and_31z"): Use zero_extend instead of and.
23065 2017-01-26 Martin Sebor <msebor@redhat.com>
23067 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
23070 2017-01-26 Martin Sebor <msebor@redhat.com>
23072 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
23073 HAVE_DFmode before using XFmode or DFmode.
23074 (parse_directive): Avoid using the z length modifier to avoid
23075 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
23077 PR middle-end/78703
23078 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
23079 to accept adjustment as an array.
23080 (get_int_range): New function.
23081 (struct directive): Make width and prec arrays.
23082 (directive::set_width, directive::set_precision): Call get_int_range.
23083 (format_integer, format_floating): Handle width and precision ranges.
23084 (format_string, parse_directive): Same.
23086 2017-01-26 Jakub Jelinek <jakub@redhat.com>
23089 * dwarf2out.c (generate_skeleton_bottom_up): For children with
23090 comdat_type_p set, just clone them, but keep the children in the
23094 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
23095 which have direct callers with -fvar-tracking-assignments enabled
23097 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
23098 inside of type units.
23100 2017-01-26 Martin Sebor <msebor@redhat.com>
23102 PR middle-end/78703
23103 * gimple-ssa-sprintf.c (struct result_range): Add likely and
23105 (struct format_result): Replace number_chars, number_chars_min,
23106 and number_chars_max with a single member of struct result_range.
23108 (format_result::operator+=): Adjust.
23109 (struct fmtresult): Remove bounded. Handle likely and unlikely
23111 (fmtresult::adjust_for_width_or_precision): New function.
23112 (fmtresult:type_max_digits): New function.
23113 (bytes_remaining): Handle likely and unlikely counters.
23114 (min_bytes_remaining): Remove.
23115 (format_percent): Simplify.
23116 (format_integer, format_floating): Set likely and unlikely counters.
23117 (get_string_length, format_character, format_string): Same.
23118 (format_plain, should_warn_p): New function.
23119 (maybe_warn): Call should_warn_p. Update diagnostic messages
23120 and handle those for all directives, including plain strings.
23121 (format_directive): Handle likely and unlikely counters.
23122 Remove unnecessary quoting from diagnostics. Add an informational
23124 (add_bytes): Remove.
23125 (pass_sprintf_length::compute_format_length): Simplify.
23126 (try_substitute_return_value): Handle likely and unlikely counters.
23128 2017-01-26 Carl Love <cel@us.ibm.com>
23130 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
23131 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
23133 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
23136 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
23137 endianess for subregs into account.
23138 * lra-constraints.c (lra_constraints): Do risky transformations
23139 always on the first iteration.
23140 * lra-lives.c (check_pseudos_live_through_calls): Add arg
23141 last_call_used_reg_set.
23142 (process_bb_lives): Define and use last_call_used_reg_set.
23143 * lra.c (lra): Always continue after lra_constraints on the first
23146 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
23148 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
23150 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
23152 2017-01-26 Jakub Jelinek <jakub@redhat.com>
23154 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
23155 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
23156 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
23157 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
23158 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
23159 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
23160 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
23161 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
23162 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
23164 2017-01-26 Marek Polacek <polacek@redhat.com>
23167 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
23168 for the third operand.
23170 2017-01-26 Jakub Jelinek <jakub@redhat.com>
23172 PR middle-end/79236
23173 * omp-low.c (struct omp_context): Add simt_stmt field.
23174 (scan_omp_for): Return omp_context *.
23175 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
23176 context to the _simt_ SIMD stmt.
23177 (lower_omp_for): For combined SIMD with sibling _simt_
23178 SIMD, make sure to use the same decls in _looptemp_
23179 clauses as in the sibling.
23181 2017-01-26 David Sherwood <david.sherwood@arm.com>
23183 PR middle-end/79212
23184 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
23187 2017-01-26 Jakub Jelinek <jakub@redhat.com>
23190 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
23191 emit fld b; fld a; if possible.
23193 * brig-builtins.def: Update copyright years.
23194 * config/arm/arm_acle_builtins.def: Update copyright years.
23196 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
23199 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
23200 constraint instead of o for the stxsd instruction.
23202 2017-01-25 Carl Love <cel@us.ibm.com>
23204 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
23205 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
23207 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
23209 * doc/invoke.texi (C++ Dialect Options): Fix typo.
23211 2017-01-25 Richard Biener <rguenther@suse.de>
23213 PR tree-optimization/69264
23214 * target.def (vector_alignment_reachable): Improve documentation.
23215 * doc/tm.texi: Regenerate.
23216 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
23218 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
23219 earlier changes with respect to TYPE_USER_ALIGN.
23220 (vector_alignment_reachable_p): Likewise. Improve dumping.
23222 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23225 * config/arm/arm.md (xordi3): Force constant operand into a register
23228 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23230 * doc/invoke.texi (-fstore-merging): Correct default optimization
23231 levels at which it is enabled.
23232 (-O): Move -fstore-merging from list to...
23235 2017-01-25 Richard Biener <rguenther@suse.de>
23238 * omp-expand.c: Include debug.h.
23239 (expand_omp_taskreg): Make sure to generate early debug before
23240 outlining anything from a function.
23241 (expand_omp_target): Likewise.
23242 (grid_expand_target_grid_body): Likewise.
23244 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
23247 * asan.c (get_translation_unit_decl): New function.
23248 (asan_add_global): Extract modules file name from globals
23249 TRANSLATION_UNIT_DECL name.
23251 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
23254 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
23255 for long calls with APCS frame and VFP.
23257 2017-01-24 David Malcolm <dmalcolm@redhat.com>
23259 * cfg.c (original_copy_tables_initialized_p): New function.
23260 * cfg.h (original_copy_tables_initialized_p): New decl.
23261 * cfgrtl.c (relink_block_chain): Guard the call to
23262 free_original_copy_tables with a call to
23263 original_copy_tables_initialized_p.
23264 * cgraph.h (symtab_node::native_rtl_p): New decl.
23265 * cgraphunit.c (symtab_node::native_rtl_p): New function.
23266 (symtab_node::needed_p): Don't assert for early assembly output
23267 for __RTL functions.
23268 (cgraph_node::finalize_function): Set "force_output" for __RTL
23270 (cgraph_node::analyze): Bail out early for __RTL functions.
23271 (analyze_functions): Update assertion to support __RTL functions.
23272 (cgraph_node::expand): Bail out early for __RTL functions.
23273 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
23275 * function.h (struct function): Update comment for field
23277 * gimple-expr.c: Include "tree-pass.h".
23278 (gimple_has_body_p): Return false for __RTL functions.
23279 * Makefile.in (OBJS): Add run-rtl-passes.o.
23280 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
23282 (gcc::pass_manager::get_clean_slate): New accessor.
23283 * passes.c: Include "insn-addr.h".
23284 (should_skip_pass_p): Add logging. Update logic for running
23285 "expand" to be compatible with both __GIMPLE and __RTL. Guard
23286 property-provider override so it is only done for gimple passes.
23288 (skip_pass): New function.
23289 (execute_one_pass): Call skip_pass when skipping passes.
23290 * read-md.c (md_reader::read_char): Support filtering
23291 the input to a subset of line numbers.
23292 (md_reader::md_reader): Initialize fields
23293 m_first_line and m_last_line.
23294 (md_reader::read_file_fragment): New function.
23295 * read-md.h (md_reader::read_file_fragment): New decl.
23296 (md_reader::m_first_line): New field.
23297 (md_reader::m_last_line): New field.
23298 * read-rtl-function.c (function_reader::create_function): Only
23299 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
23300 curr_properties. Set DECL_INITIAL to a dummy block.
23301 (read_rtl_function_body_from_file_range): New function.
23302 * read-rtl-function.h (read_rtl_function_body_from_file_range):
23304 * run-rtl-passes.c: New file.
23305 * run-rtl-passes.h: New file.
23307 2017-01-24 Jeff Law <law@redhat.com>
23309 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
23312 2017-01-24 Bin Cheng <bin.cheng@arm.com>
23314 PR tree-optimization/79159
23315 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
23316 (record_nonwrapping_iv): Improve boundary using above function if no
23317 value range information.
23319 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
23320 Martin Jambor <mjambor@suse.cz>
23322 * brig-builtins.def: New file.
23323 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
23324 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
23325 (DEF_HSAIL_SAT_BUILTIN): Likewise.
23326 (DEF_HSAIL_INTR_BUILTIN): Likewise.
23327 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
23328 * builtin-types.def (BT_INT8): New.
23329 (BT_INT16): Likewise.
23330 (BT_UINT8): Likewise.
23331 (BT_UINT16): Likewise.
23332 (BT_FN_ULONG): Likewise.
23333 (BT_FN_UINT_INT): Likewise.
23334 (BT_FN_UINT_ULONG): Likewise.
23335 (BT_FN_UINT_LONG): Likewise.
23336 (BT_FN_UINT_PTR): Likewise.
23337 (BT_FN_ULONG_PTR): Likewise.
23338 (BT_FN_INT8_FLOAT): Likewise.
23339 (BT_FN_INT16_FLOAT): Likewise.
23340 (BT_FN_UINT32_FLOAT): Likewise.
23341 (BT_FN_UINT16_FLOAT): Likewise.
23342 (BT_FN_UINT8_FLOAT): Likewise.
23343 (BT_FN_UINT64_FLOAT): Likewise.
23344 (BT_FN_UINT16_UINT32): Likewise.
23345 (BT_FN_UINT32_UINT16): Likewise.
23346 (BT_FN_UINT16_UINT16_UINT16): Likewise.
23347 (BT_FN_INT_PTR_INT): Likewise.
23348 (BT_FN_UINT_PTR_UINT): Likewise.
23349 (BT_FN_LONG_PTR_LONG): Likewise.
23350 (BT_FN_ULONG_PTR_ULONG): Likewise.
23351 (BT_FN_VOID_UINT64_UINT64): Likewise.
23352 (BT_FN_UINT8_UINT8_UINT8): Likewise.
23353 (BT_FN_INT8_INT8_INT8): Likewise.
23354 (BT_FN_INT16_INT16_INT16): Likewise.
23355 (BT_FN_INT_INT_INT): Likewise.
23356 (BT_FN_UINT_FLOAT_UINT): Likewise.
23357 (BT_FN_FLOAT_UINT_UINT): Likewise.
23358 (BT_FN_ULONG_UINT_UINT): Likewise.
23359 (BT_FN_ULONG_UINT_PTR): Likewise.
23360 (BT_FN_ULONG_ULONG_ULONG): Likewise.
23361 (BT_FN_UINT_UINT_UINT): Likewise.
23362 (BT_FN_VOID_UINT_PTR): Likewise.
23363 (BT_FN_UINT_UINT_PTR: Likewise.
23364 (BT_FN_UINT32_UINT64_PTR): Likewise.
23365 (BT_FN_INT_INT_UINT_UINT): Likewise.
23366 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
23367 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
23368 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
23369 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
23370 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
23371 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
23372 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
23373 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
23374 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
23375 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
23376 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
23377 * doc/frontends.texi: List BRIG FE.
23378 * doc/install.texi (Testing): Add BRIG tesring requirements.
23379 * doc/invoke.texi (Overall Options): Mention BRIG.
