1 2017-02-06 Julian Brown <julian@codesourcery.com>
2 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
3 Virendra Pathak <virendra.pathak@broadcom.com>
5 * config/aarch64/aarch64-cores.def: Change the scheduler
7 * config/aarch64/aarch64.md: Include thunderx2t99.md.
8 * config/aarch64/thunderx2t99.md: New file.
10 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
12 * doc/standards.texi (Go Language): Update link to language
15 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
17 * tree-eh.c (lower_resx): Sanitize profile.
18 (cleanup_empty_eh_move_lp): Likewise.
20 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
23 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
25 * cfgloopmanip.h (loop_version): Update prototype.
26 * modulo-sched.c (sms_schedule): Update call of loop_version.
27 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
28 * tree-parloops.c (gen_parallel_loop): Likewise.
29 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
30 * tree-ssa-loop-split.c (split_loop): Likewise.
31 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
32 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
34 2017-02-05 Martin Liska <mliska@suse.cz>
37 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
39 (print_operand_address): Initialize a struct to zero.
41 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
43 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
44 garbage collector only in textual form.
46 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
48 * doc/extend.texi (x86 specific memory model extensions for
49 transactional memory): Simplify a phrase.
51 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
54 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
55 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
56 (atomic_storedi_1): Likewise.
58 2017-02-04 Jakub Jelinek <jakub@redhat.com>
60 PR tree-optimization/79338
61 * tree-parloops.c (gather_scalar_reductions): Don't call
62 vect_analyze_loop_form for loop->inner before destroying loop's
65 2017-02-03 Martin Sebor <msebor@redhat.com>
67 PR tree-optimization/79327
68 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
69 when precision has resulted in leading zeros.
70 (format_integer): Adjust the likely counter to assume an unknown
71 argument that may be zero is non-zero.
73 2017-02-03 Jason Merrill <jason@redhat.com>
76 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
77 avoid copying non-taken branch.
79 2017-02-03 Jakub Jelinek <jakub@redhat.com>
81 PR tree-optimization/79340
82 * tree-vect-loop.c (vectorizable_reduction): Release
83 vec_defs elements after safe_splicing them into other vectors.
86 PR tree-optimization/79327
87 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
88 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
90 (format_integer): Use wide_int_to_tree instead of build_int_cst
91 + to_?hwi. If argmin is NULL, just set argmin and argmax to
92 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
93 of shortest and longest sequence.
95 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
97 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
98 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
100 2017-02-03 Walter Lee <walt@tilera.com>
103 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
104 after initial stackframe link reg save.
105 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
107 2017-02-03 Jakub Jelinek <jakub@redhat.com>
110 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
111 wu for stxssp alternative.
113 2017-02-03 Martin Sebor <msebor@redhat.com>
115 PR tree-optimization/79352
116 * gimple-fold.c (get_range_strlen): Add argument.
117 (get_range_strlen): Change return type to bool.
118 (get_maxval_strlen): Pass in a dummy argument.
119 * gimple-fold.h (get_range_strlen): Change return type to bool.
120 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
121 * tree.h (array_at_struct_end_p): Add argument.
122 * tree.c (array_at_struct_end_p): Handle it.
124 2017-02-03 Martin Liska <mliska@suse.cz>
127 * multiple_target.c (create_dispatcher_calls): Redirect edge
128 from a caller of a dispatcher.
129 (expand_target_clones): Make the clones local.
130 (ipa_target_clone): Do both target clones and resolvers.
131 (ipa_dispatcher_calls): Remove the pass.
132 (pass_dispatcher_calls::gate): Likewise.
133 (make_pass_dispatcher_calls): Likewise.
134 * passes.def (pass_target_clone): Put as very first IPA early
137 2017-02-03 Martin Liska <mliska@suse.cz>
139 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
140 in case of a function with ifunc attribute.
142 2017-02-03 Martin Liska <mliska@suse.cz>
144 * cgraph.c (cgraph_node::dump): Dump function version info.
145 * symtab.c (symtab_node::dump_base): Add missing new line.
147 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
149 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
150 (ifcombine_ifandif): Use it.
152 2017-02-03 Martin Liska <mliska@suse.cz>
154 * doc/invoke.texi: Document default value for
155 use-after-scope-direct-emission-threshold.
157 2017-02-03 Martin Liska <mliska@suse.cz>
159 PR tree-optimization/79339
160 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
161 (format_floating): Likewise.
163 2017-02-03 Martin Liska <mliska@suse.cz>
166 * ipa-prop.c (ipa_node_params_t::insert): Remove current
168 (ipa_node_params_t::remove): Likewise.
169 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
170 initialization from removed ipa_node_params_t::insert.
171 (ipa_node_params::~ipa_node_params): Move from removed
172 ipa_node_params_t::release.
173 * symbol-summary.h (symbol_summary::m_released): New member.
174 Do not release a summary twice. Do not allow to call finalizer
175 for types of a summary that live in GGC memory.
177 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
179 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
182 2017-02-02 Martin Sebor <msebor@redhat.com>
185 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
186 (format_string): Tighten up the range of output for non-constant
187 strings and correct the expected range for wide non-constant strings.
189 2017-02-02 Martin Sebor <msebor@redhat.com>
191 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
194 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
196 (-fdump-tree-@var): Add to index and document how to come up
197 with pass-specific option and dump file names.
198 (-fdump-passes): Clarify where to look for output.
200 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
203 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
204 statistics of the analyzed path; allow threading for speed when
205 any of BBs along the path are optimized for speed.
207 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
210 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
211 settings of the virtual registers.
214 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
216 * explow.c (get_dynamic_stack_size): Take known alignment of stack
217 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
220 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
222 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
223 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
225 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
227 * config/s390/s390.md: Add missing comments with the expanded
229 * config/s390/vector.md: Likewise.
230 * config/s390/vx-builtins.md: Likewise.
232 2017-02-02 Jakub Jelinek <jakub@redhat.com>
235 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
236 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
237 conditions on a single line.
239 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
241 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
242 __S390_VX__ to __VX__.
244 2017-02-01 Andrew Pinski <apinski@cavium.com>
246 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
247 stmt_info to record_stmt_cost.
248 (vect_get_known_peeling_cost): Pass stmt_info if known to
250 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
251 cpu_vector_cost field into
252 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
253 field into vec_int_stmt_cost and vec_fp_stmt_cost.
254 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
255 splitting of scalar_stmt_cost and vec_stmt_cost.
256 (thunderx_vector_cost): Likewise.
257 (cortexa57_vector_cost): LIkewise.
258 (exynosm1_vector_cost): Likewise.
259 (xgene1_vector_cost): Likewise.
260 (thunderx2t99_vector_cost): Improve after the splitting of the two
262 (aarch64_builtin_vectorization_cost): Update for the splitting of
263 scalar_stmt_cost and vec_stmt_cost.
265 2017-02-01 Torvald Riegel <triegel@redhat.com>
266 Richard Henderson <rth@redhat.com>
268 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
269 conditional on existance of a fast atomic load.
270 * optabs-query.c (can_atomic_load_p): New function.
271 * optabs-query.h (can_atomic_load_p): Declare it.
272 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
273 no fast atomic load is available for the particular size of access.
274 (expand_atomic_compare_and_swap): Likewise.
275 (expand_atomic_load): Likewise.
276 (expand_atomic_store): Likewise.
277 (expand_atomic_fetch_op): Likewise.
278 * testsuite/lib/target-supports.exp
279 (check_effective_target_sync_int_128): Remove x86 because it provides
281 (check_effective_target_sync_int_128_runtime): Likewise.
283 2017-02-01 Richard Biener <rguenther@suse.de>
285 * graphite.c: Include tree-vectorizer.h for find_loop_location.
286 (graphite_transform_loops): Provide opt-info for optimized nests.
287 * tree-parloop.c (parallelize_loops): Provide opt-info for
290 2017-02-01 Richard Biener <rguenther@suse.de>
293 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
296 2017-02-01 Richard Biener <rguenther@suse.de>
298 PR tree-optimization/71824
299 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
300 Verify the loops are valid in the merged SESE region.
301 (scop_detection::can_represent_loop_1): Check analyzing the
302 evolution of the number of iterations in the region succeeds.
304 2017-01-31 Ian Lance Taylor <iant@golang.org>
306 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
307 REG_ARGS_SIZE note to 32-bit push insns and call insn.
309 2017-01-31 David Malcolm <dmalcolm@redhat.com>
311 PR preprocessor/79210
312 * input.c (get_substring_ranges_for_loc): Replace line_width
313 assertion with error-handling.
315 2017-01-31 Richard Biener <rguenther@suse.de>
317 PR tree-optimization/77318
318 * graphite-sese-to-poly.c (extract_affine): Fix assert.
319 (create_pw_aff_from_tree): Take loop parameter.
320 (add_condition_to_pbb): Pass loop of the condition to
321 create_pw_aff_from_tree.
323 2017-01-31 Jakub Jelinek <jakub@redhat.com>
325 * config/s390/s390.c (s390_asan_shadow_offset): New function.
326 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
328 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
332 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
334 (convert_int_to_float128): Likewise.
335 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
336 (convert_int_to_float128): Likewise.
337 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
338 (UNSPEC_IEEE128_CONVERT): Likewise.
339 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
340 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
341 Use local variables for IBM extended format.
342 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
343 (fix_trunc<mode>si2_fprs): Likewise.
344 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
345 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
346 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
347 to know that we can now have integers of all sizes in vector
349 (fix<uns>_<mode>di2_hw): Likewise.
350 (float<uns>_<mode>si2_hw): Likewise.
351 (fix_<mode>si2_hw): Likewise.
352 (fixuns_<mode>si2_hw): Likewise.
353 (float<uns>_<mode>di2_hw): Likewise.
354 (float_<mode>di2_hw): Likewise.
355 (float_<mode>si2_hw): Likewise.
356 (floatuns_<mode>di2_hw): Likewise.
357 (floatuns_<mode>si2_hw): Likewise.
358 (xscvqp<su>wz_<mode>): Delete, no longer used.
359 (xscvqp<su>dz_<mode>): Likewise.
360 (xscv<su>dqp_<mode>): Likewise.
361 (ieee128_mfvsrd_64bit): Likewise.
362 (ieee128_mfvsrd_32bit): Likewise.
363 (ieee128_mfvsrwz): Likewise.
364 (ieee128_mtvsrw): Likewise.
365 (ieee128_mtvsrd_64bit): Likewise.
366 (ieee128_mtvsrd_32bit): Likewise.
368 2017-01-31 Martin Liska <mliska@suse.cz>
371 * ipa-prop.c (ipa_free_all_node_params): Call release method
372 instead of ~sumbol_summary to not to trigger double times
375 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
377 PR tree-optimization/71691
378 * bitmap.h (class auto_bitmap): New.
379 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
380 is_maybe_undefined instead of ssa_undefined_value_p.
382 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
384 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
385 __S390_ARCH_LEVEL__ to __ARCH__.
387 2017-01-31 Jakub Jelinek <jakub@redhat.com>
389 PR tree-optimization/79267
390 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
391 if should_remove_lhs_p is true.
393 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
396 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
397 (add_alignment_attribute): New.
398 (base_type_die): Add alignment attribute.
399 (subrange_type_die): Likewise.
400 (modified_type_die): Likewise.
401 (gen_array_type_die): Likewise.
402 (gen_descr_array_type_die: Likewise.
403 (gen_enumeration_type_die): Likewise.
404 (gen_subprogram_die): Likewise.
405 (gen_variable_die): Likewise.
406 (gen_field_die): Likewise.
407 (gen_ptr_to_mbr_type_die): Likewise.
408 (gen_struct_or_union_type_die): Likewise.
409 (gen_subroutine_type_die): Likewise.
410 (gen_typedef_die): Likewise.
411 (base_type_cmp): Compare alignment attribute.
413 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
416 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
417 (setb_unsigned) New pattern for setb with CCUNS.
418 * config/rs6000/rs6000.c (expand_block_compare): Use a different
419 subfc./subfe sequence to avoid overflow problems. Generate a
420 shorter sequence with cmpld/setb for power9.
421 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
422 for generating subfc. instruction.
423 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
424 now uses this instruction.
426 2017-01-30 Ian Lance Taylor <iant@google.com>
429 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
430 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
432 2017-01-30 Martin Sebor <msebor@redhat.com>
434 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
435 Move constant to the right of a relational operator.
436 (get_mpfr_format_length, format_character, format_string): Ditto.
437 (should_warn_p, maybe_warn): Same.
439 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
441 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
444 * asan.c (get_translation_unit_decl): Remove function.
445 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
447 2017-01-30 Martin Liska <mliska@suse.cz>
449 PR gcov-profile/79259
450 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
453 2017-01-30 Martin Liska <mliska@suse.cz>
456 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
457 Initialize variables with NULL value.
459 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
462 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
464 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
466 2017-01-30 Richard Biener <rguenther@suse.de>
468 PR tree-optimization/79276
469 * tree-vrp.c (process_assert_insertions): Properly adjust common
470 when removing a duplicate.
472 * gcc.dg/torture/pr79276.c: New testcase.
474 2017-01-30 Richard Biener <rguenther@suse.de>
476 PR tree-optimization/79256
477 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
478 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
480 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
482 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
485 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
486 ("*r<noxa>sbg_<mode>_sll_bitmask")
487 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
488 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
489 Use contiguous_bitmask_nowrap_operand.
491 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
494 * config/rs6000/altivec.h (vec_xl): Revise #define.
497 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
499 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
501 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
503 PR rtl-optimization/79194
504 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
505 traps before call to bypass_conditional_jumps.
507 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
509 PR tree-optimization/71374
510 * lra-constraints.c (check_conflict_input_operands): New.
511 (match_reload): Use it.
513 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
516 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
517 account to calculate conflict_set.
519 2017-01-27 Bin Cheng <bin.cheng@arm.com>
521 PR rtl-optimization/78559
522 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
523 other_insn in combine.
525 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
527 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
528 uint16_type_node for BT_UINT16.
530 2017-01-27 David Malcolm <dmalcolm@redhat.com>
532 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
534 (GIMPLE Tests): New node.
535 (RTL Tests): New node.
537 2017-01-27 Richard Biener <rguenther@suse.de>
539 PR tree-optimization/79245
540 * tree-loop-distribution.c (distribute_loop): Apply cost
541 modeling also to detected patterns.
543 2017-01-27 Richard Biener <rguenther@suse.de>
545 PR tree-optimization/71433
546 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
547 (compare_assert_loc): New function.
548 (process_assert_insertions): Sort and optimize assert locations
549 to remove duplicates and push down identical assertions on
550 edges to their destination block.
552 2017-01-27 Richard Biener <rguenther@suse.de>
554 PR tree-optimization/79244
555 * tree-vrp.c (remove_range_assertions): Forcefully propagate
556 out SSA names even if abnormal.
558 2017-01-27 Jakub Jelinek <jakub@redhat.com>
560 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
561 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
562 instead of MPFR_RNDN.
564 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
567 * arm.c (arm_option_override): Don't call build_target_option_node
568 until after doing all option overrides.
569 (arm_valid_target_attribute_tree): Likewise.
571 2017-01-27 Martin Liska <mliska@suse.cz>
573 * doc/invoke.texi (-fprofile-arcs): Document profiling support
574 for {cd}tors and C++ {cd}tors.
576 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
578 * config/s390/s390.md ("*setmem_long_and")
579 ("*setmem_long_and_31z"): Use zero_extend instead of and.
581 2017-01-26 Martin Sebor <msebor@redhat.com>
583 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
586 2017-01-26 Martin Sebor <msebor@redhat.com>
588 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
589 HAVE_DFmode before using XFmode or DFmode.
590 (parse_directive): Avoid using the z length modifier to avoid
591 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
594 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
595 to accept adjustment as an array.
596 (get_int_range): New function.
597 (struct directive): Make width and prec arrays.
598 (directive::set_width, directive::set_precision): Call get_int_range.
599 (format_integer, format_floating): Handle width and precision ranges.
600 (format_string, parse_directive): Same.
602 2017-01-26 Jakub Jelinek <jakub@redhat.com>
605 * dwarf2out.c (generate_skeleton_bottom_up): For children with
606 comdat_type_p set, just clone them, but keep the children in the
610 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
611 which have direct callers with -fvar-tracking-assignments enabled
613 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
614 inside of type units.
616 2017-01-26 Martin Sebor <msebor@redhat.com>
619 * gimple-ssa-sprintf.c (struct result_range): Add likely and
621 (struct format_result): Replace number_chars, number_chars_min,
622 and number_chars_max with a single member of struct result_range.
624 (format_result::operator+=): Adjust.
625 (struct fmtresult): Remove bounded. Handle likely and unlikely
627 (fmtresult::adjust_for_width_or_precision): New function.
628 (fmtresult:type_max_digits): New function.
629 (bytes_remaining): Handle likely and unlikely counters.
630 (min_bytes_remaining): Remove.
631 (format_percent): Simplify.
632 (format_integer, format_floating): Set likely and unlikely counters.
633 (get_string_length, format_character, format_string): Same.
634 (format_plain, should_warn_p): New function.
635 (maybe_warn): Call should_warn_p. Update diagnostic messages
636 and handle those for all directives, including plain strings.
637 (format_directive): Handle likely and unlikely counters.
638 Remove unnecessary quoting from diagnostics. Add an informational
641 (pass_sprintf_length::compute_format_length): Simplify.
642 (try_substitute_return_value): Handle likely and unlikely counters.
644 2017-01-26 Carl Love <cel@us.ibm.com>
646 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
647 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
649 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
652 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
653 endianess for subregs into account.
654 * lra-constraints.c (lra_constraints): Do risky transformations
655 always on the first iteration.
656 * lra-lives.c (check_pseudos_live_through_calls): Add arg
657 last_call_used_reg_set.
658 (process_bb_lives): Define and use last_call_used_reg_set.
659 * lra.c (lra): Always continue after lra_constraints on the first
662 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
664 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
666 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
668 2017-01-26 Jakub Jelinek <jakub@redhat.com>
670 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
671 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
672 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
673 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
674 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
675 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
676 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
677 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
678 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
680 2017-01-26 Marek Polacek <polacek@redhat.com>
683 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
684 for the third operand.
686 2017-01-26 Jakub Jelinek <jakub@redhat.com>
689 * omp-low.c (struct omp_context): Add simt_stmt field.
690 (scan_omp_for): Return omp_context *.
691 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
692 context to the _simt_ SIMD stmt.
693 (lower_omp_for): For combined SIMD with sibling _simt_
694 SIMD, make sure to use the same decls in _looptemp_
695 clauses as in the sibling.
697 2017-01-26 David Sherwood <david.sherwood@arm.com>
700 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
703 2017-01-26 Jakub Jelinek <jakub@redhat.com>
706 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
707 emit fld b; fld a; if possible.
709 * brig-builtins.def: Update copyright years.
710 * config/arm/arm_acle_builtins.def: Update copyright years.
712 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
715 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
716 constraint instead of o for the stxsd instruction.
718 2017-01-25 Carl Love <cel@us.ibm.com>
720 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
721 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
723 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
725 * doc/invoke.texi (C++ Dialect Options): Fix typo.
727 2017-01-25 Richard Biener <rguenther@suse.de>
729 PR tree-optimization/69264
730 * target.def (vector_alignment_reachable): Improve documentation.
731 * doc/tm.texi: Regenerate.
732 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
734 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
735 earlier changes with respect to TYPE_USER_ALIGN.
736 (vector_alignment_reachable_p): Likewise. Improve dumping.
738 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
741 * config/arm/arm.md (xordi3): Force constant operand into a register
744 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
746 * doc/invoke.texi (-fstore-merging): Correct default optimization
747 levels at which it is enabled.
748 (-O): Move -fstore-merging from list to...
751 2017-01-25 Richard Biener <rguenther@suse.de>
754 * omp-expand.c: Include debug.h.
755 (expand_omp_taskreg): Make sure to generate early debug before
756 outlining anything from a function.
757 (expand_omp_target): Likewise.
758 (grid_expand_target_grid_body): Likewise.
760 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
763 * asan.c (get_translation_unit_decl): New function.
764 (asan_add_global): Extract modules file name from globals
765 TRANSLATION_UNIT_DECL name.
767 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
770 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
771 for long calls with APCS frame and VFP.
773 2017-01-24 David Malcolm <dmalcolm@redhat.com>
775 * cfg.c (original_copy_tables_initialized_p): New function.
776 * cfg.h (original_copy_tables_initialized_p): New decl.
777 * cfgrtl.c (relink_block_chain): Guard the call to
778 free_original_copy_tables with a call to
779 original_copy_tables_initialized_p.
780 * cgraph.h (symtab_node::native_rtl_p): New decl.
781 * cgraphunit.c (symtab_node::native_rtl_p): New function.
782 (symtab_node::needed_p): Don't assert for early assembly output
784 (cgraph_node::finalize_function): Set "force_output" for __RTL
786 (cgraph_node::analyze): Bail out early for __RTL functions.
787 (analyze_functions): Update assertion to support __RTL functions.
788 (cgraph_node::expand): Bail out early for __RTL functions.
789 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
791 * function.h (struct function): Update comment for field
793 * gimple-expr.c: Include "tree-pass.h".
794 (gimple_has_body_p): Return false for __RTL functions.
795 * Makefile.in (OBJS): Add run-rtl-passes.o.
796 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
798 (gcc::pass_manager::get_clean_slate): New accessor.
799 * passes.c: Include "insn-addr.h".
800 (should_skip_pass_p): Add logging. Update logic for running
801 "expand" to be compatible with both __GIMPLE and __RTL. Guard
802 property-provider override so it is only done for gimple passes.
804 (skip_pass): New function.
805 (execute_one_pass): Call skip_pass when skipping passes.
806 * read-md.c (md_reader::read_char): Support filtering
807 the input to a subset of line numbers.
808 (md_reader::md_reader): Initialize fields
809 m_first_line and m_last_line.
810 (md_reader::read_file_fragment): New function.
811 * read-md.h (md_reader::read_file_fragment): New decl.
812 (md_reader::m_first_line): New field.
813 (md_reader::m_last_line): New field.
814 * read-rtl-function.c (function_reader::create_function): Only
815 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
816 curr_properties. Set DECL_INITIAL to a dummy block.
817 (read_rtl_function_body_from_file_range): New function.
818 * read-rtl-function.h (read_rtl_function_body_from_file_range):
820 * run-rtl-passes.c: New file.
821 * run-rtl-passes.h: New file.
823 2017-01-24 Jeff Law <law@redhat.com>
825 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
828 2017-01-24 Bin Cheng <bin.cheng@arm.com>
830 PR tree-optimization/79159
831 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
832 (record_nonwrapping_iv): Improve boundary using above function if no
833 value range information.
835 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
836 Martin Jambor <mjambor@suse.cz>
838 * brig-builtins.def: New file.
839 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
840 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
841 (DEF_HSAIL_SAT_BUILTIN): Likewise.
842 (DEF_HSAIL_INTR_BUILTIN): Likewise.
843 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
844 * builtin-types.def (BT_INT8): New.
845 (BT_INT16): Likewise.
846 (BT_UINT8): Likewise.
847 (BT_UINT16): Likewise.
848 (BT_FN_ULONG): Likewise.
849 (BT_FN_UINT_INT): Likewise.
850 (BT_FN_UINT_ULONG): Likewise.
851 (BT_FN_UINT_LONG): Likewise.
852 (BT_FN_UINT_PTR): Likewise.
853 (BT_FN_ULONG_PTR): Likewise.
854 (BT_FN_INT8_FLOAT): Likewise.
855 (BT_FN_INT16_FLOAT): Likewise.
856 (BT_FN_UINT32_FLOAT): Likewise.