23380 * doc/standards.texi (Standards): Doucment BRIG HSA version.
23382 2017-01-24 Richard Biener <rguenther@suse.de>
23384 PR translation/79208
23385 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
23387 2017-01-24 Martin Jambor <mjambor@suse.cz>
23390 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
23391 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
23392 and known_contexts.
23394 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
23396 PR middle-end/79123
23397 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
23398 casts from signed to unsigned really don't have a range.
23400 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
23402 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
23403 GMP_RNDx for compatiblity.
23405 2017-01-24 Martin Liska <mliska@suse.cz>
23408 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
23409 that would prevent us to call alloca with -1 as argument.
23411 2017-01-24 Jakub Jelinek <jakub@redhat.com>
23413 * dwarf2out.c (output_compilation_unit_header, output_file_names):
23414 Avoid -Wformat-security warning.
23416 2017-01-23 Andrew Pinski <apinski@cavium.com>
23418 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
23421 2017-01-23 Martin Sebor <msebor@redhat.com>
23423 PR middle-end/78703
23424 * gimple-ssa-sprintf.c (warn_level): New global.
23425 (format_integer): Use it here and throughout the rest of the file.
23426 Use the same switch to compute sign as base.
23427 (maybe_warn): New function.
23428 (format_directive): Factor out warnings into maybe_warn.
23429 Add debugging output. Use warn_level.
23430 (add_bytes): Use warn_level.
23431 (pass_sprintf_length::compute_format_length): Add debugging output.
23432 (try_substitute_return_value): Same.
23433 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
23435 PR middle-end/78703
23436 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
23437 (struct fmtresult, format_integer, format_floating): Adjust.
23438 (fmtresult::fmtresult): Set max correctly in two argument ctor.
23439 (get_string_length, format_string,format_directive): Same.
23440 (pass_sprintf_length::compute_format_length): Same.
23441 (try_substitute_return_value): Simplify slightly.
23443 PR middle-end/78703
23444 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
23445 (fmtresult::operator+=): Outlined.
23446 (struct fmtresult): Add ctors.
23447 (struct conversion_spec): Rename...
23448 (struct directive): ...to this. Add and remove data members.
23449 (directive::set_width, directive::set_precision): New functions.
23450 (format_percent): Use fmtresult ctor.
23451 (get_width_and_precision): Remove.
23452 (format_integer): Make naming changes. Avoid computing width and
23454 (format_floating): Same. Adjust indentation.
23455 (format_character, format_none): New functions.
23456 (format_string): Moved character handling to format_character.
23457 (format_directive): Remove arguments, change return type.
23458 (parse_directive): New function.
23459 (pass_sprintf_length::compute_format_length): Move directive
23460 parsing to parse_directive.
23462 2017-01-23 Jakub Jelinek <jakub@redhat.com>
23464 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
23465 (assign_assembler_name_if_needed): ... this.
23466 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
23467 (assign_assembler_name_if_needed): ... this.
23468 (free_lang_data_in_cgraph): Adjust callers.
23469 * cgraphunit.c (cgraph_node::analyze): Likewise.
23470 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
23473 2017-01-23 Richard Biener <rguenther@suse.de>
23475 PR tree-optimization/79088
23476 PR tree-optimization/79188
23477 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
23478 resetting loop bounds after last path deletion. Reset loop
23479 bounds of the target loop, make code match the comments.
23480 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
23481 Make sure loops need no fixups.
23483 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
23485 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
23486 exponent support with double type for first argument.
23487 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
23488 type returned by __builtin_vec_extract_sig,
23489 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
23490 functions from "vector int" to "vector unsigned int" or from
23491 "vector long long int" to "vector unsigned long long int".
23492 Changed type returned by __builtin_vec_extract_exp,
23493 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
23494 functions from "vector int" to "vector unsigned int" or from
23495 "vector long long int" to "vector unsigned long long int".
23496 Changed return type of __builtin_vec_test_data_class,
23497 __builtin_vec_test_data_class_sp, and
23498 __builtin_vec_test_data_class_dp from "vector int" to
23499 "vector bool int" or from "vector long long int" to "vector bool
23500 long long int" and changed second argument type from "unsigned
23501 int" to "int". Added new overloaded function forms "vector float
23502 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
23503 "vector float __builtin_vec_insert_exp_sp (vector float, vector
23504 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
23505 double, vector unsigned long long int)" and "vector double
23506 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
23507 long int)". Changed return type of
23508 __builtin_scalar_test_data_class and
23509 __builtin_scalar_test_data_class_sp and
23510 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
23511 int" and changed second argument from "unsigned int" to "int".
23512 Changed type returned by __builtin_scalar_test_neg,
23513 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
23514 from "int" to "bool int". Added new overloaded function form
23515 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
23516 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
23517 exponent double-precision with floating point first argument.
23518 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
23519 documentation of scalar_test_data_class, scalar_test_neg,
23520 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
23521 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
23522 vec_test_data_class built-in functions to reflect refinements in
23523 their type signatures.
23525 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
23527 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
23529 (aarch64_elf_asm_destructor): Likewise.
23531 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
23533 PR rtl-optimization/78634
23534 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
23535 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
23536 * ifcvt.c (noce_try_cmove): Add missing cost check.
23538 PR rtl-optimization/71724
23539 * combine.c (if_then_else_cond): Look for situations where it is
23540 beneficial to undo the work of one of the recursive calls.
23542 2017-01-23 Bin Cheng <bin.cheng@arm.com>
23544 PR tree-optimization/70754
23545 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
23546 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
23547 combined stmt before it if not NULL.
23548 (combine_chains): Process refs reversely and compute dominance point
23551 2017-01-23 Martin Liska <mliska@suse.cz>
23553 PR tree-optimization/79196
23554 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
23555 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
23557 (strlen_optimize_stmt): Call the renamed function.
23559 2017-01-23 Michael Matz <matz@suse.de>
23561 PR tree-optimization/78384
23562 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
23564 2017-01-23 Richard Biener <rguenther@suse.de>
23566 PR tree-optimization/79186
23567 * tree-vrp.c (register_new_assert_for): Make sure we've seen
23568 both incoming edges before moving an assert.
23570 2017-01-23 Martin Jambor <mjambor@suse.cz>
23572 * ipa-prop.c (load_from_param_1): Removed.
23573 (load_from_unmodified_param): Bits from load_from_param_1 put back
23575 (load_from_param): Removed.
23576 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
23577 with stmt. Reverted back to use of load_from_unmodified_param.
23579 2017-01-23 Martin Jambor <mjambor@suse.cz>
23582 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
23583 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
23584 field a pointer to garbage collected vector, mark lattices and
23585 ipcp_orig_node with GTY((skip)).
23586 (ipa_get_param_count): Adjust to descriptors being a pointer.
23587 (ipa_get_param): Likewise.
23588 (ipa_get_type): Likewise.
23589 (ipa_get_param_move_cost): Likewise.
23590 (ipa_set_param_used): Likewise.
23591 (ipa_get_controlled_uses): Likewise.
23592 (ipa_set_controlled_uses): Likewise.
23593 (ipa_is_param_used): Likewise.
23594 (ipa_node_params_t): Move into garbage collector. New methods insert
23596 (ipa_node_params_sum): Annotate wth GTY(()).
23597 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
23599 (ipa_load_from_parm_agg): Adjust declaration.
23600 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
23601 * ipa-profile.c (ipa_profile): Likewise.
23602 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
23603 (ipa_populate_param_decls): Make descriptors parameter garbage
23605 (ipa_dump_param): Adjust to descriptors being a pointer.
23606 (ipa_alloc_node_params): Likewise.
23607 (ipa_initialize_node_params): Likewise.
23608 (load_from_param_1): Make descriptors parameter garbage collected.
23609 (load_from_unmodified_param): Likewise.
23610 (load_from_param): Likewise.
23611 (ipa_load_from_parm_agg): Likewise.
23612 (ipa_node_params::~ipa_node_params): Removed.
23613 (ipa_free_all_node_params): Remove call to delete operator.
23614 (ipa_node_params_t::insert): New.
23615 (ipa_node_params_t::remove): Likewise.
23616 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
23617 copy known_csts and known_contexts vectors.
23618 (ipa_read_node_info): Adjust to descriptors being a pointer.
23619 (ipcp_modif_dom_walker): Make m_descriptors field garbage
23621 (ipcp_transform_function): Make descriptors variable garbage
23624 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
23626 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
23627 * config/i386/avx512dqintrin.h: Ditto.
23628 * config/i386/avx512fintrin.h: Ditto.
23629 * config/i386/i386.c: Handle new builtins.
23630 * config/i386/i386-builtin.def: Add new builtins.
23631 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
23632 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
23634 2017-01-23 Jakub Jelinek <jakub@redhat.com>
23635 Martin Liska <mliska@suse.cz>
23637 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
23638 * asan.c (asan_expand_poison_ifn): Support stores and use
23639 appropriate ASAN report function.
23640 * internal-fn.c (expand_ASAN_POISON_USE): New function.
23641 * internal-fn.def (ASAN_POISON_USE): Declare.
23642 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
23643 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
23644 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
23645 ASAN_POISON calls w/o LHS.
23646 * tree-ssa.c (execute_update_addresses_taken): Create clobber
23647 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
23648 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
23649 * gimplify.c (asan_poison_variables): Add attribute
23650 use_after_scope_memory to variables that really needs to live
23652 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
23653 having the attribute.
23655 2017-01-23 Martin Liska <mliska@suse.cz>
23657 * asan.c (create_asan_shadow_var): New function.
23658 (asan_expand_poison_ifn): Likewise.
23659 * asan.h (asan_expand_poison_ifn): New declaration.
23660 * internal-fn.c (expand_ASAN_POISON): Likewise.
23661 * internal-fn.def (ASAN_POISON): New builtin.
23662 * sanopt.c (pass_sanopt::execute): Expand
23663 asan_expand_poison_ifn.
23664 * tree-inline.c (copy_decl_for_dup_finish): Make function
23666 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
23667 * tree-ssa.c (is_asan_mark_p): New function.
23668 (execute_update_addresses_taken): Rewrite local variables
23669 (identified just by use-after-scope as addressable) into SSA.
23671 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
23673 * doc/install.texi (Specific): opensource.apple.com uses https
23674 now. Remove trailing slash.
23676 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
23678 * README.Portability: Remove note on an Irix compatibility issue.
23680 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
23682 * gcov.c (INCLUDE_ALGORITHM): Define.
23683 (INCLUDE_VECTOR): Define.
23684 No longer include <vector> and <algorithm> directly.
23686 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
23688 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
23690 * doc/invoke.texi (Code Gen Options): Ditto.
23692 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
23695 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
23697 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
23699 rtl-optimization/79125
23700 * cprop.c (local_cprop_pass): Handle cases where we make an
23701 unconditional trap.
23703 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
23707 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
23708 read from, for big endian.
23710 2017-01-20 Jiong Wang <jiong.wang@arm.com>
23712 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
23713 register pauth builtins for LP64 only.
23715 2017-01-20 Marek Polacek <polacek@redhat.com>
23718 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
23721 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
23723 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
23725 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
23726 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
23727 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
23729 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23732 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
23733 in big-endian mode when they are not a single duplicated value.
23735 2017-01-20 Richard Biener <rguenther@suse.de>
23737 * BASE-VER: Bump to 7.0.1.
23739 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
23741 * omp-low.c (omplow_simd_context): New struct. Use it...
23742 (lower_rec_simd_input_clauses): ...here and...
23743 (lower_rec_input_clauses): ...here to hold common data. Adjust all
23744 references to idx, lane, max_vf, is_simt.
23746 2017-01-20 Graham Markall <graham.markall@embecosm.com>
23748 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
23751 2017-01-20 Martin Jambor <mjambor@suse.cz>
23753 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
23754 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
23756 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
23757 (GTFILES): Rename hsa.c to hsa-common.c.
23758 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
23759 * hsa-dump.c: Likewise.
23760 * hsa-gen.c: Likewise.
23761 * hsa-regalloc.c: Likewise.
23762 * ipa-hsa.c: Likewise.
23763 * omp-expand.c: Likewise.
23764 * omp-low.c: Likewise.
23765 * toplev.c: Likewise.
23767 2017-01-20 Marek Polacek <polacek@redhat.com>
23770 * doc/invoke.texi: Document -Wduplicated-branches.
23771 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
23772 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
23773 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
23774 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
23775 return 0 only when not OEP_LEXICOGRAPHIC.
23776 (fold_build_cleanup_point_expr): Use the expression
23777 location when building CLEANUP_POINT_EXPR.
23778 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
23779 * tree.c (add_expr): Handle error_mark_node.
23781 2017-01-20 Martin Liska <mliska@suse.cz>
23784 * tree-profile.c (init_ic_make_global_vars): Do not call
23786 (gimple_init_gcov_profiler): Likewise.
23788 2017-01-20 Martin Liska <mliska@suse.cz>
23791 * cgraph.h (maybe_create_reference): Remove argument and
23793 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
23795 * ipa-cp.c (create_specialized_node): Likewise.
23796 * symtab.c (symtab_node::maybe_create_reference): Handle
23797 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
23799 2017-01-20 Martin Liska <mliska@suse.cz>
23801 * read-rtl-function.c (function_reader::create_function): Use
23802 build_decl instread of build_decl_stat.
23804 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
23806 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
23807 * config/i386/avx512dqintrin.h: Ditto.
23808 * config/i386/avx512fintrin.h: Ditto.
23809 * config/i386/i386-builtin-types.def: Add new types.
23810 * config/i386/i386.c: Handle new types.
23811 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
23812 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
23813 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
23814 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
23815 (__builtin_ia32_kshiftridi): New.
23816 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
23818 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
23822 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
23823 define to rs6000_init_stack_protect_guard.
23824 (rs6000_init_stack_protect_guard): New function.
23826 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
23827 Yunqiang Su <yunqiang.su@imgtec.com>
23829 * config.gcc (supported_defaults): Add madd4.
23830 (with_madd4): Add validation.
23831 (all_defaults): Add madd4.
23832 * config/mips/mips.opt (mmadd4): New option.
23833 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
23835 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
23837 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
23838 (ISA_HAS_FUSED_MADD4): Likewise.
23839 * doc/invoke.texi (-mmadd4): Document the new option.
23840 * doc/install.texi (--with-madd4): Document the new option.
23842 2017-01-19 Jiong Wang <jiong.wang@arm.com>
23844 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
23845 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
23846 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
23847 (aarch64_init_pauth_hint_builtins): New.
23848 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
23849 (aarch64_expand_builtin): Expand new builtins.
23851 2017-01-19 Jiong Wang <jiong.wang@arm.com>
23853 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
23854 * combine-stack-adj.c (no_unhandled_cfa): Handle
23855 REG_CFA_TOGGLE_RA_MANGLE.
23856 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
23857 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
23858 info for return address signing.
23859 (aarch64_expand_epilogue): Likewise.
23861 2017-01-19 Jiong Wang <jiong.wang@arm.com>
23863 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
23864 * config/aarch64/aarch64-protos.h
23865 (aarch64_return_address_signing_enabled): New declaration.
23866 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
23868 (aarch64_expand_prologue): Sign return address before it's pushed onto
23870 (aarch64_expand_epilogue): Authenticate return address fetched from
23872 (aarch64_override_options): Sanity check for ILP32 and ISA level.
23873 (aarch64_attributes): New function attributes for "sign-return-address".
23874 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
23875 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
23876 ("*do_return"): Generate combined instructions according to key index.
23877 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
23878 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
23880 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
23881 * config/aarch64/aarch64.opt (msign-return-address=): New.
23882 * doc/extend.texi (AArch64 Function Attributes): Documents
23883 "sign-return-address=".
23884 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
23886 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
23888 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
23889 overall option summary.
23891 2017-01-19 Jiong Wang <jiong.wang@arm.com>
23893 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
23894 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
23895 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
23896 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
23898 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
23900 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
23901 -mpower9-minmax by default for -mcpu=power9.
23902 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
23903 128-bit floating point.
23905 2017-01-20 Alan Modra <amodra@gmail.com>
23907 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
23908 optimizing for size.
23910 2017-01-20 Alan Modra <amodra@gmail.com>
23913 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
23914 for strcmp and strncmp from corresponding builtin decl.
23916 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
23918 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
23919 instead of i386/rtems-64.h.
23920 * config/i386/rtems-64.h: Remove.
23922 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
23926 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
23928 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
23930 2017-01-19 Tamar Christina <tamar.christina@arm.com>
23932 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
23933 Change int to HOST_WIDE_INT.
23934 * config/aarch64/aarch64-protos.h
23935 (aarch64_simd_gen_const_vector_dup): Likewise.
23936 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
23938 2017-01-19 David Malcolm <dmalcolm@redhat.com>
23940 * langhooks-def.h (lhd_type_for_size): New decl.
23941 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
23942 * langhooks.c (lhd_type_for_size): New function, taken from
23945 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
23947 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
23948 define_bypass for CR latency.
23949 (power9-cracked-alu): Update bypass latency and remove power9-branch.
23950 (power9-alu2): Add define_bypass for CR latency.
23952 (power9-mul): Update insn latency.
23953 (power9-mul-compare): Update insn latency, bypass latency and remove
23956 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23958 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
23960 * config/aarch64/aarch64.md
23961 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
23962 aarch64_nopcrelative_literal_loads.
23963 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
23965 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
23967 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
23968 TARGET_LOONGSON_3A.
23969 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
23971 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
23974 * config.gcc (supported_defaults): Add lxc1-sxc1.
23975 (with_lxc1_sxc1): Add validation.
23976 (all_defaults): Add lxc1-sxc1.
23977 * config/mips/mips.opt (mlxc1-sxc1): New option.
23978 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
23980 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
23981 __mips_no_lxc1_sxc1.
23982 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
23983 * doc/invoke.texi (-mlxc1-sxc1): Document the new option.
23984 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
23986 2017-01-19 Richard Biener <rguenther@suse.de>
23988 PR tree-optimization/72488
23989 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
23990 sure to restore SSA info.
23991 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
23993 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
23995 PR rtl-optimization/79121
23996 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
23997 of the inner type when shifting an extended value.
23999 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
24002 * symtab.c (symtab_node::equal_address_to): Fix comparing of
24003 interposable aliases.
24005 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
24008 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
24009 Use the evmergelohi instruction.
24010 (mov_si<mode>_e500_subreg4_2_le): Likewise.
24011 (mov_sitf_e500_subreg8_2_be): Likewise.
24012 (mov_sitf_e500_subreg12_2_le): Likewise.
24013 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
24014 (mov_si<mode>_e500_subreg4_2_be): Likewise.
24015 (mov_sitf_e500_subreg8_2_le): Likewise.
24016 (mov_sitf_e500_subreg12_2_be): Likewise.
24018 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24020 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
24021 attribute from vecsimple to vecperm.
24022 (altivec_vbpermq2): Likewise.
24024 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24027 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
24029 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
24030 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
24031 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
24032 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
24033 case where N arg is SIZE_MAX.
24034 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
24035 (cmpstrsi): Add pattern.
24037 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
24039 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
24040 __builtin_vec_revb builtins.
24041 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
24042 built-in functions to support generation of the ISA 3.0 XXBR<x>
24043 vector byte reverse instructions.
24044 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
24045 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
24046 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
24047 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
24048 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
24049 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
24050 (P9V_BUILTIN_VEC_REVB): Likewise.
24051 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
24052 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
24053 (p9_xxbrq_v16qi): Likewise.
24054 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
24055 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
24056 (p9_xxbrh_v8hi): Likewise.
24057 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
24058 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
24059 vec_revb built-in functions.
24061 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
24063 PR rtl-optimization/78952
24064 * config/i386/i386.md (any_extract): New code iterator.
24065 (*insvqi_2): Use any_extract for source operand.
24066 (*insvqi_3): Use any_shiftrt for source operand.
24068 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
24070 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
24072 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
24074 2017-01-18 Matthias Klose <doko@ubuntu.com>
24076 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
24078 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24080 * config/rs6000/altivec.h (vec_bperm): Change #define.
24081 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
24082 (altivec_vbpermq2): New define_insn.
24083 (altivec_vbpermd): Likewise.
24084 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
24085 function interface.
24086 (VBPERMD): Likewise.
24087 (VBPERM): New polymorphic function interface.
24088 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
24089 Add entries for P9V_BUILTIN_VEC_VBPERM.
24090 * doc/extend.texi: Add interfaces for vec_bperm.
24092 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
24094 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
24095 first letter of error messages.
24096 (s390_resolve_overloaded_builtin): Likewise.
24097 * config/s390/s390.c (s390_expand_builtin): Likewise.
24098 (s390_invalid_arg_for_unprototyped_fn): Likewise.
24099 (s390_valid_target_attribute_inner_p): Likewise.
24100 * config/s390/s390.md ("tabort"): Likewise.
24102 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
24104 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
24105 (ISA_AVOID_DIV_HILO): New macro.
24106 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
24107 (ISA_HAS_DDIV): Likewise.
24109 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
24111 * doc/invoke.texi (fabi-version): Correct number of occurrences.
24113 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
24115 * doc/invoke.texi (fabi-version): Spelling fix.
24117 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
24120 * doc/invoke.texi (fabi-version): Mention mangling fix for
24123 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
24126 * doc/invoke.texi (fabi-version): Document discriminator mangling.
24128 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
24131 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
24132 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
24134 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
24136 (stack_protect_test): Ditto.
24137 * config/rs6000/rs6000.opt (mstack-protector-guard=,
24138 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
24140 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
24141 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
24142 -mstack-protector-guard-offset=.
24143 (RS/6000 and PowerPC Options): Ditto.
24145 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
24147 * config/i386/i386.h (MASK_CLASS_P): New define.
24148 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
24149 there are no registers from different register sets also when
24150 mask registers are used. Update function comment.
24151 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
24152 to (*k/*r) and (*k/*km) alternatives.
24154 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
24156 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
24157 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
24158 (EH_RETURN_HANDLER_RTX): New define.
24159 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
24160 Force frame pointer in EH return functions.
24161 (aarch64_expand_epilogue): Add barrier for eh_return.