857 (BT_FN_UINT16_FLOAT): Likewise.
858 (BT_FN_UINT8_FLOAT): Likewise.
859 (BT_FN_UINT64_FLOAT): Likewise.
860 (BT_FN_UINT16_UINT32): Likewise.
861 (BT_FN_UINT32_UINT16): Likewise.
862 (BT_FN_UINT16_UINT16_UINT16): Likewise.
863 (BT_FN_INT_PTR_INT): Likewise.
864 (BT_FN_UINT_PTR_UINT): Likewise.
865 (BT_FN_LONG_PTR_LONG): Likewise.
866 (BT_FN_ULONG_PTR_ULONG): Likewise.
867 (BT_FN_VOID_UINT64_UINT64): Likewise.
868 (BT_FN_UINT8_UINT8_UINT8): Likewise.
869 (BT_FN_INT8_INT8_INT8): Likewise.
870 (BT_FN_INT16_INT16_INT16): Likewise.
871 (BT_FN_INT_INT_INT): Likewise.
872 (BT_FN_UINT_FLOAT_UINT): Likewise.
873 (BT_FN_FLOAT_UINT_UINT): Likewise.
874 (BT_FN_ULONG_UINT_UINT): Likewise.
875 (BT_FN_ULONG_UINT_PTR): Likewise.
876 (BT_FN_ULONG_ULONG_ULONG): Likewise.
877 (BT_FN_UINT_UINT_UINT): Likewise.
878 (BT_FN_VOID_UINT_PTR): Likewise.
879 (BT_FN_UINT_UINT_PTR: Likewise.
880 (BT_FN_UINT32_UINT64_PTR): Likewise.
881 (BT_FN_INT_INT_UINT_UINT): Likewise.
882 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
883 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
884 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
885 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
886 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
887 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
888 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
889 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
890 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
891 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
892 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
893 * doc/frontends.texi: List BRIG FE.
894 * doc/install.texi (Testing): Add BRIG tesring requirements.
895 * doc/invoke.texi (Overall Options): Mention BRIG.
896 * doc/standards.texi (Standards): Doucment BRIG HSA version.
898 2017-01-24 Richard Biener <rguenther@suse.de>
901 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
903 2017-01-24 Martin Jambor <mjambor@suse.cz>
906 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
907 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
910 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
913 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
914 casts from signed to unsigned really don't have a range.
916 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
918 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
919 GMP_RNDx for compatiblity.
921 2017-01-24 Martin Liska <mliska@suse.cz>
924 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
925 that would prevent us to call alloca with -1 as argument.
927 2017-01-24 Jakub Jelinek <jakub@redhat.com>
929 * dwarf2out.c (output_compilation_unit_header, output_file_names):
930 Avoid -Wformat-security warning.
932 2017-01-23 Andrew Pinski <apinski@cavium.com>
934 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
937 2017-01-23 Martin Sebor <msebor@redhat.com>
940 * gimple-ssa-sprintf.c (warn_level): New global.
941 (format_integer): Use it here and throughout the rest of the file.
942 Use the same switch to compute sign as base.
943 (maybe_warn): New function.
944 (format_directive): Factor out warnings into maybe_warn.
945 Add debugging output. Use warn_level.
946 (add_bytes): Use warn_level.
947 (pass_sprintf_length::compute_format_length): Add debugging output.
948 (try_substitute_return_value): Same.
949 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
952 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
953 (struct fmtresult, format_integer, format_floating): Adjust.
954 (fmtresult::fmtresult): Set max correctly in two argument ctor.
955 (get_string_length, format_string,format_directive): Same.
956 (pass_sprintf_length::compute_format_length): Same.
957 (try_substitute_return_value): Simplify slightly.
960 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
961 (fmtresult::operator+=): Outlined.
962 (struct fmtresult): Add ctors.
963 (struct conversion_spec): Rename...
964 (struct directive): ...to this. Add and remove data members.
965 (directive::set_width, directive::set_precision): New functions.
966 (format_percent): Use fmtresult ctor.
967 (get_width_and_precision): Remove.
968 (format_integer): Make naming changes. Avoid computing width and
970 (format_floating): Same. Adjust indentation.
971 (format_character, format_none): New functions.
972 (format_string): Moved character handling to format_character.
973 (format_directive): Remove arguments, change return type.
974 (parse_directive): New function.
975 (pass_sprintf_length::compute_format_length): Move directive
976 parsing to parse_directive.
978 2017-01-23 Jakub Jelinek <jakub@redhat.com>
980 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
981 (assign_assembler_name_if_needed): ... this.
982 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
983 (assign_assembler_name_if_needed): ... this.
984 (free_lang_data_in_cgraph): Adjust callers.
985 * cgraphunit.c (cgraph_node::analyze): Likewise.
986 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
989 2017-01-23 Richard Biener <rguenther@suse.de>
991 PR tree-optimization/79088
992 PR tree-optimization/79188
993 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
994 resetting loop bounds after last path deletion. Reset loop
995 bounds of the target loop, make code match the comments.
996 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
997 Make sure loops need no fixups.
999 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1001 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1002 exponent support with double type for first argument.
1003 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1004 type returned by __builtin_vec_extract_sig,
1005 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1006 functions from "vector int" to "vector unsigned int" or from
1007 "vector long long int" to "vector unsigned long long int".
1008 Changed type returned by __builtin_vec_extract_exp,
1009 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1010 functions from "vector int" to "vector unsigned int" or from
1011 "vector long long int" to "vector unsigned long long int".
1012 Changed return type of __builtin_vec_test_data_class,
1013 __builtin_vec_test_data_class_sp, and
1014 __builtin_vec_test_data_class_dp from "vector int" to
1015 "vector bool int" or from "vector long long int" to "vector bool
1016 long long int" and changed second argument type from "unsigned
1017 int" to "int". Added new overloaded function forms "vector float
1018 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1019 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1020 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1021 double, vector unsigned long long int)" and "vector double
1022 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1023 long int)". Changed return type of
1024 __builtin_scalar_test_data_class and
1025 __builtin_scalar_test_data_class_sp and
1026 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1027 int" and changed second argument from "unsigned int" to "int".
1028 Changed type returned by __builtin_scalar_test_neg,
1029 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1030 from "int" to "bool int". Added new overloaded function form
1031 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1032 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1033 exponent double-precision with floating point first argument.
1034 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1035 documentation of scalar_test_data_class, scalar_test_neg,
1036 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1037 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1038 vec_test_data_class built-in functions to reflect refinements in
1039 their type signatures.
1041 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1043 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1045 (aarch64_elf_asm_destructor): Likewise.
1047 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1049 PR rtl-optimization/78634
1050 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1051 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1052 * ifcvt.c (noce_try_cmove): Add missing cost check.
1054 PR rtl-optimization/71724
1055 * combine.c (if_then_else_cond): Look for situations where it is
1056 beneficial to undo the work of one of the recursive calls.
1058 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1060 PR tree-optimization/70754
1061 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1062 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1063 combined stmt before it if not NULL.
1064 (combine_chains): Process refs reversely and compute dominance point
1067 2017-01-23 Martin Liska <mliska@suse.cz>
1069 PR tree-optimization/79196
1070 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1071 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1073 (strlen_optimize_stmt): Call the renamed function.
1075 2017-01-23 Michael Matz <matz@suse.de>
1077 PR tree-optimization/78384
1078 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
1080 2017-01-23 Richard Biener <rguenther@suse.de>
1082 PR tree-optimization/79186
1083 * tree-vrp.c (register_new_assert_for): Make sure we've seen
1084 both incoming edges before moving an assert.
1086 2017-01-23 Martin Jambor <mjambor@suse.cz>
1088 * ipa-prop.c (load_from_param_1): Removed.
1089 (load_from_unmodified_param): Bits from load_from_param_1 put back
1091 (load_from_param): Removed.
1092 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
1093 with stmt. Reverted back to use of load_from_unmodified_param.
1095 2017-01-23 Martin Jambor <mjambor@suse.cz>
1098 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
1099 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
1100 field a pointer to garbage collected vector, mark lattices and
1101 ipcp_orig_node with GTY((skip)).
1102 (ipa_get_param_count): Adjust to descriptors being a pointer.
1103 (ipa_get_param): Likewise.
1104 (ipa_get_type): Likewise.
1105 (ipa_get_param_move_cost): Likewise.
1106 (ipa_set_param_used): Likewise.
1107 (ipa_get_controlled_uses): Likewise.
1108 (ipa_set_controlled_uses): Likewise.
1109 (ipa_is_param_used): Likewise.
1110 (ipa_node_params_t): Move into garbage collector. New methods insert
1112 (ipa_node_params_sum): Annotate wth GTY(()).
1113 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
1115 (ipa_load_from_parm_agg): Adjust declaration.
1116 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
1117 * ipa-profile.c (ipa_profile): Likewise.
1118 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
1119 (ipa_populate_param_decls): Make descriptors parameter garbage
1121 (ipa_dump_param): Adjust to descriptors being a pointer.
1122 (ipa_alloc_node_params): Likewise.
1123 (ipa_initialize_node_params): Likewise.
1124 (load_from_param_1): Make descriptors parameter garbage collected.
1125 (load_from_unmodified_param): Likewise.
1126 (load_from_param): Likewise.
1127 (ipa_load_from_parm_agg): Likewise.
1128 (ipa_node_params::~ipa_node_params): Removed.
1129 (ipa_free_all_node_params): Remove call to delete operator.
1130 (ipa_node_params_t::insert): New.
1131 (ipa_node_params_t::remove): Likewise.
1132 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
1133 copy known_csts and known_contexts vectors.
1134 (ipa_read_node_info): Adjust to descriptors being a pointer.
1135 (ipcp_modif_dom_walker): Make m_descriptors field garbage
1137 (ipcp_transform_function): Make descriptors variable garbage
1140 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1142 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1143 * config/i386/avx512dqintrin.h: Ditto.
1144 * config/i386/avx512fintrin.h: Ditto.
1145 * gcc/config/i386/i386.c: Handle new builtins.
1146 * config/i386/i386-builtin.def: Add new builtins.
1147 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1148 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1150 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1151 Martin Liska <mliska@suse.cz>
1153 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1154 * asan.c (asan_expand_poison_ifn): Support stores and use
1155 appropriate ASAN report function.
1156 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1157 * internal-fn.def (ASAN_POISON_USE): Declare.
1158 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1159 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1160 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1161 ASAN_POISON calls w/o LHS.
1162 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1163 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1164 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1165 * gimplify.c (asan_poison_variables): Add attribute
1166 use_after_scope_memory to variables that really needs to live
1168 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1169 having the attribute.
1171 2017-01-23 Martin Liska <mliska@suse.cz>
1173 * asan.c (create_asan_shadow_var): New function.
1174 (asan_expand_poison_ifn): Likewise.
1175 * asan.h (asan_expand_poison_ifn): New declaration.
1176 * internal-fn.c (expand_ASAN_POISON): Likewise.
1177 * internal-fn.def (ASAN_POISON): New builtin.
1178 * sanopt.c (pass_sanopt::execute): Expand
1179 asan_expand_poison_ifn.
1180 * tree-inline.c (copy_decl_for_dup_finish): Make function
1182 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1183 * tree-ssa.c (is_asan_mark_p): New function.
1184 (execute_update_addresses_taken): Rewrite local variables
1185 (identified just by use-after-scope as addressable) into SSA.
1187 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1189 * doc/install.texi (Specific): opensource.apple.com uses https
1190 now. Remove trailing slash.
1192 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1194 * README.Portability: Remove note on an Irix compatibility issue.
1196 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1198 * gcov.c (INCLUDE_ALGORITHM): Define.
1199 (INCLUDE_VECTOR): Define.
1200 No longer include <vector> and <algorithm> directly.
1202 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1204 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1206 * doc/invoke.texi (Code Gen Options): Ditto.
1208 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1211 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1213 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1215 rtl-optimization/79125
1216 * cprop.c (local_cprop_pass): Handle cases where we make an
1219 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1223 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1224 read from, for big endian.