24162 (aarch64_final_eh_return_addr): Remove.
24163 (aarch64_eh_return_handler_rtx): New function.
24164 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
24166 (aarch64_eh_return_handler_rtx): New prototype.
24168 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24170 * config/rs6000/altivec.h (vec_rlmi): New #define.
24171 (vec_vrlnm): Likewise.
24172 (vec_rlnm): Likewise.
24173 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
24174 (UNSPEC_VRLNM): Likewise.
24175 (VIlong): New mode iterator.
24176 (altivec_vrl<VI_char>mi): New define_insn.
24177 (altivec_vrl<VI_char>nm): Likewise.
24178 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
24180 (VRLDNM): Likewise.
24181 (RLNM): New polymorphic function entry.
24182 (VRLWMI): New monomorphic function entry.
24183 (VRLDMI): Likewise.
24184 (RLMI): New polymorphic function entry.
24185 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
24186 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
24187 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
24190 2017-01-17 Jakub Jelinek <jakub@redhat.com>
24193 * dwarf2out.c (field_byte_offset): Restore the
24194 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
24195 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
24196 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
24199 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
24202 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
24204 2017-01-17 Jakub Jelinek <jakub@redhat.com>
24207 * dwarf2out.c (add_data_member_location_attribute): For constant
24208 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
24209 instead of DW_AT_data_member_location, DW_AT_bit_offset and
24210 DW_AT_byte_size attributes.
24212 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
24214 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
24215 after forcing to constant memory when the code model is medium.
24217 2017-01-17 Julia Koval <julia.koval@intel.com>
24220 * config/i386/avx512fintrin.h
24221 (_mm512_i32gather_ps): Change __addr type to void const*.
24222 (_mm512_mask_i32gather_ps): Ditto.
24223 (_mm512_i32gather_pd): Ditto.
24224 (_mm512_mask_i32gather_pd): Ditto.
24225 (_mm512_i64gather_ps): Ditto.
24226 (_mm512_mask_i64gather_ps): Ditto.
24227 (_mm512_i64gather_pd): Ditto.
24228 (_mm512_mask_i64gather_pd): Ditto.
24229 (_mm512_i32gather_epi32): Ditto.
24230 (_mm512_mask_i32gather_epi32): Ditto.
24231 (_mm512_i32gather_epi64): Ditto.
24232 (_mm512_mask_i32gather_epi64): Ditto.
24233 (_mm512_i64gather_epi32): Ditto.
24234 (_mm512_mask_i64gather_epi32): Ditto.
24235 (_mm512_i64gather_epi64): Ditto.
24236 (_mm512_mask_i64gather_epi64): Ditto.
24237 (_mm512_i32scatter_ps): Change __addr type to void*.
24238 (_mm512_mask_i32scatter_ps): Ditto.
24239 (_mm512_i32scatter_pd): Ditto.
24240 (_mm512_mask_i32scatter_pd): Ditto.
24241 (_mm512_i64scatter_ps): Ditto.
24242 (_mm512_mask_i64scatter_ps): Ditto.
24243 (_mm512_i64scatter_pd): Ditto.
24244 (_mm512_mask_i64scatter_pd): Ditto.
24245 (_mm512_i32scatter_epi32): Ditto.
24246 (_mm512_mask_i32scatter_epi32): Ditto.
24247 (_mm512_i32scatter_epi64): Ditto.
24248 (_mm512_mask_i32scatter_epi64): Ditto.
24249 (_mm512_i64scatter_epi32): Ditto.
24250 (_mm512_mask_i64scatter_epi32): Ditto.
24251 (_mm512_i64scatter_epi64): Ditto.
24252 (_mm512_mask_i64scatter_epi64): Ditto.
24253 * config/i386/avx512pfintrin.h
24254 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
24255 (_mm512_mask_prefetch_i32gather_ps): Ditto.
24256 (_mm512_mask_prefetch_i64gather_pd): Ditto.
24257 (_mm512_mask_prefetch_i64gather_ps): Ditto.
24258 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
24259 (_mm512_prefetch_i32scatter_ps): Ditto.
24260 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
24261 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
24262 (_mm512_prefetch_i64scatter_pd): Ditto.
24263 (_mm512_prefetch_i64scatter_ps): Ditto.
24264 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
24265 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
24266 * config/i386/avx512vlintrin.h
24267 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
24268 (_mm_mmask_i32gather_ps): Ditto.
24269 (_mm256_mmask_i32gather_pd): Ditto.
24270 (_mm_mmask_i32gather_pd): Ditto.
24271 (_mm256_mmask_i64gather_ps): Ditto.
24272 (_mm_mmask_i64gather_ps): Ditto.
24273 (_mm256_mmask_i64gather_pd): Ditto.
24274 (_mm_mmask_i64gather_pd): Ditto.
24275 (_mm256_mmask_i32gather_epi32): Ditto.
24276 (_mm_mmask_i32gather_epi32): Ditto.
24277 (_mm256_mmask_i32gather_epi64): Ditto.
24278 (_mm_mmask_i32gather_epi64): Ditto.
24279 (_mm256_mmask_i64gather_epi32): Ditto.
24280 (_mm_mmask_i64gather_epi32): Ditto.
24281 (_mm256_mmask_i64gather_epi64): Ditto.
24282 (_mm_mmask_i64gather_epi64): Ditto.
24283 (_mm256_i32scatter_ps): Change __addr type to void*.
24284 (_mm256_mask_i32scatter_ps): Ditto.
24285 (_mm_i32scatter_ps): Ditto.
24286 (_mm_mask_i32scatter_ps): Ditto.
24287 (_mm256_i32scatter_pd): Ditto.
24288 (_mm256_mask_i32scatter_pd): Ditto.
24289 (_mm_i32scatter_pd): Ditto.
24290 (_mm_mask_i32scatter_pd): Ditto.
24291 (_mm256_i64scatter_ps): Ditto.
24292 (_mm256_mask_i64scatter_ps): Ditto.
24293 (_mm_i64scatter_ps): Ditto.
24294 (_mm_mask_i64scatter_ps): Ditto.
24295 (_mm256_i64scatter_pd): Ditto.
24296 (_mm256_mask_i64scatter_pd): Ditto.
24297 (_mm_i64scatter_pd): Ditto.
24298 (_mm_mask_i64scatter_pd): Ditto.
24299 (_mm256_i32scatter_epi32): Ditto.
24300 (_mm256_mask_i32scatter_epi32): Ditto.
24301 (_mm_i32scatter_epi32): Ditto.
24302 (_mm_mask_i32scatter_epi32): Ditto.
24303 (_mm256_i32scatter_epi64): Ditto.
24304 (_mm256_mask_i32scatter_epi64): Ditto.
24305 (_mm_i32scatter_epi64): Ditto.
24306 (_mm_mask_i32scatter_epi64): Ditto.
24307 (_mm256_i64scatter_epi32): Ditto.
24308 (_mm256_mask_i64scatter_epi32): Ditto.
24309 (_mm_i64scatter_epi32): Ditto.
24310 (_mm_mask_i64scatter_epi32): Ditto.
24311 (_mm256_i64scatter_epi64): Ditto.
24312 (_mm256_mask_i64scatter_epi64): Ditto.
24313 (_mm_i64scatter_epi64): Ditto.
24314 (_mm_mask_i64scatter_epi64): Ditto.
24315 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
24316 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
24317 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
24318 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
24319 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
24320 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
24321 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
24322 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
24323 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
24324 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
24325 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
24326 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
24327 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
24328 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
24329 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
24330 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
24331 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
24332 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
24333 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
24334 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
24335 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
24336 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
24337 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
24338 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
24339 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
24340 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
24341 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
24342 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
24343 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
24344 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
24345 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
24346 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
24347 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
24348 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
24349 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
24350 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
24351 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
24352 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
24353 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
24354 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
24355 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
24356 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
24357 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
24358 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
24359 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
24360 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
24361 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
24362 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
24363 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
24364 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
24365 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
24366 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
24367 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
24368 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
24369 definitions accordingly.
24371 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
24372 Kuan-Lin Chen <kuanlinchentw@gmail.com>
24375 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
24378 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
24381 * ira-conflicts.c (ira_build_conflicts): Update total conflict
24382 hard regs for inner regno.
24384 2017-01-17 Martin Liska <mliska@suse.cz>
24387 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
24388 assumption and add comment.
24390 2017-01-17 Nathan Sidwell <nathan@acm.org>
24392 * ipa-visibility.c (localize_node): New function, broken out of ...
24393 (function_and_variable_visibility): ... here. Call it.
24395 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
24397 PR middle-end/77445
24398 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
24399 correctly set frequency of oudgoing edge.
24400 (duplicate_thread_path): Fix profile updating.
24402 2017-01-17 Jakub Jelinek <jakub@redhat.com>
24405 * configure.ac: Add GCC_BASE_VER.
24406 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
24407 version from BASE-VER file.
24408 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
24409 (gcc.o): Depend on $(BASEVER).
24410 * common.opt (dumpfullversion): New option.
24411 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
24412 * doc/invoke.texi: Document -dumpfullversion.
24413 * doc/install.texi: Document --with-gcc-major-version-only.
24414 * configure: Regenerated.
24416 2017-01-17 Richard Biener <rguenther@suse.de>
24418 PR tree-optimization/71433
24419 * tree-vrp.c (register_new_assert_for): Merge same asserts
24420 on all incoming edges.
24421 (process_assert_insertions_for): Handle insertions at the
24424 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
24426 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
24427 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
24429 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
24432 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
24435 2017-01-17 Alan Modra <amodra@gmail.com>
24438 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
24439 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
24440 symbolic stack limit when pic.
24442 2017-01-16 Martin Sebor <msebor@redhat.com>
24444 PR tree-optimization/78608
24445 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
24447 2017-01-16 Jeff Law <law@redhat.com>
24450 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
24451 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
24452 for several include directories that may be relative to sysroot.
24453 * config/i386/x-mingw32 (gplus_includedir): Define.
24454 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
24455 (native_system_includedir): Likewise.
24456 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
24457 override if TARGET_SYSTEM_ROOT is defined.
24458 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
24460 PR tree-optimization/79090
24461 PR tree-optimization/33562
24462 PR tree-optimization/61912
24463 PR tree-optimization/77485
24464 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
24465 and computed trims into the dump file.
24467 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
24469 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
24471 2017-01-16 Jakub Jelinek <jakub@redhat.com>
24474 * gimplify.c (gimplify_init_constructor): If want_value and
24475 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
24479 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
24480 sequence. Formatting fixes.
24481 (doloop_optimize): Formatting fixes.
24484 * gcc.c (debug_level_greater_than_spec_func): New function.
24485 (static_spec_functions): Add debug-level-gt spec function.