1226 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1228 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1229 register pauth builtins for LP64 only.
1231 2017-01-20 Marek Polacek <polacek@redhat.com>
1234 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1237 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1239 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1241 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1242 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1243 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1245 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1248 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1249 in big-endian mode when they are not a single duplicated value.
1251 2017-01-20 Richard Biener <rguenther@suse.de>
1253 * BASE-VER: Bump to 7.0.1.
1255 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1257 * omp-low.c (omplow_simd_context): New struct. Use it...
1258 (lower_rec_simd_input_clauses): ...here and...
1259 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1260 references to idx, lane, max_vf, is_simt.
1262 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1264 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1267 2017-01-20 Martin Jambor <mjambor@suse.cz>
1269 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1270 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1272 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1273 (GTFILES): Rename hsa.c to hsa-common.c.
1274 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1275 * hsa-dump.c: Likewise.
1276 * hsa-gen.c: Likewise.
1277 * hsa-regalloc.c: Likewise.
1278 * ipa-hsa.c: Likewise.
1279 * omp-expand.c: Likewise.
1280 * omp-low.c: Likewise.
1281 * toplev.c: Likewise.
1283 2017-01-20 Marek Polacek <polacek@redhat.com>
1286 * doc/invoke.texi: Document -Wduplicated-branches.
1287 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1288 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1289 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1290 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1291 return 0 only when not OEP_LEXICOGRAPHIC.
1292 (fold_build_cleanup_point_expr): Use the expression
1293 location when building CLEANUP_POINT_EXPR.
1294 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1295 * tree.c (add_expr): Handle error_mark_node.
1297 2017-01-20 Martin Liska <mliska@suse.cz>
1300 * tree-profile.c (init_ic_make_global_vars): Do not call
1302 (gimple_init_gcov_profiler): Likewise.
1304 2017-01-20 Martin Liska <mliska@suse.cz>
1307 * cgraph.h (maybe_create_reference): Remove argument and
1309 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1311 * ipa-cp.c (create_specialized_node): Likewise.
1312 * symtab.c (symtab_node::maybe_create_reference): Handle
1313 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1315 2017-01-20 Martin Liska <mliska@suse.cz>
1317 * read-rtl-function.c (function_reader::create_function): Use
1318 build_decl instread of build_decl_stat.
1320 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1322 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1323 * config/i386/avx512dqintrin.h: Ditto.
1324 * config/i386/avx512fintrin.h: Ditto.
1325 * config/i386/i386-builtin-types.def: Add new types.
1326 * gcc/config/i386/i386.c: Handle new types.
1327 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1328 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1329 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1330 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1331 (__builtin_ia32_kshiftridi): New.
1332 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1334 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1338 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1339 define to rs6000_init_stack_protect_guard.
1340 (rs6000_init_stack_protect_guard): New function.
1342 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1343 Yunqiang Su <yunqiang.su@imgtec.com>
1345 * config.gcc (supported_defaults): Add madd4.
1346 (with_madd4): Add validation.
1347 (all_defaults): Add madd4.
1348 * config/mips/mips.opt (mmadd4): New option.
1349 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1351 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1353 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1354 (ISA_HAS_FUSED_MADD4): Likewise.
1355 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1356 * gcc/doc/install.texi (--with-madd4): Document the new option.
1358 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1360 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1361 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1362 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1363 (aarch64_init_pauth_hint_builtins): New.
1364 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1365 (aarch64_expand_builtin): Expand new builtins.
1367 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1369 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1370 * combine-stack-adj.c (no_unhandled_cfa): Handle
1371 REG_CFA_TOGGLE_RA_MANGLE.
1372 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1373 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1374 info for return address signing.
1375 (aarch64_expand_epilogue): Likewise.
1377 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1379 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1380 * config/aarch64/aarch64-protos.h
1381 (aarch64_return_address_signing_enabled): New declaration.
1382 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1384 (aarch64_expand_prologue): Sign return address before it's pushed onto
1386 (aarch64_expand_epilogue): Authenticate return address fetched from
1388 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1389 (aarch64_attributes): New function attributes for "sign-return-address".
1390 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1391 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1392 ("*do_return"): Generate combined instructions according to key index.
1393 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1394 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1396 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1397 * config/aarch64/aarch64.opt (msign-return-address=): New.
1398 * doc/extend.texi (AArch64 Function Attributes): Documents
1399 "sign-return-address=".
1400 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1402 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1404 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1405 overall option summary.
1407 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1409 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1410 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1411 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1412 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1414 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1416 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1417 -mpower9-minmax by default for -mcpu=power9.
1418 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1419 128-bit floating point.
1421 2017-01-20 Alan Modra <amodra@gmail.com>
1423 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1424 optimizing for size.
1426 2017-01-20 Alan Modra <amodra@gmail.com>
1429 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1430 for strcmp and strncmp from corresponding builtin decl.
1432 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1434 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1435 instead of i386/rtems-64.h.
1436 * config/i386/rtems-64.h: Remove.
1438 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1442 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1444 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1446 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1448 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1449 Change int to HOST_WIDE_INT.
1450 * config/aarch64/aarch64-protos.h
1451 (aarch64_simd_gen_const_vector_dup): Likewise.
1452 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1454 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1456 * langhooks-def.h (lhd_type_for_size): New decl.
1457 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1458 * langhooks.c (lhd_type_for_size): New function, taken from
1461 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1463 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1464 define_bypass for CR latency.
1465 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1466 (power9-alu2): Add define_bypass for CR latency.
1468 (power9-mul): Update insn latency.
1469 (power9-mul-compare): Update insn latency, bypass latency and remove
1472 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1474 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1476 * config/aarch64/aarch64.md
1477 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1478 aarch64_nopcrelative_literal_loads.
1479 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1481 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1483 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1485 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1487 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1490 * config.gcc (supported_defaults): Add lxc1-sxc1.
1491 (with_lxc1_sxc1): Add validation.
1492 (all_defaults): Add lxc1-sxc1.
1493 * config/mips/mips.opt (mlxc1-sxc1): New option.
1494 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1496 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1497 __mips_no_lxc1_sxc1.
1498 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1499 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1500 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1502 2017-01-19 Richard Biener <rguenther@suse.de>
1504 PR tree-optimization/72488
1505 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1506 sure to restore SSA info.
1507 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1509 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1511 PR rtl-optimization/79121
1512 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1513 of the inner type when shifting an extended value.
1515 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1518 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1519 interposable aliases.
1521 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1524 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1525 Use the evmergelohi instruction.
1526 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1527 (mov_sitf_e500_subreg8_2_be): Likewise.
1528 (mov_sitf_e500_subreg12_2_le): Likewise.
1529 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1530 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1531 (mov_sitf_e500_subreg8_2_le): Likewise.
1532 (mov_sitf_e500_subreg12_2_be): Likewise.
1534 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1536 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1537 attribute from vecsimple to vecperm.
1538 (altivec_vbpermq2): Likewise.
1540 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1543 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1545 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1546 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1547 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1548 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1549 case where N arg is SIZE_MAX.
1550 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1551 (cmpstrsi): Add pattern.
1553 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1555 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1556 __builtin_vec_revb builtins.
1557 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1558 built-in functions to support generation of the ISA 3.0 XXBR<x>
1559 vector byte reverse instructions.
1560 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1561 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1562 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1563 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1564 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1565 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1566 (P9V_BUILTIN_VEC_REVB): Likewise.
1567 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1568 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1569 (p9_xxbrq_v16qi): Likewise.
1570 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1571 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1572 (p9_xxbrh_v8hi): Likewise.
1573 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1574 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1575 vec_revb built-in functions.
1577 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1579 PR rtl-optimization/78952
1580 * config/i386/i386.md (any_extract): New code iterator.
1581 (*insvqi_2): Use any_extract for source operand.
1582 (*insvqi_3): Use any_shiftrt for source operand.
1584 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1586 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1588 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1590 2017-01-18 Matthias Klose <doko@ubuntu.com>
1592 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1594 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1596 * config/rs6000/altivec.h (vec_bperm): Change #define.
1597 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1598 (altivec_vbpermq2): New define_insn.
1599 (altivec_vbpermd): Likewise.
1600 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1602 (VBPERMD): Likewise.
1603 (VBPERM): New polymorphic function interface.
1604 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1605 Add entries for P9V_BUILTIN_VEC_VBPERM.
1606 * doc/extend.texi: Add interfaces for vec_bperm.
1608 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1610 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1611 first letter of error messages.
1612 (s390_resolve_overloaded_builtin): Likewise.
1613 * config/s390/s390.c (s390_expand_builtin): Likewise.
1614 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1615 (s390_valid_target_attribute_inner_p): Likewise.
1616 * config/s390/s390.md ("tabort"): Likewise.
1618 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1620 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1621 (ISA_AVOID_DIV_HILO): New macro.
1622 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1623 (ISA_HAS_DDIV): Likewise.
1625 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1627 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1629 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1631 * doc/invoke.texi (fabi-version): Spelling fix.
1633 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1636 * doc/invoke.texi (fabi-version): Mention mangling fix for
1639 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1642 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1644 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1647 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1648 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1650 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1652 (stack_protect_test): Ditto.
1653 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1654 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1656 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1657 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1658 -mstack-protector-guard-offset=.
1659 (RS/6000 and PowerPC Options): Ditto.
1661 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1663 * config/i386/i386.h (MASK_CLASS_P): New define.
1664 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1665 there are no registers from different register sets also when
1666 mask registers are used. Update function comment.
1667 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1668 to (*k/*r) and (*k/*km) alternatives.
1670 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1672 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1673 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1674 (EH_RETURN_HANDLER_RTX): New define.
1675 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1676 Force frame pointer in EH return functions.
1677 (aarch64_expand_epilogue): Add barrier for eh_return.
1678 (aarch64_final_eh_return_addr): Remove.
1679 (aarch64_eh_return_handler_rtx): New function.
1680 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1682 (aarch64_eh_return_handler_rtx): New prototype.
1684 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1686 * config/rs6000/altivec.h (vec_rlmi): New #define.
1687 (vec_vrlnm): Likewise.
1688 (vec_rlnm): Likewise.
1689 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1690 (UNSPEC_VRLNM): Likewise.
1691 (VIlong): New mode iterator.
1692 (altivec_vrl<VI_char>mi): New define_insn.
1693 (altivec_vrl<VI_char>nm): Likewise.
1694 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1697 (RLNM): New polymorphic function entry.
1698 (VRLWMI): New monomorphic function entry.
1700 (RLMI): New polymorphic function entry.
1701 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1702 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1703 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1706 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1709 * dwarf2out.c (field_byte_offset): Restore the
1710 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1711 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1712 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1715 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1718 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1720 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1723 * dwarf2out.c (add_data_member_location_attribute): For constant
1724 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1725 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1726 DW_AT_byte_size attributes.
1728 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1730 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1731 after forcing to constant memory when the code model is medium.
1733 2017-01-17 Julia Koval <julia.koval@intel.com>
1736 * config/i386/avx512fintrin.h
1737 (_mm512_i32gather_ps): Change __addr type to void const*.
1738 (_mm512_mask_i32gather_ps): Ditto.
1739 (_mm512_i32gather_pd): Ditto.
1740 (_mm512_mask_i32gather_pd): Ditto.
1741 (_mm512_i64gather_ps): Ditto.
1742 (_mm512_mask_i64gather_ps): Ditto.
1743 (_mm512_i64gather_pd): Ditto.
1744 (_mm512_mask_i64gather_pd): Ditto.
1745 (_mm512_i32gather_epi32): Ditto.
1746 (_mm512_mask_i32gather_epi32): Ditto.
1747 (_mm512_i32gather_epi64): Ditto.
1748 (_mm512_mask_i32gather_epi64): Ditto.
1749 (_mm512_i64gather_epi32): Ditto.
1750 (_mm512_mask_i64gather_epi32): Ditto.