24486 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
24488 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
24489 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
24490 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
24491 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
24492 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
24493 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
24495 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
24497 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
24498 QImode fixups to general and mask registers only.
24500 2017-01-16 Carl Love <cel@us.ibm.com>
24502 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
24503 for built-in functions
24504 vector signed char vec_nabs (vector signed char)
24505 vector signed short vec_nabs (vector signed short)
24506 vector signed int vec_nabs (vector signed int)
24507 vector signed long long vec_nabs (vector signed long long)
24508 vector float vec_nabs (vector float)
24509 vector double vec_nabs (vector double)
24510 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
24512 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
24513 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
24514 * doc/extend.texi: Update the documentation file for the new built-in
24517 2017-01-16 Martin Sebor <msebor@redhat.com>
24519 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
24522 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24524 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
24525 UNSPEC_VSX__XXSPLTD to require special splat handling.
24527 2017-01-16 David Malcolm <dmalcolm@redhat.com>
24530 * system.h: Poison strndup.
24532 2017-01-16 Alan Modra <amodra@gmail.com>
24535 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
24538 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
24540 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
24542 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
24544 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
24545 call recog here. Assert that INSN_CODE (insn) is non-negative.
24547 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
24550 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
24552 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
24553 in the currently scheduled RTL fragment.
24555 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
24557 PR rtl-optimization/78751
24558 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
24561 2017-01-14 Jeff Law <law@redhat.com>
24563 PR tree-optimization/79090
24564 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
24565 variable length stores.
24566 (compute_trims): Delete dead assignment to *trim_tail.
24567 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
24570 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
24572 PR rtl-optimization/78626
24573 PR rtl-optimization/78727
24574 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
24575 of a block, and split such blocks after everything else is finished.
24577 2017-01-14 Alan Modra <amodra@gmail.com>
24580 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
24581 target legitimate_combined_insn.
24582 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
24583 (rs6000_legitimate_combined_insn): New function.
24584 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
24586 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
24587 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
24588 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
24590 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
24592 * doc/frontends.texi (G++ and GCC): Remove references to Java.
24594 2017-01-13 Jeff Law <law@redhat.com>
24596 PR tree-optimization/33562
24597 PR tree-optimization/61912
24598 PR tree-optimization/77485
24599 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
24601 (delete_dead_assignment): Likewise.
24602 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
24603 statement to delete_dead_call and delete_dead_assignment.
24605 2017-01-13 David Malcolm <dmalcolm@redhat.com>
24608 * substring-locations.c (format_warning_va): Strengthen case 1 so
24609 that both endpoints of the substring must be within the format
24610 range for just the substring to be printed.
24612 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
24614 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
24615 * config/i386/i386.c (ix86_target_string): Add missing options
24616 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
24617 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
24618 flags_other and ix86_target_other to flags2_other. Display unknown
24620 (ix86_valid_target_attribute_inner_p): Add missing options and
24621 reorder options by implied ISAs, as in ix86_target_string.
24623 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
24625 * hash-table.h (hash_table::too_empty_p): New function.
24626 (hash_table::expand): Use it.
24627 (hash_table::traverse): Likewise.
24628 (hash_table::empty_slot): Use sizeof (value_type) instead of
24629 sizeof (PTR) to convert bytes to elements. Shrink the table
24630 if the current size is excessive for the current number of
24633 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
24635 * ira-costs.c (record_reg_classes): Break from the inner loop
24636 early once alt_fail is known to be true. Update outer loop
24637 handling accordingly.
24639 2017-01-13 Jeff Law <law@redhat.com>
24641 * tree-ssa-dse.c (decrement_count): New function.
24642 (increment_start_addr, maybe_trim_memstar_call): Likewise.
24643 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
24644 when we know the partially dead statement is a mem* function.
24646 PR tree-optimization/61912
24647 PR tree-optimization/77485
24648 * tree-ssa-dse.c: Include expr.h.
24649 (maybe_trim_constructor_store): New function.
24650 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
24652 PR tree-optimization/33562
24653 PR tree-optimization/61912
24654 PR tree-optimization/77485
24655 * doc/invoke.texi: Document new dse-max-object-size param.
24656 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
24657 * tree-ssa-dse.c: Include params.h.
24658 (dse_store_status): New enum.
24659 (initialize_ao_ref_for_dse): New, partially extracted from
24661 (valid_ao_ref_for_dse, normalize_ref): New.
24662 (setup_live_bytes_from_ref, compute_trims): Likewise.
24663 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
24664 (maybe_trim_partially_dead_store): Likewise.
24665 (maybe_trim_complex_store): Likewise.
24666 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
24667 Track what bytes live from the original store. Return tri-state
24668 for dead, partially dead or live.
24669 (dse_dom_walker): Add constructor, destructor and new private members.
24670 (delete_dead_call, delete_dead_assignment): New extracted from
24672 (dse_optimize_stmt): Make a member of dse_dom_walker.
24673 Use initialize_ao_ref_for_dse.
24675 PR tree-optimization/33562
24676 PR tree-optimization/61912
24677 PR tree-optimization/77485
24678 * sbitmap.h (bitmap_count_bits): Prototype.
24679 (bitmap_clear_range, bitmap_set_range): Likewise.
24680 * sbitmap.c (bitmap_clear_range): New function.
24681 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
24683 2017-01-13 Martin Liska <mliska@suse.cz>
24686 * function.c (set_cfun): Add new argument force.
24687 * function.h (set_cfun): Likewise.
24688 * ipa-inline-transform.c (inline_call): Use the function when
24689 strict alising from is dropped for function we inline to.
24691 2017-01-13 Richard Biener <rguenther@suse.de>
24693 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
24694 for dumping GIMPLE INTEGER_CSTs.
24696 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24698 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
24699 to 201112L since C++17.
24701 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
24704 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
24705 if -fsanitize=kernel-address is present.
24707 2017-01-13 Richard Biener <rguenther@suse.de>
24709 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
24710 as _Literal ( type ) number in case usual suffixes do not
24711 preserve all information.
24713 2017-01-13 Richard Biener <rguenther@suse.de>
24715 PR tree-optimization/77283
24716 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
24717 and ssa-iterators.h.
24718 (is_feasible_trace): Implement a cost model based on joiner
24721 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
24724 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
24725 char or short to __float128/_Float128 directly.
24727 2017-01-12 Martin Sebor <msebor@redhat.com>
24729 to -Wformat-overflow.
24730 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
24731 (min_bytes_remaining): Same.
24732 (get_string_length): Same.
24733 (format_string): Same.
24734 (format_directive): Same.
24736 (pass_sprintf_length::handle_gimple_call): Same.
24738 2017-01-12 Jakub Jelinek <jakub@redhat.com>
24740 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
24741 info.nowrite calls with no lhs that can't throw. Return bool
24742 whether gsi_remove has been called or not.
24743 (pass_sprintf_length::handle_gimple_call): Return bool whether
24744 try_substitute_return_value called gsi_remove. Formatting fix.
24745 (pass_sprintf_length::execute): Don't use gsi_remove if
24746 handle_gimple_call returned true.
24749 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
24750 be removed due to side-effects, don't remove following barrier nor
24751 turn the successor edge into fallthru edge.
24753 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24756 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
24757 element-reversing loads and stores as not swappable.
24759 2017-01-12 Nathan Sidwell <nathan@acm.org>
24760 Nicolai Stange <nicstange@gmail.com>
24762 * combine.c (try_combine): Don't ignore result of overlap checking
24763 loop. Combine overlap & asm check into single loop.
24765 2017-01-12 Richard Biener <rguenther@suse.de>
24767 * tree-pretty-print.c (dump_generic_node): Provide -gimple
24768 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
24770 2017-01-12 Richard Biener <rguenther@suse.de>
24772 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
24773 and TS_TARGET_OPTION directly derive from TS_BASE.
24774 * tree-core.h (tree_optimization_option): Derive from tree_base.
24775 (tree_target_option): Likewise.
24777 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
24779 * config/i386/i386.c (memory_address_length): Increase len
24780 only when rip_relative_addr_p returns false.
24782 2017-01-11 Julia Koval <julia.koval@intel.com>
24784 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
24785 (OPTION_MASK_ISA_SGX_SET): New.
24786 (ix86_handle_option): Handle OPT_msgx.
24787 * config.gcc: Added sgxintrin.h.
24788 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
24789 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
24790 * config/i386/i386.c (ix86_target_string): Add -msgx.
24792 (ix86_option_override_internal): Handle new options.
24793 (ix86_valid_target_attribute_inner_p): Add sgx.
24794 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
24795 * config/i386/i386.opt: Add msgx.
24796 * config/i386/sgxintrin.h: New file.
24797 * config/i386/x86intrin.h: Add sgxintrin.h.
24799 2017-01-11 Jakub Jelinek <jakub@redhat.com>
24802 * fold-const.c (maybe_nonzero_address): Return 1 for function
24804 (tree_single_nonzero_warnv_p): Don't handle function local objects
24808 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
24811 2017-01-11 David Malcolm <dmalcolm@redhat.com>
24814 * opts.c: Include "spellcheck.h"
24815 (struct string_fragment): New struct.
24816 (struct edit_distance_traits<const string_fragment &>): New
24818 (get_closest_sanitizer_option): New function.
24819 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
24821 2017-01-11 Jakub Jelinek <jakub@redhat.com>
24823 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
24825 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
24826 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
24827 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
24828 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
24829 for initial die_offset if dwarf_split_debug_info.
24830 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
24831 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
24833 (output_skeleton_debug_sections): Formatting fix. Use
24834 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
24835 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
24837 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
24839 * config/arm/cortex-a53.md: Add bypasses for
24840 cortex_a53_r2f_cvt.
24841 (cortex_a53_r2f): Only use for transfers.
24842 (cortex_a53_f2r): Likewise.
24843 (cortex_a53_r2f_cvt): Add reservation for conversions.
24844 (cortex_a53_f2r_cvt): Likewise.
24846 2017-01-11 Tamar Christina <tamar.christina@arm.com>
24848 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
24849 to all inlined functions, change static to extern.
24851 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
24854 * config/arm/arm.c (legitimize_pic_address): Handle reference to
24856 (arm_assemble_integer): Likewise.
24858 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
24860 * config.gcc: Use new awk script to check CPU, FPU and architecture
24861 parameters for --with-... options.
24862 * config/arm/parsecpu.awk: New file
24863 * config/arm/arm-cpus.in: New file.
24864 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
24866 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
24868 * config/arm/t-arm: Update dependency rules.
24869 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
24870 of processing .def files.
24871 * config/arm/genopt.sh: Deleted.
24872 * config/arm/gentune.sh: Deleted.
24873 * config/arm/arm-cores.def: Deleted.
24874 * config/arm/arm-arches.def: Deleted.
24875 * config/arm/arm-fpus.def: Deleted.
24876 * config/arm/arm-tune.md: Regenerated.
24877 * config/arm/arm-tables.opt: Regenerated.
24878 * config/arm/arm-cpu.h: New generated file.
24879 * config/arm/arm-cpu-data.h: New generated file.
24880 * config/arm/arm-cpu-cdata.h: New generated file.
24882 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
24885 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
24887 (input_varpool_node): Unpack dynamically_initialized bit.
24889 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
24891 PR rtl-optimization/79032
24892 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
24893 the alignment of the adjusted memory reference against that of MODE,
24894 instead of the alignment of the original memory reference.
24896 2017-01-11 Martin Jambor <mjambor@suse.cz>
24898 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
24900 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
24901 decorated functions.
24903 2017-01-11 Richard Biener <rguenther@suse.de>
24905 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
24906 set range/nonnull info for PHI results. Do not set it on
24907 stmts marked for removal.