1751 (_mm512_i64gather_epi64): Ditto.
1752 (_mm512_mask_i64gather_epi64): Ditto.
1753 (_mm512_i32scatter_ps): Change __addr type to void*.
1754 (_mm512_mask_i32scatter_ps): Ditto.
1755 (_mm512_i32scatter_pd): Ditto.
1756 (_mm512_mask_i32scatter_pd): Ditto.
1757 (_mm512_i64scatter_ps): Ditto.
1758 (_mm512_mask_i64scatter_ps): Ditto.
1759 (_mm512_i64scatter_pd): Ditto.
1760 (_mm512_mask_i64scatter_pd): Ditto.
1761 (_mm512_i32scatter_epi32): Ditto.
1762 (_mm512_mask_i32scatter_epi32): Ditto.
1763 (_mm512_i32scatter_epi64): Ditto.
1764 (_mm512_mask_i32scatter_epi64): Ditto.
1765 (_mm512_i64scatter_epi32): Ditto.
1766 (_mm512_mask_i64scatter_epi32): Ditto.
1767 (_mm512_i64scatter_epi64): Ditto.
1768 (_mm512_mask_i64scatter_epi64): Ditto.
1769 * config/i386/avx512pfintrin.h
1770 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1771 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1772 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1773 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1774 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1775 (_mm512_prefetch_i32scatter_ps): Ditto.
1776 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1777 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1778 (_mm512_prefetch_i64scatter_pd): Ditto.
1779 (_mm512_prefetch_i64scatter_ps): Ditto.
1780 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1781 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1782 * config/i386/avx512vlintrin.h
1783 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1784 (_mm_mmask_i32gather_ps): Ditto.
1785 (_mm256_mmask_i32gather_pd): Ditto.
1786 (_mm_mmask_i32gather_pd): Ditto.
1787 (_mm256_mmask_i64gather_ps): Ditto.
1788 (_mm_mmask_i64gather_ps): Ditto.
1789 (_mm256_mmask_i64gather_pd): Ditto.
1790 (_mm_mmask_i64gather_pd): Ditto.
1791 (_mm256_mmask_i32gather_epi32): Ditto.
1792 (_mm_mmask_i32gather_epi32): Ditto.
1793 (_mm256_mmask_i32gather_epi64): Ditto.
1794 (_mm_mmask_i32gather_epi64): Ditto.
1795 (_mm256_mmask_i64gather_epi32): Ditto.
1796 (_mm_mmask_i64gather_epi32): Ditto.
1797 (_mm256_mmask_i64gather_epi64): Ditto.
1798 (_mm_mmask_i64gather_epi64): Ditto.
1799 (_mm256_i32scatter_ps): Change __addr type to void*.
1800 (_mm256_mask_i32scatter_ps): Ditto.
1801 (_mm_i32scatter_ps): Ditto.
1802 (_mm_mask_i32scatter_ps): Ditto.
1803 (_mm256_i32scatter_pd): Ditto.
1804 (_mm256_mask_i32scatter_pd): Ditto.
1805 (_mm_i32scatter_pd): Ditto.
1806 (_mm_mask_i32scatter_pd): Ditto.
1807 (_mm256_i64scatter_ps): Ditto.
1808 (_mm256_mask_i64scatter_ps): Ditto.
1809 (_mm_i64scatter_ps): Ditto.
1810 (_mm_mask_i64scatter_ps): Ditto.
1811 (_mm256_i64scatter_pd): Ditto.
1812 (_mm256_mask_i64scatter_pd): Ditto.
1813 (_mm_i64scatter_pd): Ditto.
1814 (_mm_mask_i64scatter_pd): Ditto.
1815 (_mm256_i32scatter_epi32): Ditto.
1816 (_mm256_mask_i32scatter_epi32): Ditto.
1817 (_mm_i32scatter_epi32): Ditto.
1818 (_mm_mask_i32scatter_epi32): Ditto.
1819 (_mm256_i32scatter_epi64): Ditto.
1820 (_mm256_mask_i32scatter_epi64): Ditto.
1821 (_mm_i32scatter_epi64): Ditto.
1822 (_mm_mask_i32scatter_epi64): Ditto.
1823 (_mm256_i64scatter_epi32): Ditto.
1824 (_mm256_mask_i64scatter_epi32): Ditto.
1825 (_mm_i64scatter_epi32): Ditto.
1826 (_mm_mask_i64scatter_epi32): Ditto.
1827 (_mm256_i64scatter_epi64): Ditto.
1828 (_mm256_mask_i64scatter_epi64): Ditto.
1829 (_mm_i64scatter_epi64): Ditto.
1830 (_mm_mask_i64scatter_epi64): Ditto.
1831 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1832 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1833 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1834 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1835 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1836 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1837 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1838 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1839 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1840 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1841 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1842 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1843 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1844 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1845 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1846 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1847 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1848 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1849 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1850 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1851 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1852 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1853 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1854 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1855 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1856 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1857 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1858 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1859 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1860 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1861 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1862 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1863 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1864 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1865 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1866 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1867 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1868 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1869 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1870 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1871 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1872 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1873 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1874 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1875 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1876 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1877 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1878 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1879 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1880 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1881 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1882 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1883 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1884 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1885 definitions accordingly.
1887 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1888 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1891 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1894 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1897 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1898 hard regs for inner regno.
1900 2017-01-17 Martin Liska <mliska@suse.cz>
1903 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1904 assumption and add comment.
1906 2017-01-17 Nathan Sidwell <nathan@acm.org>
1908 * ipa-visibility.c (localize_node): New function, broken out of ...
1909 (function_and_variable_visibility): ... here. Call it.
1911 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1914 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1915 correctly set frequency of oudgoing edge.
1916 (duplicate_thread_path): Fix profile updating.
1918 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1921 * configure.ac: Add GCC_BASE_VER.
1922 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1923 version from BASE-VER file.
1924 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1925 (gcc.o): Depend on $(BASEVER).
1926 * common.opt (dumpfullversion): New option.
1927 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1928 * doc/invoke.texi: Document -dumpfullversion.
1929 * doc/install.texi: Document --with-gcc-major-version-only.
1930 * configure: Regenerated.
1932 2017-01-17 Richard Biener <rguenther@suse.de>
1934 PR tree-optimization/71433
1935 * tree-vrp.c (register_new_assert_for): Merge same asserts
1936 on all incoming edges.
1937 (process_assert_insertions_for): Handle insertions at the
1940 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1942 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1943 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1945 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1948 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1951 2017-01-17 Alan Modra <amodra@gmail.com>
1954 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1955 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1956 symbolic stack limit when pic.
1958 2017-01-16 Martin Sebor <msebor@redhat.com>
1960 PR tree-optimization/78608
1961 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1963 2017-01-16 Jeff Law <law@redhat.com>
1966 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1967 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1968 for several include directories that may be relative to sysroot.
1969 * config/i386/x-mingw32 (gplus_includedir): Define.
1970 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1971 (native_system_includedir): Likewise.
1972 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1973 override if TARGET_SYSTEM_ROOT is defined.
1974 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1976 PR tree-optimization/79090
1977 PR tree-optimization/33562
1978 PR tree-optimization/61912
1979 PR tree-optimization/77485
1980 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1981 and computed trims into the dump file.
1983 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1985 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1987 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1990 * gimplify.c (gimplify_init_constructor): If want_value and
1991 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1995 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1996 sequence. Formatting fixes.
1997 (doloop_optimize): Formatting fixes.
2000 * gcc.c (debug_level_greater_than_spec_func): New function.
2001 (static_spec_functions): Add debug-level-gt spec function.
2002 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2004 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2005 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2006 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2007 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2008 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2009 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2011 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2013 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2014 QImode fixups to general and mask registers only.
2016 2017-01-16 Carl Love <cel@us.ibm.com>
2018 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2019 for built-in functions
2020 vector signed char vec_nabs (vector signed char)
2021 vector signed short vec_nabs (vector signed short)
2022 vector signed int vec_nabs (vector signed int)
2023 vector signed long long vec_nabs (vector signed long long)
2024 vector float vec_nabs (vector float)
2025 vector double vec_nabs (vector double)
2026 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2028 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2029 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2030 * doc/extend.texi: Update the documentation file for the new built-in
2033 2017-01-16 Martin Sebor <msebor@redhat.com>
2035 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2038 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2040 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2041 UNSPEC_VSX__XXSPLTD to require special splat handling.
2043 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2046 * system.h: Poison strndup.
2048 2017-01-16 Alan Modra <amodra@gmail.com>
2051 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2054 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2056 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2058 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2060 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2061 call recog here. Assert that INSN_CODE (insn) is non-negative.
2063 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2066 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2068 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2069 in the currently scheduled RTL fragment.
2071 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2073 PR rtl-optimization/78751
2074 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2077 2017-01-14 Jeff Law <law@redhat.com>
2079 PR tree-optimization/79090
2080 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
2081 variable length stores.
2082 (compute_trims): Delete dead assignment to *trim_tail.
2083 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
2086 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
2088 PR rtl-optimization/78626
2089 PR rtl-optimization/78727
2090 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
2091 of a block, and split such blocks after everything else is finished.
2093 2017-01-14 Alan Modra <amodra@gmail.com>
2096 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
2097 target legitimate_combined_insn.
2098 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
2099 (rs6000_legitimate_combined_insn): New function.
2100 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
2102 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
2103 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
2104 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
2106 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
2108 * doc/frontends.texi (G++ and GCC): Remove references to Java.
2110 2017-01-13 Jeff Law <law@redhat.com>
2112 PR tree-optimization/33562
2113 PR tree-optimization/61912
2114 PR tree-optimization/77485
2115 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
2117 (delete_dead_assignment): Likewise.
2118 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
2119 statement to delete_dead_call and delete_dead_assignment.
2121 2017-01-13 David Malcolm <dmalcolm@redhat.com>
2124 * substring-locations.c (format_warning_va): Strengthen case 1 so
2125 that both endpoints of the substring must be within the format
2126 range for just the substring to be printed.
2128 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
2130 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
2131 * config/i386/i386.c (ix86_target_string): Add missing options
2132 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
2133 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
2134 flags_other and ix86_target_other to flags2_other. Display unknown
2136 (ix86_valid_target_attribute_inner_p): Add missing options and
2137 reorder options by implied ISAs, as in ix86_target_string.
2139 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2141 * hash-table.h (hash_table::too_empty_p): New function.
2142 (hash_table::expand): Use it.
2143 (hash_table::traverse): Likewise.
2144 (hash_table::empty_slot): Use sizeof (value_type) instead of
2145 sizeof (PTR) to convert bytes to elements. Shrink the table
2146 if the current size is excessive for the current number of
2149 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2151 * ira-costs.c (record_reg_classes): Break from the inner loop
2152 early once alt_fail is known to be true. Update outer loop
2153 handling accordingly.
2155 2017-01-13 Jeff Law <law@redhat.com>
2157 * tree-ssa-dse.c (decrement_count): New function.
2158 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2159 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2160 when we know the partially dead statement is a mem* function.
2162 PR tree-optimization/61912
2163 PR tree-optimization/77485
2164 * tree-ssa-dse.c: Include expr.h.
2165 (maybe_trim_constructor_store): New function.
2166 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2168 PR tree-optimization/33562
2169 PR tree-optimization/61912
2170 PR tree-optimization/77485
2171 * doc/invoke.texi: Document new dse-max-object-size param.
2172 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2173 * tree-ssa-dse.c: Include params.h.
2174 (dse_store_status): New enum.
2175 (initialize_ao_ref_for_dse): New, partially extracted from
2177 (valid_ao_ref_for_dse, normalize_ref): New.
2178 (setup_live_bytes_from_ref, compute_trims): Likewise.
2179 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2180 (maybe_trim_partially_dead_store): Likewise.
2181 (maybe_trim_complex_store): Likewise.
2182 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2183 Track what bytes live from the original store. Return tri-state
2184 for dead, partially dead or live.
2185 (dse_dom_walker): Add constructor, destructor and new private members.
2186 (delete_dead_call, delete_dead_assignment): New extracted from
2188 (dse_optimize_stmt): Make a member of dse_dom_walker.
2189 Use initialize_ao_ref_for_dse.
2191 PR tree-optimization/33562
2192 PR tree-optimization/61912
2193 PR tree-optimization/77485
2194 * sbitmap.h (bitmap_count_bits): Prototype.
2195 (bitmap_clear_range, bitmap_set_range): Likewise.
2196 * sbitmap.c (bitmap_clear_range): New function.
2197 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2199 2017-01-13 Martin Liska <mliska@suse.cz>
2202 * function.c (set_cfun): Add new argument force.
2203 * function.h (set_cfun): Likewise.
2204 * ipa-inline-transform.c (inline_call): Use the function when
2205 strict alising from is dropped for function we inline to.
2207 2017-01-13 Richard Biener <rguenther@suse.de>
2209 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2210 for dumping GIMPLE INTEGER_CSTs.
2212 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2214 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2215 to 201112L since C++17.
2217 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2220 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2221 if -fsanitize=kernel-address is present.
2223 2017-01-13 Richard Biener <rguenther@suse.de>
2225 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2226 as _Literal ( type ) number in case usual suffixes do not
2227 preserve all information.
2229 2017-01-13 Richard Biener <rguenther@suse.de>
2231 PR tree-optimization/77283
2232 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2233 and ssa-iterators.h.
2234 (is_feasible_trace): Implement a cost model based on joiner
2237 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2240 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2241 char or short to __float128/_Float128 directly.
2243 2017-01-12 Martin Sebor <msebor@redhat.com>
2245 to -Wformat-overflow.
2246 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2247 (min_bytes_remaining): Same.
2248 (get_string_length): Same.
2249 (format_string): Same.
2250 (format_directive): Same.
2252 (pass_sprintf_length::handle_gimple_call): Same.
2254 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2256 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2257 info.nowrite calls with no lhs that can't throw. Return bool
2258 whether gsi_remove has been called or not.
2259 (pass_sprintf_length::handle_gimple_call): Return bool whether
2260 try_substitute_return_value called gsi_remove. Formatting fix.
2261 (pass_sprintf_length::execute): Don't use gsi_remove if
2262 handle_gimple_call returned true.
2265 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2266 be removed due to side-effects, don't remove following barrier nor
2267 turn the successor edge into fallthru edge.
2269 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2272 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2273 element-reversing loads and stores as not swappable.
2275 2017-01-12 Nathan Sidwell <nathan@acm.org>
2276 Nicolai Stange <nicstange@gmail.com>
2278 * combine.c (try_combine): Don't ignore result of overlap checking
2279 loop. Combine overlap & asm check into single loop.
2281 2017-01-12 Richard Biener <rguenther@suse.de>
2283 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2284 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2286 2017-01-12 Richard Biener <rguenther@suse.de>
2288 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2289 and TS_TARGET_OPTION directly derive from TS_BASE.
2290 * tree-core.h (tree_optimization_option): Derive from tree_base.
2291 (tree_target_option): Likewise.
2293 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2295 * config/i386/i386.c (memory_address_length): Increase len
2296 only when rip_relative_addr_p returns false.
2298 2017-01-11 Julia Koval <julia.koval@intel.com>
2300 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2301 (OPTION_MASK_ISA_SGX_SET): New.
2302 (ix86_handle_option): Handle OPT_msgx.
2303 * config.gcc: Added sgxintrin.h.
2304 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2305 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2306 * config/i386/i386.c (ix86_target_string): Add -msgx.
2308 (ix86_option_override_internal): Handle new options.
2309 (ix86_valid_target_attribute_inner_p): Add sgx.
2310 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2311 * config/i386/i386.opt: Add msgx.
2312 * config/i386/sgxintrin.h: New file.
2313 * config/i386/x86intrin.h: Add sgxintrin.h.
2315 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2318 * fold-const.c (maybe_nonzero_address): Return 1 for function
2320 (tree_single_nonzero_warnv_p): Don't handle function local objects
2324 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2327 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2330 * opts.c: Include "spellcheck.h"
2331 (struct string_fragment): New struct.
2332 (struct edit_distance_traits<const string_fragment &>): New
2334 (get_closest_sanitizer_option): New function.
2335 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2337 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2339 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2341 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2342 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2343 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2344 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2345 for initial die_offset if dwarf_split_debug_info.
2346 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2347 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2349 (output_skeleton_debug_sections): Formatting fix. Use
2350 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2351 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2353 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2355 * config/arm/cortex-a53.md: Add bypasses for
2357 (cortex_a53_r2f): Only use for transfers.
2358 (cortex_a53_f2r): Likewise.
2359 (cortex_a53_r2f_cvt): Add reservation for conversions.
2360 (cortex_a53_f2r_cvt): Likewise.
2362 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2364 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2365 to all inlined functions, change static to extern.
2367 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2370 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2372 (arm_assemble_integer): Likewise.
2374 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2376 * config.gcc: Use new awk script to check CPU, FPU and architecture
2377 parameters for --with-... options.
2378 * config/arm/parsecpu.awk: New file
2379 * config/arm/arm-cpus.in: New file.
2380 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2382 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2384 * config/arm/t-arm: Update dependency rules.
2385 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2386 of processing .def files.
2387 * config/arm/genopt.sh: Deleted.
2388 * config/arm/gentune.sh: Deleted.
2389 * config/arm/arm-cores.def: Deleted.
2390 * config/arm/arm-arches.def: Deleted.
2391 * config/arm/arm-fpus.def: Deleted.
2392 * config/arm/arm-tune.md: Regenerated.
2393 * config/arm/arm-tables.opt: Regenerated.
2394 * config/arm/arm-cpu.h: New generated file.
2395 * config/arm/arm-cpu-data.h: New generated file.
2396 * config/arm/arm-cpu-cdata.h: New generated file.
2398 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2401 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2403 (input_varpool_node): Unpack dynamically_initialized bit.
2405 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2407 PR rtl-optimization/79032
2408 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2409 the alignment of the adjusted memory reference against that of MODE,
2410 instead of the alignment of the original memory reference.
2412 2017-01-11 Martin Jambor <mjambor@suse.cz>
2414 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2416 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2417 decorated functions.
2419 2017-01-11 Richard Biener <rguenther@suse.de>
2421 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2422 set range/nonnull info for PHI results. Do not set it on
2423 stmts marked for removal.
2425 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2427 * expr.c (store_field): In the bitfield case, fetch the return value
2428 from the registers before applying a single big-endian adjustment.
2429 Always do a final load for a BLKmode value not larger than a word.
2431 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2434 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2435 that we correctly handle column numbers greater than
2436 LINE_MAP_MAX_COLUMN_NUMBER.
2438 2017-01-10 Martin Sebor <msebor@redhat.com>
2441 * gimple-ssa-sprintf.c (get_destination_size): Call
2442 {init,fini}object_sizes.
2443 * tree-object-size.c (addr_object_size): Adjust.
2444 (pass_through_call): Adjust.
2445 (pass_object_sizes::execute): Adjust.
2446 * tree-object-size.h (fini_object_sizes): Declare.
2448 2017-01-10 Martin Sebor <msebor@redhat.com>
2450 PR tree-optimization/78775
2451 * builtins.c (get_size_range): Move...
2452 * calls.c: ...to here.
2453 (alloc_max_size): Accept zero argument.
2454 (operand_signed_p): Remove.
2455 (maybe_warn_alloc_args_overflow): Call get_size_range.
2456 * calls.h (get_size_range): Declare.
2458 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2460 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2461 from TI's devices.csv file as of September 2016.
2462 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2464 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2466 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2467 * doc/invoke.texi: Likewise.
2468 * doc/md.texi: Likewise.
2469 * doc/objc.texi: Likewise.
2471 2017-01-10 Joshua Conner <joshconner@google.com>
2473 * config/arm/fuchsia-elf.h: New file.
2474 * config/fuchsia.h: New file.
2475 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2476 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2478 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2480 2016-01-10 Richard Biener <rguenther@suse.de>
2482 PR tree-optimization/79034
2483 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2484 Propagate out degenerate PHIs in the joiner.
2486 2017-01-10 Martin Liska <mliska@suse.cz>
2488 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2489 (sort_congruence_classes_by_decl_uid): Likewise.
2490 (sort_congruence_class_groups_by_decl_uid): Likewise.
2491 (sem_item_optimizer::merge_classes): Sort class, groups in these
2492 classes and members in the groups by DECL_UID of declarations.
2493 This would make merge operations stable.
2495 2017-01-10 Martin Liska <mliska@suse.cz>
2497 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2498 usage of m_classes_vec.
2499 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2500 (sem_item_optimizer::get_group_by_hash): Likewise.
2501 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2502 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2503 (sem_item_optimizer::verify_classes): Likewise.
2504 (sem_item_optimizer::process_cong_reduction): Likewise.
2505 (sem_item_optimizer::dump_cong_classes): Likewise.
2506 (sem_item_optimizer::merge_classes): Likewise.
2507 * ipa-icf.h (congruence_class_hash): Rename from
2508 congruence_class_group_hash. Remove declaration of m_classes_vec.
2510 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2512 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2513 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2514 * config.gcc: Add avx512vpopcntdqintrin.h.
2515 * config/i386/avx512vpopcntdqintrin.h: New.
2516 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2517 * config/i386/i386-builtin-types.def: Add new types.
2518 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2519 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2520 __builtin_ia32_vpopcountq_v8di_mask): New.
2521 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2522 __AVX512VPOPCNTDQ__.
2523 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2524 (PTA_AVX512VPOPCNTDQ): Define.
2525 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2526 TARGET_AVX512VPOPCNTDQ_P): Define.
2527 * config/i386/i386.opt: Add mavx512vpopcntdq.
2528 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2529 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2531 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2534 * predict.def (PRED_CALL): Set to 67.
2536 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2538 * expr.c (store_field): In the bitfield case, if the value comes from
2539 a function call and is of an aggregate type returned in registers, do
2540 not modify the field mode; extract the value in all cases if the mode
2541 is BLKmode and the size is not larger than a word.
2543 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2546 * config/i386/cpuid.h: Fix undefined behavior.
2548 2017-01-04 Jeff Law <law@redhat.com>
2550 PR tree-optimization/79007
2551 PR tree-optimization/67955
2552 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2553 conservative for pt.null when flag_non_call_exceptions is on.
2555 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2557 PR translation/79019
2558 PR translation/79020
2559 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2560 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2561 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2563 * config/avr/avr.opt (maccumulate-args): Likewise.
2564 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2565 * common.opt (freport-bug): Likewise.
2566 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2567 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2568 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2569 translatable string.
2570 * config/i386/i386.c (function_value_32): Likewise.
2571 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2572 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2574 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2575 * common/config/msp430/msp430-common.c (msp430_handle_option):
2577 * symtab.c (symtab_node::verify_base): Likewise.
2578 * opts.c (set_debug_level): Likewise.
2579 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2580 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2581 missing whitespace to translatable strings.
2582 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2583 * config/sh/superh.h: Likewise.
2584 * config/i386/xopintrin.h: Likewise.
2585 * config/i386/znver1.md: Likewise.
2586 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2587 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2588 * double-int.h (struct double_int): Likewise.
2589 * double-int.c (div_and_round_double): Likewise.
2590 * wide-int.cc: Likewise.
2591 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2592 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2593 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2594 (crossjumps_occurred): ... this.
2595 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2598 PR tree-optimization/78899
2599 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2600 returning bool return struct loop *, NULL for failure and the new
2602 (versionable_outer_loop_p): Don't version outer loop if it has
2603 dont_vectorized bit set.
2604 (tree_if_conversion): When versioning outer loop, ensure
2605 tree_if_conversion is performed also on the inner loop of the
2606 non-vectorizable outer loop copy.
2607 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2608 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2609 prevent vectorization of it.