24909 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
24911 * expr.c (store_field): In the bitfield case, fetch the return value
24912 from the registers before applying a single big-endian adjustment.
24913 Always do a final load for a BLKmode value not larger than a word.
24915 2017-01-10 David Malcolm <dmalcolm@redhat.com>
24918 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
24919 that we correctly handle column numbers greater than
24920 LINE_MAP_MAX_COLUMN_NUMBER.
24922 2017-01-10 Martin Sebor <msebor@redhat.com>
24924 PR middle-end/78245
24925 * gimple-ssa-sprintf.c (get_destination_size): Call
24926 {init,fini}object_sizes.
24927 * tree-object-size.c (addr_object_size): Adjust.
24928 (pass_through_call): Adjust.
24929 (pass_object_sizes::execute): Adjust.
24930 * tree-object-size.h (fini_object_sizes): Declare.
24932 2017-01-10 Martin Sebor <msebor@redhat.com>
24934 PR tree-optimization/78775
24935 * builtins.c (get_size_range): Move...
24936 * calls.c: ...to here.
24937 (alloc_max_size): Accept zero argument.
24938 (operand_signed_p): Remove.
24939 (maybe_warn_alloc_args_overflow): Call get_size_range.
24940 * calls.h (get_size_range): Declare.
24942 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
24944 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
24945 from TI's devices.csv file as of September 2016.
24946 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
24948 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
24950 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
24951 * doc/invoke.texi: Likewise.
24952 * doc/md.texi: Likewise.
24953 * doc/objc.texi: Likewise.
24955 2017-01-10 Joshua Conner <joshconner@google.com>
24957 * config/arm/fuchsia-elf.h: New file.
24958 * config/fuchsia.h: New file.
24959 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
24960 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
24962 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
24964 2016-01-10 Richard Biener <rguenther@suse.de>
24966 PR tree-optimization/79034
24967 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
24968 Propagate out degenerate PHIs in the joiner.
24970 2017-01-10 Martin Liska <mliska@suse.cz>
24972 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
24973 (sort_congruence_classes_by_decl_uid): Likewise.
24974 (sort_congruence_class_groups_by_decl_uid): Likewise.
24975 (sem_item_optimizer::merge_classes): Sort class, groups in these
24976 classes and members in the groups by DECL_UID of declarations.
24977 This would make merge operations stable.
24979 2017-01-10 Martin Liska <mliska@suse.cz>
24981 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
24982 usage of m_classes_vec.
24983 (sem_item_optimizer::~sem_item_optimizer): Likewise.
24984 (sem_item_optimizer::get_group_by_hash): Likewise.
24985 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
24986 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
24987 (sem_item_optimizer::verify_classes): Likewise.
24988 (sem_item_optimizer::process_cong_reduction): Likewise.
24989 (sem_item_optimizer::dump_cong_classes): Likewise.
24990 (sem_item_optimizer::merge_classes): Likewise.
24991 * ipa-icf.h (congruence_class_hash): Rename from
24992 congruence_class_group_hash. Remove declaration of m_classes_vec.
24994 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
24996 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
24997 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
24998 * config.gcc: Add avx512vpopcntdqintrin.h.
24999 * config/i386/avx512vpopcntdqintrin.h: New.
25000 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
25001 * config/i386/i386-builtin-types.def: Add new types.
25002 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
25003 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
25004 __builtin_ia32_vpopcountq_v8di_mask): New.
25005 * config/i386/i386-c.c (ix86_target_macros_internal): Define
25006 __AVX512VPOPCNTDQ__.
25007 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
25008 (PTA_AVX512VPOPCNTDQ): Define.
25009 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
25010 TARGET_AVX512VPOPCNTDQ_P): Define.
25011 * config/i386/i386.opt: Add mavx512vpopcntdq.
25012 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
25013 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
25015 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
25017 PR middle-end/77484
25018 * predict.def (PRED_CALL): Set to 67.
25020 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
25022 * expr.c (store_field): In the bitfield case, if the value comes from
25023 a function call and is of an aggregate type returned in registers, do
25024 not modify the field mode; extract the value in all cases if the mode
25025 is BLKmode and the size is not larger than a word.
25027 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
25030 * config/i386/cpuid.h: Fix undefined behavior.
25032 2017-01-04 Jeff Law <law@redhat.com>
25034 PR tree-optimization/79007
25035 PR tree-optimization/67955
25036 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
25037 conservative for pt.null when flag_non_call_exceptions is on.
25039 2017-01-09 Jakub Jelinek <jakub@redhat.com>
25041 PR translation/79019
25042 PR translation/79020
25043 * params.def (PARAM_INLINE_MIN_SPEEDUP,
25044 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
25045 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
25047 * config/avr/avr.opt (maccumulate-args): Likewise.
25048 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
25049 * common.opt (freport-bug): Likewise.
25050 * cif-code.def (CIF_FINAL_ERROR): Likewise.
25051 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
25052 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
25053 translatable string.
25054 * config/i386/i386.c (function_value_32): Likewise.
25055 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
25056 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
25058 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
25059 * common/config/msp430/msp430-common.c (msp430_handle_option):
25061 * symtab.c (symtab_node::verify_base): Likewise.
25062 * opts.c (set_debug_level): Likewise.
25063 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
25064 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
25065 missing whitespace to translatable strings.
25066 * config/avr/avr.md (bswapsi2): Fix typo in comment.
25067 * config/sh/superh.h: Likewise.
25068 * config/i386/xopintrin.h: Likewise.
25069 * config/i386/znver1.md: Likewise.
25070 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
25071 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
25072 * double-int.h (struct double_int): Likewise.
25073 * double-int.c (div_and_round_double): Likewise.
25074 * wide-int.cc: Likewise.
25075 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
25076 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
25077 * cfgcleanup.c (crossjumps_occured): Renamed to ...
25078 (crossjumps_occurred): ... this.
25079 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
25082 PR tree-optimization/78899
25083 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
25084 returning bool return struct loop *, NULL for failure and the new
25086 (versionable_outer_loop_p): Don't version outer loop if it has
25087 dont_vectorized bit set.
25088 (tree_if_conversion): When versioning outer loop, ensure
25089 tree_if_conversion is performed also on the inner loop of the
25090 non-vectorizable outer loop copy.
25091 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
25092 LOOP_VECTORIZED in inner loop of the scalar outer loop and
25093 prevent vectorization of it.
25094 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
25095 the outer loop vectorization of the non-scalar version is attempted
25096 before vectorization of the inner loop in scalar version. If
25097 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
25098 vectorization of its inner loop.
25099 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
25100 has 2 inner loops, rename also on edges from bb whose single pred
25101 is outer_loop->header. Fix typo in function comment.
25103 2017-01-09 Martin Sebor <msebor@redhat.com>
25106 * asan.c (asan_emit_stack_protection): Increase local buffer size
25107 to avoid snprintf truncation warning.
25109 2017-01-09 Andrew Pinski <apinski@cavium.com>
25111 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
25112 to reference thunderx2t99 for the tuning structure
25113 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
25115 (thunderx2t99_extra_costs): This.
25116 * config/aarch64/aarch64-tune.md: Regenerate.
25117 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
25118 (vulcan_addrcost_table): This.
25119 (vulcan_regmove_cost): Rename to ...
25120 (thunderx2t99_regmove_cost): This.
25121 (vulcan_vector_cost): Rename to ...
25122 (thunderx2t99_vector_cost): this.
25123 (vulcan_branch_cost): Rename to ...
25124 (thunderx2t99_branch_cost): This.
25125 (vulcan_tunings): Rename to ...
25126 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
25127 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
25129 2017-01-09 Martin Jambor <mjambor@suse.cz>
25133 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
25134 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
25135 (propagate_vr_accross_jump_function): Use the above function for all
25136 value range computations for pass-through jump functions and type
25137 converasion from explicit value range values.
25138 (ipcp_propagate_stage): Do not attempt to deduce types of formal
25139 parameters from TYPE_ARG_TYPES.
25140 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
25141 (ipa_write_node_info): Stream type of the actual argument.
25142 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
25144 2017-01-09 Martin Liska <mliska@suse.cz>
25147 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
25148 (lookup_compiler): Do not show error message with have_E.
25150 2017-01-09 Jakub Jelinek <jakub@redhat.com>
25152 PR tree-optimization/78938
25153 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
25154 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
25155 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
25156 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
25159 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25161 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
25164 2017-01-09 Richard Biener <rguenther@suse.de>
25166 PR tree-optimization/78997
25167 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
25168 name condition properly.
25170 2017-01-09 Richard Biener <rguenther@suse.de>
25173 * dwarf2out.c (is_cxx): New overload with context.
25174 (is_naming_typedef_decl): Use it.
25176 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
25178 * invoke.texi (Option Summary): Correct spacing in option lists
25179 and add line breaks to fix over-long lines.
25181 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
25183 PR middle-end/17660
25185 * extend.texi (Common Variable Attributes): Add xref to GCC
25186 Internals manual to explain mode attribute keywords.
25188 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
25191 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
25192 and Preprocessor Options.
25193 (Options for Linking): Document -pthread here....
25194 (RS/6000 and PowerPC Options): ...not here.
25195 (Solaris 2 Options): ...or here.
25196 * doc/cppopts.texi: Document -pthread.
25198 2017-01-08 Martin Sebor <msebor@redhat.com>
25200 PR middle-end/77708
25201 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
25202 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
25203 New member functions.
25204 (format_directive): Used them.
25206 (pass_sprintf_length::handle_gimple_call): Same.
25207 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
25208 to avoid truncation for any argument.
25209 (extract_affine_mul): Same.
25210 * tree.c (get_file_function_name): Same.
25212 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
25214 PR middle-end/77484
25215 * predict.def (PRED_INDIR_CALL): Set to 86.
25217 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
25219 PR preprocessor/54124
25220 * doc/cppopts.texi: Reformat -d subtable to list the full name
25221 of the options. Add cross-reference to the docs for the general
25222 compiler -d options.
25223 * doc/invoke.texi (Developer Options): Add cross-reference to the
25224 preprocessor-specific -d option documentation.
25226 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
25228 PR preprocessor/13498
25229 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
25230 redudant material, and reflect new command-line options.
25231 (System Headers): Likewise.
25233 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
25235 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
25236 -isystem, and -idirafter. Copy-edit.
25237 * doc/cppopts.texi: Copy-edit. Remove contradiction about
25238 default for -ftrack-macro-expansion. Delete obsolete and
25239 badly-formatted implementation details about -fdebug-cpp output.
25240 * doc/cppwarnopts.texi: Copy-edit.
25242 2017-01-07 David Malcolm <dmalcolm@redhat.com>
25245 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
25246 that the transition from a max line width >= 1<<10 to narrower
25247 lines works correctly.
25249 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
25251 * doc/options.texi (PerFunction): New.
25252 * opt-functions.awk (switch_flags): Map both Optimization and
25253 PerFunction to CL_OPTIMIZATION.
25254 * opth-gen.awk: Test for PerFunction flag along with
25256 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
25257 it only when the latter is present. Skip those that don't in
25258 the hash function generator.