2610 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2611 the outer loop vectorization of the non-scalar version is attempted
2612 before vectorization of the inner loop in scalar version. If
2613 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2614 vectorization of its inner loop.
2615 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2616 has 2 inner loops, rename also on edges from bb whose single pred
2617 is outer_loop->header. Fix typo in function comment.
2619 2017-01-09 Martin Sebor <msebor@redhat.com>
2622 * asan.c (asan_emit_stack_protection): Increase local buffer size
2623 to avoid snprintf truncation warning.
2625 2017-01-09 Andrew Pinski <apinski@cavium.com>
2627 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2628 to reference thunderx2t99 for the tuning structure
2629 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2631 (thunderx2t99_extra_costs): This.
2632 * config/aarch64/aarch64-tune.md: Regenerate.
2633 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2634 (vulcan_addrcost_table): This.
2635 (vulcan_regmove_cost): Rename to ...
2636 (thunderx2t99_regmove_cost): This.
2637 (vulcan_vector_cost): Rename to ...
2638 (thunderx2t99_vector_cost): this.
2639 (vulcan_branch_cost): Rename to ...
2640 (thunderx2t99_branch_cost): This.
2641 (vulcan_tunings): Rename to ...
2642 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2643 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2645 2017-01-09 Martin Jambor <mjambor@suse.cz>
2649 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2650 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2651 (propagate_vr_accross_jump_function): Use the above function for all
2652 value range computations for pass-through jump functions and type
2653 converasion from explicit value range values.
2654 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2655 parameters from TYPE_ARG_TYPES.
2656 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2657 (ipa_write_node_info): Stream type of the actual argument.
2658 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2660 2017-01-09 Martin Liska <mliska@suse.cz>
2663 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2664 (lookup_compiler): Do not show error message with have_E.
2666 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2668 PR tree-optimization/78938
2669 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2670 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2671 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2672 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2675 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2677 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2680 2017-01-09 Richard Biener <rguenther@suse.de>
2682 PR tree-optimization/78997
2683 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2684 name condition properly.
2686 2017-01-09 Richard Biener <rguenther@suse.de>
2689 * dwarf2out.c (is_cxx): New overload with context.
2690 (is_naming_typedef_decl): Use it.
2692 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2694 * invoke.texi (Option Summary): Correct spacing in option lists
2695 and add line breaks to fix over-long lines.
2697 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2701 * extend.texi (Common Variable Attributes): Add xref to GCC
2702 Internals manual to explain mode attribute keywords.
2704 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2707 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2708 and Preprocessor Options.
2709 (Options for Linking): Document -pthread here....
2710 (RS/6000 and PowerPC Options): ...not here.
2711 (Solaris 2 Options): ...or here.
2712 * doc/cppopts.texi: Document -pthread.
2714 2017-01-08 Martin Sebor <msebor@redhat.com>
2717 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2718 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2719 New member functions.
2720 (format_directive): Used them.
2722 (pass_sprintf_length::handle_gimple_call): Same.
2723 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2724 to avoid truncation for any argument.
2725 (extract_affine_mul): Same.
2726 * tree.c (get_file_function_name): Same.
2728 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2731 * predict.def (PRED_INDIR_CALL): Set to 86.
2733 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2735 PR preprocessor/54124
2736 * doc/cppopts.texi: Reformat -d subtable to list the full name
2737 of the options. Add cross-reference to the docs for the general
2738 compiler -d options.
2739 * doc/invoke.texi (Developer Options): Add cross-reference to the
2740 preprocessor-specific -d option documentation.
2742 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2744 PR preprocessor/13498
2745 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2746 redudant material, and reflect new command-line options.
2747 (System Headers): Likewise.
2749 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2751 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2752 -isystem, and -idirafter. Copy-edit.
2753 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2754 default for -ftrack-macro-expansion. Delete obsolete and
2755 badly-formatted implementation details about -fdebug-cpp output.
2756 * doc/cppwarnopts.texi: Copy-edit.
2758 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2761 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2762 that the transition from a max line width >= 1<<10 to narrower
2763 lines works correctly.
2765 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2767 * doc/options.texi (PerFunction): New.
2768 * opt-functions.awk (switch_flags): Map both Optimization and
2769 PerFunction to CL_OPTIMIZATION.
2770 * opth-gen.awk: Test for PerFunction flag along with
2772 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2773 it only when the latter is present. Skip those that don't in
2774 the hash function generator.
2775 * common.opt (fvar-tracking): Mark as PerFunction instead of
2777 (fvar-tracking-assignments): Likewise.
2778 (fvar-tracking-assignments-toggle): Likewise.
2779 (fvar-tracking-uninit): Likewise.
2781 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2783 PR translation/79018
2784 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2787 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2790 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2791 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2792 TARGET_LONG_JUMP_TABLE_OFFSETS.
2793 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2794 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2795 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2796 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2797 * config/m68k/m68k.md (tablejump expander): Likewise.
2798 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2799 TARGET_LONG_JUMP_TABLE_OFFSETS.
2800 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2801 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2803 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2804 David Holsgrove <david.holsgrove@xilinx.com>
2806 * common/config/microblaze/microblaze-common.c
2807 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2808 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2810 * config/microblaze/microblaze.c (microblaze_must_save_register)
2811 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2813 (microblaze_eh_return): New function.
2814 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2815 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2816 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2817 * config/microblaze/microblaze.md (eh_return): New pattern.
2819 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2821 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2822 GCC_DIAGNOSTIC_STRINGIFY): Define.
2824 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2826 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2828 * config/arm/arm.md (<mcrr>): New.
2830 * config/arm/arm.c (arm_arch5te): New.
2831 (arm_option_override): Set arm_arch5te.
2832 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2834 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2835 (arm_mcrr_qualifiers): ... this. New.
2836 (MRRC_QUALIFIERS): Define to...
2837 (arm_mrrc_qualifiers): ... this. New.
2838 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2840 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2841 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2842 (MRRCI, mrrc, MRRC): New.
2843 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2844 VUNSPEC_MRRC2): New.
2846 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2848 * config/arm/arm.md (<mcr>): New.
2850 * config/arm/arm.c (arm_coproc_builtin_available): Add
2851 support for mcr, mrc, mcr2 and mrc2.
2852 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2853 (arm_mcr_qualifiers): ... this. New.
2854 (MRC_QUALIFIERS): Define to ...
2855 (arm_mrc_qualifiers): ... this. New.
2856 (MCR_QUALIFIERS): Define to ...
2857 (arm_mcr_qualifiers): ... this. New.
2858 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2860 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2861 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2862 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2865 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2867 * config/arm/arm.md (*ldc): New.
2871 * config/arm/arm.c (arm_coproc_builtin_available): Add
2872 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2873 (arm_coproc_ldc_stc_legitimate_address): New.
2874 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2875 'qualifier_const_pointer'.
2876 (LDC_QUALIFIERS): Define to...
2877 (arm_ldc_qualifiers): ... this. New.
2878 (STC_QUALIFIERS): Define to...
2879 (arm_stc_qualifiers): ... this. New.
2880 * config/arm/arm-protos.h
2881 (arm_coproc_ldc_stc_legitimate_address): New.
2882 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2883 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2884 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2885 stc2, stcl, stc2l): New.
2886 * config/arm/constraints.md (Uz): New.
2887 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2888 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2889 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2890 VUNSPEC_STC2L): New.
2892 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2894 * config/arm/arm.md (<cdp>): New.
2895 * config/arm/arm.c (neon_const_bounds): Rename this ...
2896 (arm_const_bounds): ... this.
2897 (arm_coproc_builtin_available): New.
2898 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2899 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2900 (CDP_QUALIFIERS): Define to...
2901 (arm_cdp_qualifiers): ... this. New.
2903 (arm_expand_builtin_args): Add case for 6 arguments.
2904 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2905 (arm_const_bounds): ... this.
2906 (arm_coproc_builtin_available): New.
2907 * config/arm/arm_acle.h (__arm_cdp): New.
2909 * config/arm/arm_acle_builtins.def (cdp): New.
2911 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2912 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2914 * config/arm/types.md (coproc): New.
2915 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2916 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2917 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2918 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2920 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2922 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2923 (UBINOP_QUALIFIERS): New.
2925 (acle_builtin_data): New. Change comment.
2926 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2927 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2928 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2929 arm_acle_builtins.def.
2930 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2931 (arm_init_acle_builtins): New.
2932 (CRC32_BUILTIN): Remove.
2933 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2934 crc32cb, crc32ch and crc32cw.
2935 (arm_init_crc32_builtins): Remove.
2936 (arm_init_builtins): Use arm_init_acle_builtins rather
2937 than arm_init_crc32_builtins.
2938 (arm_expand_acle_builtin): New.
2939 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2940 * config/arm/arm_acle_builtins.def: New.
2942 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2944 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2945 (arm_builtin_datum): ... this.
2946 (arm_init_neon_builtin): Rename to ...
2947 (arm_init_builtin): ... this. Add a new parameters PREFIX
2948 and USE_SIG_IN_NAME.
2949 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2950 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2951 'arm_builtin_datum'.
2952 (arm_init_vfp_builtins): Likewise.
2953 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2954 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2955 (arm_expand_neon_args): Rename to ...
2956 (arm_expand_builtin_args): ... this. Rename builtin_arg
2957 enum values and differentiate between ARG_BUILTIN_MEMORY
2958 and ARG_BUILTIN_NEON_MEMORY.
2959 (arm_expand_neon_builtin_1): Rename to ...
2960 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2961 values, arm_expand_builtin_args and add bool parameter NEON.
2962 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2963 (arm_expand_vfp_builtin): Likewise.
2964 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2966 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2969 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2970 * predict.c (tree_estimate_probability_bb): Reverse direction of
2971 polymorphic call predictor.
2973 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2975 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2976 (determine_pass_name_match): ...this new function and...
2977 (should_skip_pass_p): ...this new function.
2979 2017-01-06 Nathan Sidwell <nathan@acm.org>
2981 * ipa-visibility.c (function_and_variable_visibility): Reformat
2982 comments and long lines. Remove extrneous if.
2983 * symtab.c (symtab_node::make_decl_local): Fix code format.
2984 (symtab_node::set_section_for_node): Fix comment typo.
2986 2017-01-06 Martin Liska <mliska@suse.cz>
2989 * lra-constraints.c: Rename invariant to lra_invariant.
2990 * predict.c (set_even_probabilities): Initialize e to NULL.
2992 2017-01-05 Martin Sebor <msebor@redhat.com>
2994 PR tree-optimization/78910
2995 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2996 (format_integer): Correct off-by-one error in the handling
2997 of precision with negative numbers in signed conversions..
2999 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3001 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3003 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3005 PR tree-optimization/71016
3006 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3007 factor_out_conditional_conversion. Formatting fix.
3008 (factor_out_conditional_conversion): Add cond_stmt argument.
3009 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3010 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3013 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3015 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3016 read-rtl-function.o, and selftest-rtl.o.
3017 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3018 (selftest::aarch64_test_loading_full_dump): New function.
3019 (selftest::aarch64_run_selftests): New function.
3020 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3021 selftest::aarch64_run_selftests.
3022 * config/i386/i386.c
3023 (selftest::ix86_test_loading_dump_fragment_1): New function.
3024 (selftest::ix86_test_loading_call_insn): New function.
3025 (selftest::ix86_test_loading_full_dump): New function.
3026 (selftest::ix86_test_loading_unspec): New function.
3027 (selftest::ix86_run_selftests): Call the new functions.
3028 * emit-rtl.c (maybe_set_max_label_num): New function.
3029 * emit-rtl.h (maybe_set_max_label_num): New decl.
3030 * function.c (instantiate_decls): Guard call to
3031 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3032 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3034 * gensupport.c (gen_reader::gen_reader): Pass "false"
3035 for new "compact" param of rtx_reader.
3036 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3037 rather than an empty string for NULL strings.
3038 * read-md.c: Potentially include config.h rather than bconfig.h.
3039 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3040 (have_error): New global, copied from errors.c.