25259 * common.opt (fvar-tracking): Mark as PerFunction instead of
25261 (fvar-tracking-assignments): Likewise.
25262 (fvar-tracking-assignments-toggle): Likewise.
25263 (fvar-tracking-uninit): Likewise.
25265 2017-01-07 Jakub Jelinek <jakub@redhat.com>
25267 PR translation/79018
25268 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
25271 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
25274 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
25275 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
25276 TARGET_LONG_JUMP_TABLE_OFFSETS.
25277 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
25278 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
25279 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
25280 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
25281 * config/m68k/m68k.md (tablejump expander): Likewise.
25282 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
25283 TARGET_LONG_JUMP_TABLE_OFFSETS.
25284 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
25285 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
25287 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25288 David Holsgrove <david.holsgrove@xilinx.com>
25290 * common/config/microblaze/microblaze-common.c
25291 (TARGET_EXCEPT_UNWIND_INFO): Remove.
25292 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
25294 * config/microblaze/microblaze.c (microblaze_must_save_register)
25295 (microblaze_expand_epilogue, microblaze_return_addr): Handle
25297 (microblaze_eh_return): New function.
25298 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
25299 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
25300 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
25301 * config/microblaze/microblaze.md (eh_return): New pattern.
25303 2017-01-06 Jakub Jelinek <jakub@redhat.com>
25305 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
25306 GCC_DIAGNOSTIC_STRINGIFY): Define.
25308 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
25310 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25312 * config/arm/arm.md (<mcrr>): New.
25314 * config/arm/arm.c (arm_arch5te): New.
25315 (arm_option_override): Set arm_arch5te.
25316 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
25318 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
25319 (arm_mcrr_qualifiers): ... this. New.
25320 (MRRC_QUALIFIERS): Define to...
25321 (arm_mrrc_qualifiers): ... this. New.
25322 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
25324 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
25325 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
25326 (MRRCI, mrrc, MRRC): New.
25327 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
25328 VUNSPEC_MRRC2): New.
25330 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25332 * config/arm/arm.md (<mcr>): New.
25334 * config/arm/arm.c (arm_coproc_builtin_available): Add
25335 support for mcr, mrc, mcr2 and mrc2.
25336 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
25337 (arm_mcr_qualifiers): ... this. New.
25338 (MRC_QUALIFIERS): Define to ...
25339 (arm_mrc_qualifiers): ... this. New.
25340 (MCR_QUALIFIERS): Define to ...
25341 (arm_mcr_qualifiers): ... this. New.
25342 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
25344 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
25345 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
25346 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
25347 VUNSPEC_MRC2): New.
25349 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25351 * config/arm/arm.md (*ldc): New.
25355 * config/arm/arm.c (arm_coproc_builtin_available): Add
25356 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
25357 (arm_coproc_ldc_stc_legitimate_address): New.
25358 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
25359 'qualifier_const_pointer'.
25360 (LDC_QUALIFIERS): Define to...
25361 (arm_ldc_qualifiers): ... this. New.
25362 (STC_QUALIFIERS): Define to...
25363 (arm_stc_qualifiers): ... this. New.
25364 * config/arm/arm-protos.h
25365 (arm_coproc_ldc_stc_legitimate_address): New.
25366 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
25367 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
25368 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
25369 stc2, stcl, stc2l): New.
25370 * config/arm/constraints.md (Uz): New.
25371 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
25372 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
25373 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
25374 VUNSPEC_STC2L): New.
25376 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25378 * config/arm/arm.md (<cdp>): New.
25379 * config/arm/arm.c (neon_const_bounds): Rename this ...
25380 (arm_const_bounds): ... this.
25381 (arm_coproc_builtin_available): New.
25382 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
25383 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
25384 (CDP_QUALIFIERS): Define to...
25385 (arm_cdp_qualifiers): ... this. New.
25387 (arm_expand_builtin_args): Add case for 6 arguments.
25388 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
25389 (arm_const_bounds): ... this.
25390 (arm_coproc_builtin_available): New.
25391 * config/arm/arm_acle.h (__arm_cdp): New.
25393 * config/arm/arm_acle_builtins.def (cdp): New.
25395 * config/arm/iterators.md (CDPI,CDP,cdp): New.
25396 * config/arm/neon.md: Rename all 'neon_const_bounds' to
25397 'arm_const_bounds'.
25398 * config/arm/types.md (coproc): New.
25399 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
25400 * doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
25401 * doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
25402 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
25404 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25406 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
25407 (UBINOP_QUALIFIERS): New.
25409 (acle_builtin_data): New. Change comment.
25410 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
25411 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
25412 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
25413 arm_acle_builtins.def.
25414 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
25415 (arm_init_acle_builtins): New.
25416 (CRC32_BUILTIN): Remove.
25417 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
25418 crc32cb, crc32ch and crc32cw.
25419 (arm_init_crc32_builtins): Remove.
25420 (arm_init_builtins): Use arm_init_acle_builtins rather
25421 than arm_init_crc32_builtins.
25422 (arm_expand_acle_builtin): New.
25423 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
25424 * config/arm/arm_acle_builtins.def: New.
25426 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
25428 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
25429 (arm_builtin_datum): ... this.
25430 (arm_init_neon_builtin): Rename to ...
25431 (arm_init_builtin): ... this. Add a new parameters PREFIX
25432 and USE_SIG_IN_NAME.
25433 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
25434 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
25435 'arm_builtin_datum'.
25436 (arm_init_vfp_builtins): Likewise.
25437 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
25438 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
25439 (arm_expand_neon_args): Rename to ...
25440 (arm_expand_builtin_args): ... this. Rename builtin_arg
25441 enum values and differentiate between ARG_BUILTIN_MEMORY
25442 and ARG_BUILTIN_NEON_MEMORY.
25443 (arm_expand_neon_builtin_1): Rename to ...
25444 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
25445 values, arm_expand_builtin_args and add bool parameter NEON.
25446 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
25447 (arm_expand_vfp_builtin): Likewise.
25448 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
25450 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
25452 PR middle-end/77484
25453 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
25454 * predict.c (tree_estimate_probability_bb): Reverse direction of
25455 polymorphic call predictor.
25457 2017-01-06 David Malcolm <dmalcolm@redhat.com>
25459 * passes.c (execute_one_pass): Split out pass-skipping logic into...
25460 (determine_pass_name_match): ...this new function and...
25461 (should_skip_pass_p): ...this new function.
25463 2017-01-06 Nathan Sidwell <nathan@acm.org>
25465 * ipa-visibility.c (function_and_variable_visibility): Reformat
25466 comments and long lines. Remove extrneous if.
25467 * symtab.c (symtab_node::make_decl_local): Fix code format.
25468 (symtab_node::set_section_for_node): Fix comment typo.
25470 2017-01-06 Martin Liska <mliska@suse.cz>
25473 * lra-constraints.c: Rename invariant to lra_invariant.
25474 * predict.c (set_even_probabilities): Initialize e to NULL.
25476 2017-01-05 Martin Sebor <msebor@redhat.com>
25478 PR tree-optimization/78910
25479 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
25480 (format_integer): Correct off-by-one error in the handling
25481 of precision with negative numbers in signed conversions..
25483 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
25485 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
25487 2017-01-05 Jakub Jelinek <jakub@redhat.com>
25489 PR tree-optimization/71016
25490 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
25491 factor_out_conditional_conversion. Formatting fix.
25492 (factor_out_conditional_conversion): Add cond_stmt argument.
25493 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
25494 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
25497 2017-01-05 David Malcolm <dmalcolm@redhat.com>
25499 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
25500 read-rtl-function.o, and selftest-rtl.o.
25501 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
25502 (selftest::aarch64_test_loading_full_dump): New function.
25503 (selftest::aarch64_run_selftests): New function.
25504 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
25505 selftest::aarch64_run_selftests.
25506 * config/i386/i386.c
25507 (selftest::ix86_test_loading_dump_fragment_1): New function.
25508 (selftest::ix86_test_loading_call_insn): New function.
25509 (selftest::ix86_test_loading_full_dump): New function.
25510 (selftest::ix86_test_loading_unspec): New function.
25511 (selftest::ix86_run_selftests): Call the new functions.
25512 * emit-rtl.c (maybe_set_max_label_num): New function.
25513 * emit-rtl.h (maybe_set_max_label_num): New decl.
25514 * function.c (instantiate_decls): Guard call to
25515 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
25516 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
25518 * gensupport.c (gen_reader::gen_reader): Pass "false"
25519 for new "compact" param of rtx_reader.
25520 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
25521 rather than an empty string for NULL strings.
25522 * read-md.c: Potentially include config.h rather than bconfig.h.
25523 Wrap include of errors.h with #ifdef GENERATOR_FILE.
25524 (have_error): New global, copied from errors.c.
25525 (md_reader::read_name): Rename to...
25526 (md_reader::read_name_1): ...this, adding "out_loc" param,
25527 and converting "missing name or number" to returning false, rather
25529 (md_reader::read_name): Reimplement in terms of read_name_1.
25530 (md_reader::read_name_or_nil): New function.
25531 (md_reader::read_string): Handle "(nil)" by returning NULL.
25532 (md_reader::md_reader): Add new param "compact".
25533 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
25534 (md_reader::read_file): New method.
25535 * read-md.h (md_reader::md_reader): Add new param "compact".
25536 (md_reader::read_file): New method.
25537 (md_reader::is_compact): New accessor.
25538 (md_reader::read_name): Convert return type from void to file_location.
25539 (md_reader::read_name_or_nil): New decl.
25540 (md_reader::read_name_1): New decl.
25541 (md_reader::m_compact): New field.
25542 (noop_reader::noop_reader): Pass "false" for new "compact" param
25544 (rtx_reader::rtx_reader): Add new "compact" param.
25545 (rtx_reader::read_rtx_operand): Make virtual and convert return
25546 type from void to rtx.
25547 (rtx_reader::read_until): New decl.
25548 (rtx_reader::handle_any_trailing_information): New virtual function.
25549 (rtx_reader::postprocess): New virtual function.
25550 (rtx_reader::finalize_string): New virtual function.
25551 (rtx_reader::m_in_call_function_usage): New field.
25552 (rtx_reader::m_reuse_rtx_by_id): New field.
25553 * read-rtl-function.c: New file.
25554 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
25555 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
25556 (selftest::verify_three_block_rtl_cfg): New decl.
25557 * read-rtl-function.h: New file.
25558 * read-rtl.c: Potentially include config.h rather than bconfig.h.
25559 For host, include function.h, memmodel.h, and emit-rtl.h.
25560 (one_time_initialization): New function.
25561 (struct compact_insn_name): New struct.
25562 (compact_insn_names): New array.
25563 (find_code): Handle insn codes in compact dumps.
25564 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
25565 (bind_subst_iter_and_attr): Likewise.
25566 (add_condition_to_string): Likewise.
25567 (add_condition_to_rtx): Likewise.
25568 (apply_attribute_uses): Likewise.
25569 (add_current_iterators): Likewise.
25570 (apply_iterators): Likewise.
25571 (initialize_iterators): Guard usage of apply_subst_iterator with
25572 #ifdef GENERATOR_FILE.