3041 (md_reader::read_name): Rename to...
3042 (md_reader::read_name_1): ...this, adding "out_loc" param,
3043 and converting "missing name or number" to returning false, rather
3045 (md_reader::read_name): Reimplement in terms of read_name_1.
3046 (md_reader::read_name_or_nil): New function.
3047 (md_reader::read_string): Handle "(nil)" by returning NULL.
3048 (md_reader::md_reader): Add new param "compact".
3049 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3050 (md_reader::read_file): New method.
3051 * read-md.h (md_reader::md_reader): Add new param "compact".
3052 (md_reader::read_file): New method.
3053 (md_reader::is_compact): New accessor.
3054 (md_reader::read_name): Convert return type from void to file_location.
3055 (md_reader::read_name_or_nil): New decl.
3056 (md_reader::read_name_1): New decl.
3057 (md_reader::m_compact): New field.
3058 (noop_reader::noop_reader): Pass "false" for new "compact" param
3060 (rtx_reader::rtx_reader): Add new "compact" param.
3061 (rtx_reader::read_rtx_operand): Make virtual and convert return
3062 type from void to rtx.
3063 (rtx_reader::read_until): New decl.
3064 (rtx_reader::handle_any_trailing_information): New virtual function.
3065 (rtx_reader::postprocess): New virtual function.
3066 (rtx_reader::finalize_string): New virtual function.
3067 (rtx_reader::m_in_call_function_usage): New field.
3068 (rtx_reader::m_reuse_rtx_by_id): New field.
3069 * read-rtl-function.c: New file.
3070 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3071 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3072 (selftest::verify_three_block_rtl_cfg): New decl.
3073 * read-rtl-function.h: New file.
3074 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3075 For host, include function.h, memmodel.h, and emit-rtl.h.
3076 (one_time_initialization): New function.
3077 (struct compact_insn_name): New struct.
3078 (compact_insn_names): New array.
3079 (find_code): Handle insn codes in compact dumps.
3080 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
3081 (bind_subst_iter_and_attr): Likewise.
3082 (add_condition_to_string): Likewise.
3083 (add_condition_to_rtx): Likewise.
3084 (apply_attribute_uses): Likewise.
3085 (add_current_iterators): Likewise.
3086 (apply_iterators): Likewise.
3087 (initialize_iterators): Guard usage of apply_subst_iterator with
3088 #ifdef GENERATOR_FILE.
3089 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
3090 (md_reader::read_mapping): Likewise.
3091 (add_define_attr_for_define_subst): Likewise.
3092 (add_define_subst_attr): Likewise.
3093 (read_subst_mapping): Likewise.
3094 (check_code_iterator): Likewise.
3095 (rtx_reader::read_rtx): Likewise. Move one-time initialization
3097 (one_time_initialization): New function.
3098 (rtx_reader::read_until): New method.
3099 (read_flags): New function.
3100 (parse_reg_note_name): New function.
3101 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
3102 Handle reuse_rtx ids.
3103 Wrap iterator lookup within #ifdef GENERATOR_FILE.
3104 Add parsing support for RTL dumps, mirroring the special-cases in
3105 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
3106 values, and calling handle_any_trailing_information.
3107 (rtx_reader::read_rtx_operand): Convert return type from void
3108 to rtx, returning return_rtx. Handle case 'e'. Call
3109 finalize_string on XSTR and XTMPL fields.
3110 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
3111 "(nil)" values were omitted. Call the postprocess vfunc on the
3113 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
3114 class ctor. Initialize m_in_call_function_usage. Call
3115 one_time_initialization.
3116 * rtl-tests.c (selftest::test_uncond_jump): Call
3117 set_new_first_and_last_insn.
3118 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
3119 * selftest-rtl.c: New file.
3120 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
3121 (selftest::get_insn_by_uid): New decl.
3122 * selftest-run-tests.c (selftest::run_tests): Call
3123 read_rtl_function_c_tests.
3124 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
3125 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
3128 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
3130 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
3131 operands in a special way. Assert that pos+len <= mode precision.
3133 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3135 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
3136 3 argument Alias with unlimited for the negative form.
3137 (fno-vect-cost-model): Removed.
3139 2017-01-05 Martin Liska <mliska@suse.cz>
3141 * hsa-gen.c (gen_hsa_divmod): New function.
3142 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3144 2017-01-05 Martin Liska <mliska@suse.cz>
3147 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3150 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3152 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3153 small constant length operands.
3155 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3157 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3158 between loop iterations.
3160 2017-01-05 Martin Liska <mliska@suse.cz>
3163 * gimplify.c (gimplify_decl_expr): Compare to
3164 asan_poisoned_variables instread of checking flags.
3165 (gimplify_target_expr): Likewise.
3166 (gimplify_expr): Likewise.
3167 (gimplify_function_tree): Conditionally initialize
3168 asan_poisoned_variables.
3170 2017-01-04 Jeff Law <law@redhat.com>
3172 PR tree-optimizatin/78812
3173 * rtl.h (contains_mem_rtx_p): Prototype.
3174 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3175 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3176 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3177 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3178 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3180 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3182 * input.c (assert_char_at_range): Default-initialize actual_range.
3184 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3186 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3187 to match the caller.
3189 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3191 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3192 insns after final jump in test to emit dummy move.
3194 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3196 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3197 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3199 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3201 * multiple_target.c (create_dispatcher_calls): Init e_next.
3202 * tree-ssa-loop-split.c (split_loop): Init border.
3203 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3206 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3211 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3212 (altivec_register_operand): Do not return true if the operand
3213 contains a SUBREG mixing SImode and SFmode.
3214 (vsx_register_operand): Likewise.
3215 (vsx_reg_sfsubreg_ok): New predicate.
3216 (vfloat_operand): Do not return true if the operand contains a
3217 SUBREG mixing SImode and SFmode.
3218 (vint_operand): Likewise.
3219 (vlogical_operand): Likewise.
3220 (gpc_reg_operand): Likewise.
3221 (int_reg_operand): Likewise.
3222 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3223 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3224 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3226 (rs6000_emit_move_si_sf_subreg): New helper function.
3227 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3228 fixup SUBREGs involving SImode and SFmode.
3229 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3230 numbers for the new peephole2 optimization.
3231 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3232 the GLIBC math library that do AND/IOR/XOR operations on single
3233 precision floating point.
3234 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3235 target macros to say whether we need to avoid SUBREGs mixing
3237 (TARGET_ALLOW_SF_SUBREG): Likewise.
3238 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3239 (UNSPEC_SI_FROM_SF): Likewise.
3240 (iorxor): Change spacing.
3241 (and_ior_xor): New iterator for AND, IOR, and XOR.
3242 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3243 (movdi_from_sf_zero_ext): Likewise.
3244 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3245 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3246 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3247 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3248 (fms<mode>4): Likewise.
3249 (fnma<mode>4): Likewise.
3250 (fnms<mode>4): Likewise.
3251 (nfma<mode>4): Likewise.
3252 (nfms<mode>4): Likewise.
3254 2017-01-04 Marek Polacek <polacek@redhat.com>
3257 * doc/invoke.texi: Document -Wpointer-compare.
3259 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3261 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3264 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3265 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3268 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3271 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3272 documentation of the powerpc_popcntb_ok attribute.
3273 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3274 code to issue warning messages if a requested CPU configuration is
3275 not supported by the binary (assembler and loader) toolchain.
3276 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3277 made to define a built-in function that has been disabled.
3278 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3279 made to define a built-in function that has been disabled.
3280 (altivec_init_builtins): Add comment explaining why definition
3281 of the DST built-in functions is not preceded by an assertion
3282 check. Add assertions to prevent ICE if attempts are made to
3283 define an altivec predicate or an abs* built-in function that has
3285 (htm_init_builtins): Add comment explaining why definition of the
3286 htm built-in functions is not preceded by an assertion check.
3288 2017-01-04 Jeff Law <law@redhat.com>
3290 PR tree-optimizatin/67955
3291 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3292 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3293 the points-to solution does not include pt_null. Use DECL_PT_UID
3296 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3298 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3299 Use gen_int_mode instead of gen_lopwart for const_int operands.
3301 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3303 PR tree-optimization/71563
3304 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3305 out of range value - has low bits known to be zero.
3307 2017-01-04 Alan Modra <amodra@gmail.com>
3309 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3310 * configure: Regenerate.
3311 * config.in: Regenerate.
3313 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3316 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3317 a substring of the message, but strcmp with the whole message. Ifdef
3318 ENABLE_NLS, translate the message first using dgettext.
3320 2017-01-03 Jeff Law <law@redhat.com>
3322 PR tree-optimizatin/78856
3323 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3324 (mark_threaded_blocks): Remove code to truncate thread paths that
3325 cross multiple loop headers. Instead invalidate the cached loop
3326 iteration information and handle case of a thread path walking
3327 into an irreducible region.
3329 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3332 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3333 assertions. Add support for doing the signbit if the IEEE 128-bit
3334 floating point value is in a GPR.
3335 * config/rs6000/rs6000.md (Fsignbit): Delete.
3336 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3337 Update the length attribute if the value is in a GPR.
3338 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3339 the sign or zero extension instruction, since the value is always 0/1.
3340 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3343 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3344 extracting SImode to a GPR register so that we can generate a
3345 store, limit the vector to be in a traditional Altivec register
3346 for the vextuwrx instruction.
3348 2017-01-03 Ian Lance Taylor <iant@google.com>
3350 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3352 2017-01-03 Martin Sebor <msebor@redhat.com>
3354 PR tree-optimization/78696
3355 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3356 precision. Use MPFR for %f for greater fidelity. Correct handling
3358 (pass_sprintf_length::compute_format_length): Set width and precision
3359 specified by asrerisk to void_node for vararg functions.
3360 (try_substitute_return_value): Adjust dump output.
3362 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3364 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3366 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3368 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3369 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3370 -mlra/-mno-lra was passed to the compiler.
3372 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3374 PR rtl-optimization/65618
3375 * emit-rtl.c (try_split): Move initialization of "before" and
3376 "after" to just before the call to emit_insn_after_setloc.
3378 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3380 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3382 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3384 * dwarf2out.c (gen_enumeration_type_die): When
3385 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3387 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3389 PR tree-optimization/78965
3390 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3391 Change first argument from const call_info & to call_info &. For %n
3392 set info.nowrite to false.
3395 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3396 possibly throwing calls.
3398 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3399 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3400 and fns handling, rather than in a separate case SSA_NAME.
3402 2017-01-02 Jeff Law <law@redhat.com>
3404 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3405 fixes for first_period and second_period variables.
3407 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3410 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3411 (*insvqi_1): New insn pattern.
3412 (*insvqi_1_mem_rex64): Ditto.
3414 (*insvqi_3): Rename from *insvqi.
3416 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3418 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3420 * doc/cfg.texi (Edges): Remove reference to Java.
3421 (Maintaining the CFG): Ditto.
3423 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3426 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3427 transparent aliases.
3429 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3432 * predict.def (PRED_CALL): Update hitrate.
3433 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3434 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3435 into direct/indirect/polymorphic variants.
3437 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3439 Update copyright years.
3441 * gcc.c (process_command): Update copyright notice dates.
3442 * gcov-dump.c (print_version): Ditto.
3443 * gcov.c (print_version): Ditto.
3444 * gcov-tool.c (print_version): Ditto.
3445 * gengtype.c (create_file): Ditto.
3446 * doc/cpp.texi: Bump @copying's copyright year.
3447 * doc/cppinternals.texi: Ditto.
3448 * doc/gcc.texi: Ditto.
3449 * doc/gccint.texi: Ditto.
3450 * doc/gcov.texi: Ditto.
3451 * doc/install.texi: Ditto.
3452 * doc/invoke.texi: Ditto.
3454 Copyright (C) 2017 Free Software Foundation, Inc.
3456 Copying and distribution of this file, with or without modification,
3457 are permitted in any medium without royalty provided the copyright
3458 notice and this notice are preserved.