25573 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
25574 (md_reader::read_mapping): Likewise.
25575 (add_define_attr_for_define_subst): Likewise.
25576 (add_define_subst_attr): Likewise.
25577 (read_subst_mapping): Likewise.
25578 (check_code_iterator): Likewise.
25579 (rtx_reader::read_rtx): Likewise. Move one-time initialization
25581 (one_time_initialization): New function.
25582 (rtx_reader::read_until): New method.
25583 (read_flags): New function.
25584 (parse_reg_note_name): New function.
25585 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
25586 Handle reuse_rtx ids.
25587 Wrap iterator lookup within #ifdef GENERATOR_FILE.
25588 Add parsing support for RTL dumps, mirroring the special-cases in
25589 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
25590 values, and calling handle_any_trailing_information.
25591 (rtx_reader::read_rtx_operand): Convert return type from void
25592 to rtx, returning return_rtx. Handle case 'e'. Call
25593 finalize_string on XSTR and XTMPL fields.
25594 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
25595 "(nil)" values were omitted. Call the postprocess vfunc on the
25597 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
25598 class ctor. Initialize m_in_call_function_usage. Call
25599 one_time_initialization.
25600 * rtl-tests.c (selftest::test_uncond_jump): Call
25601 set_new_first_and_last_insn.
25602 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
25603 * selftest-rtl.c: New file.
25604 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
25605 (selftest::get_insn_by_uid): New decl.
25606 * selftest-run-tests.c (selftest::run_tests): Call
25607 read_rtl_function_c_tests.
25608 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
25609 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
25612 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
25614 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
25615 operands in a special way. Assert that pos+len <= mode precision.
25617 2017-01-05 Jakub Jelinek <jakub@redhat.com>
25619 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
25620 3 argument Alias with unlimited for the negative form.
25621 (fno-vect-cost-model): Removed.
25623 2017-01-05 Martin Liska <mliska@suse.cz>
25625 * hsa-gen.c (gen_hsa_divmod): New function.
25626 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
25628 2017-01-05 Martin Liska <mliska@suse.cz>
25631 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
25634 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
25636 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
25637 small constant length operands.
25639 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
25641 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
25642 between loop iterations.
25644 2017-01-05 Martin Liska <mliska@suse.cz>
25647 * gimplify.c (gimplify_decl_expr): Compare to
25648 asan_poisoned_variables instread of checking flags.
25649 (gimplify_target_expr): Likewise.
25650 (gimplify_expr): Likewise.
25651 (gimplify_function_tree): Conditionally initialize
25652 asan_poisoned_variables.
25654 2017-01-04 Jeff Law <law@redhat.com>
25656 PR tree-optimizatin/78812
25657 * rtl.h (contains_mem_rtx_p): Prototype.
25658 * ifcvt.c (containts_mem_rtx_p): Move from here to...
25659 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
25660 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
25661 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
25662 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
25664 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
25666 * input.c (assert_char_at_range): Default-initialize actual_range.
25668 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
25670 * df-scan.c (df_ref_create_structure): Make regno unsigned,
25671 to match the caller.
25673 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
25675 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
25676 insns after final jump in test to emit dummy move.
25678 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
25680 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
25681 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
25683 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
25685 * multiple_target.c (create_dispatcher_calls): Init e_next.
25686 * tree-ssa-loop-split.c (split_loop): Init border.
25687 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
25690 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
25695 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
25696 (altivec_register_operand): Do not return true if the operand
25697 contains a SUBREG mixing SImode and SFmode.
25698 (vsx_register_operand): Likewise.
25699 (vsx_reg_sfsubreg_ok): New predicate.
25700 (vfloat_operand): Do not return true if the operand contains a
25701 SUBREG mixing SImode and SFmode.
25702 (vint_operand): Likewise.
25703 (vlogical_operand): Likewise.
25704 (gpc_reg_operand): Likewise.
25705 (int_reg_operand): Likewise.
25706 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
25707 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
25708 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
25710 (rs6000_emit_move_si_sf_subreg): New helper function.
25711 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
25712 fixup SUBREGs involving SImode and SFmode.
25713 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
25714 numbers for the new peephole2 optimization.
25715 (peephole2 for SFmode unions): New peephole2 to optimize cases in
25716 the GLIBC math library that do AND/IOR/XOR operations on single
25717 precision floating point.
25718 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
25719 target macros to say whether we need to avoid SUBREGs mixing
25721 (TARGET_ALLOW_SF_SUBREG): Likewise.
25722 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
25723 (UNSPEC_SI_FROM_SF): Likewise.
25724 (iorxor): Change spacing.
25725 (and_ior_xor): New iterator for AND, IOR, and XOR.
25726 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
25727 (movdi_from_sf_zero_ext): Likewise.
25728 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
25729 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
25730 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
25731 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
25732 (fms<mode>4): Likewise.
25733 (fnma<mode>4): Likewise.
25734 (fnms<mode>4): Likewise.
25735 (nfma<mode>4): Likewise.
25736 (nfms<mode>4): Likewise.
25738 2017-01-04 Marek Polacek <polacek@redhat.com>
25741 * doc/invoke.texi: Document -Wpointer-compare.
25743 2017-01-04 Jakub Jelinek <jakub@redhat.com>
25745 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
25748 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
25749 descriptions for -gdwarf-5 and emit them as uleb128 instead of
25752 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
25755 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
25756 documentation of the powerpc_popcntb_ok attribute.
25757 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
25758 code to issue warning messages if a requested CPU configuration is
25759 not supported by the binary (assembler and loader) toolchain.
25760 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
25761 made to define a built-in function that has been disabled.
25762 (paired_init_builtins): Add assertion to prevent ICE if attempt is
25763 made to define a built-in function that has been disabled.
25764 (altivec_init_builtins): Add comment explaining why definition
25765 of the DST built-in functions is not preceded by an assertion
25766 check. Add assertions to prevent ICE if attempts are made to
25767 define an altivec predicate or an abs* built-in function that has
25769 (htm_init_builtins): Add comment explaining why definition of the
25770 htm built-in functions is not preceded by an assertion check.
25772 2017-01-04 Jeff Law <law@redhat.com>
25774 PR tree-optimizatin/67955
25775 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
25776 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
25777 the points-to solution does not include pt_null. Use DECL_PT_UID
25780 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
25782 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
25783 Use gen_int_mode instead of gen_lopwart for const_int operands.
25785 2017-01-04 Jakub Jelinek <jakub@redhat.com>
25787 PR tree-optimization/71563
25788 * match.pd: Simplify X << Y into X if Y is known to be 0 or
25789 out of range value - has low bits known to be zero.
25791 2017-01-04 Alan Modra <amodra@gmail.com>
25793 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
25794 * configure: Regenerate.
25795 * config.in: Regenerate.
25797 2017-01-04 Jakub Jelinek <jakub@redhat.com>
25800 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
25801 a substring of the message, but strcmp with the whole message. Ifdef
25802 ENABLE_NLS, translate the message first using dgettext.
25804 2017-01-03 Jeff Law <law@redhat.com>
25806 PR tree-optimizatin/78856
25807 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
25808 (mark_threaded_blocks): Remove code to truncate thread paths that
25809 cross multiple loop headers. Instead invalidate the cached loop
25810 iteration information and handle case of a thread path walking
25811 into an irreducible region.
25813 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
25816 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
25817 assertions. Add support for doing the signbit if the IEEE 128-bit
25818 floating point value is in a GPR.
25819 * config/rs6000/rs6000.md (Fsignbit): Delete.
25820 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
25821 Update the length attribute if the value is in a GPR.
25822 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
25823 the sign or zero extension instruction, since the value is always 0/1.
25824 (signbit<mode>2_dm2): Delete using <Fsignbit>.
25827 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
25828 extracting SImode to a GPR register so that we can generate a
25829 store, limit the vector to be in a traditional Altivec register
25830 for the vextuwrx instruction.
25832 2017-01-03 Ian Lance Taylor <iant@google.com>
25834 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
25836 2017-01-03 Martin Sebor <msebor@redhat.com>
25838 PR tree-optimization/78696
25839 * gimple-ssa-sprintf.c (format_floating): Correct handling of
25840 precision. Use MPFR for %f for greater fidelity. Correct handling
25842 (pass_sprintf_length::compute_format_length): Set width and precision
25843 specified by asrerisk to void_node for vararg functions.
25844 (try_substitute_return_value): Adjust dump output.
25846 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
25848 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
25850 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
25852 * doc/invoke.texi (SPARC options): Document -mlra as the default.
25853 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
25854 -mlra/-mno-lra was passed to the compiler.
25856 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
25858 PR rtl-optimization/65618
25859 * emit-rtl.c (try_split): Move initialization of "before" and
25860 "after" to just before the call to emit_insn_after_setloc.
25862 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
25864 * doc/md.texi (Standard Names): Remove reference to Java frontend.
25866 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
25868 * dwarf2out.c (gen_enumeration_type_die): When
25869 -gno-strict-dwarf, add a DW_AT_encoding attribute.
25871 2017-01-03 Jakub Jelinek <jakub@redhat.com>
25873 PR tree-optimization/78965
25874 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
25875 Change first argument from const call_info & to call_info &. For %n
25876 set info.nowrite to false.
25878 PR middle-end/78901
25879 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
25880 possibly throwing calls.
25882 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
25883 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
25884 and fns handling, rather than in a separate case SSA_NAME.
25886 2017-01-02 Jeff Law <law@redhat.com>
25888 * config/darwin-driver.c (darwin_driver_init): Const-correctness
25889 fixes for first_period and second_period variables.
25891 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
25894 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
25895 (*insvqi_1): New insn pattern.
25896 (*insvqi_1_mem_rex64): Ditto.
25897 (*insvqi_2): Ditto.
25898 (*insvqi_3): Rename from *insvqi.
25900 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
25902 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
25904 * doc/cfg.texi (Edges): Remove reference to Java.
25905 (Maintaining the CFG): Ditto.
25907 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
25909 PR middle-end/77674
25910 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
25911 transparent aliases.
25913 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
25915 PR middle-end/77484
25916 * predict.def (PRED_CALL): Update hitrate.
25917 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
25918 * predict.c (tree_estimate_probability_bb): Split CALL predictor
25919 into direct/indirect/polymorphic variants.
25921 2017-01-01 Jakub Jelinek <jakub@redhat.com>
25923 Update copyright years.
25925 * gcc.c (process_command): Update copyright notice dates.
25926 * gcov-dump.c (print_version): Ditto.
25927 * gcov.c (print_version): Ditto.
25928 * gcov-tool.c (print_version): Ditto.
25929 * gengtype.c (create_file): Ditto.
25930 * doc/cpp.texi: Bump @copying's copyright year.
25931 * doc/cppinternals.texi: Ditto.
25932 * doc/gcc.texi: Ditto.
25933 * doc/gccint.texi: Ditto.
25934 * doc/gcov.texi: Ditto.
25935 * doc/install.texi: Ditto.
25936 * doc/invoke.texi: Ditto.
25938 Copyright (C) 2017 Free Software Foundation, Inc.
25940 Copying and distribution of this file, with or without modification,
25941 are permitted in any medium without royalty provided the copyright
25942 notice and this notice are preserved.