1 2017-02-05 Martin Liska <mliska@suse.cz>
4 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
6 (print_operand_address): Initialize a struct to zero.
8 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
10 * doc/extend.texi (x86 specific memory model extensions for
11 transactional memory): Simplify a phrase.
13 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
16 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
17 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
18 (atomic_storedi_1): Likewise.
20 2017-02-04 Jakub Jelinek <jakub@redhat.com>
22 PR tree-optimization/79338
23 * tree-parloops.c (gather_scalar_reductions): Don't call
24 vect_analyze_loop_form for loop->inner before destroying loop's
27 2017-02-03 Martin Sebor <msebor@redhat.com>
29 PR tree-optimization/79327
30 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
31 when precision has resulted in leading zeros.
32 (format_integer): Adjust the likely counter to assume an unknown
33 argument that may be zero is non-zero.
35 2017-02-03 Jason Merrill <jason@redhat.com>
38 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
39 avoid copying non-taken branch.
41 2017-02-03 Jakub Jelinek <jakub@redhat.com>
43 PR tree-optimization/79340
44 * tree-vect-loop.c (vectorizable_reduction): Release
45 vec_defs elements after safe_splicing them into other vectors.
48 PR tree-optimization/79327
49 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
50 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
52 (format_integer): Use wide_int_to_tree instead of build_int_cst
53 + to_?hwi. If argmin is NULL, just set argmin and argmax to
54 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
55 of shortest and longest sequence.
57 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
59 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
60 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
62 2017-02-03 Walter Lee <walt@tilera.com>
65 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
66 after initial stackframe link reg save.
67 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
69 2017-02-03 Jakub Jelinek <jakub@redhat.com>
72 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
73 wu for stxssp alternative.
75 2017-02-03 Martin Sebor <msebor@redhat.com>
77 PR tree-optimization/79352
78 * gimple-fold.c (get_range_strlen): Add argument.
79 (get_range_strlen): Change return type to bool.
80 (get_maxval_strlen): Pass in a dummy argument.
81 * gimple-fold.h (get_range_strlen): Change return type to bool.
82 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
83 * tree.h (array_at_struct_end_p): Add argument.
84 * tree.c (array_at_struct_end_p): Handle it.
86 2017-02-03 Martin Liska <mliska@suse.cz>
89 * multiple_target.c (create_dispatcher_calls): Redirect edge
90 from a caller of a dispatcher.
91 (expand_target_clones): Make the clones local.
92 (ipa_target_clone): Do both target clones and resolvers.
93 (ipa_dispatcher_calls): Remove the pass.
94 (pass_dispatcher_calls::gate): Likewise.
95 (make_pass_dispatcher_calls): Likewise.
96 * passes.def (pass_target_clone): Put as very first IPA early
99 2017-02-03 Martin Liska <mliska@suse.cz>
101 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
102 in case of a function with ifunc attribute.
104 2017-02-03 Martin Liska <mliska@suse.cz>
106 * cgraph.c (cgraph_node::dump): Dump function version info.
107 * symtab.c (symtab_node::dump_base): Add missing new line.
109 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
111 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
112 (ifcombine_ifandif): Use it.
114 2017-02-03 Martin Liska <mliska@suse.cz>
116 * doc/invoke.texi: Document default value for
117 use-after-scope-direct-emission-threshold.
119 2017-02-03 Martin Liska <mliska@suse.cz>
121 PR tree-optimization/79339
122 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
123 (format_floating): Likewise.
125 2017-02-03 Martin Liska <mliska@suse.cz>
128 * ipa-prop.c (ipa_node_params_t::insert): Remove current
130 (ipa_node_params_t::remove): Likewise.
131 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
132 initialization from removed ipa_node_params_t::insert.
133 (ipa_node_params::~ipa_node_params): Move from removed
134 ipa_node_params_t::release.
135 * symbol-summary.h (symbol_summary::m_released): New member.
136 Do not release a summary twice. Do not allow to call finalizer
137 for types of a summary that live in GGC memory.
139 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
141 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
144 2017-02-02 Martin Sebor <msebor@redhat.com>
147 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
148 (format_string): Tighten up the range of output for non-constant
149 strings and correct the expected range for wide non-constant strings.
151 2017-02-02 Martin Sebor <msebor@redhat.com>
153 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
156 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
158 (-fdump-tree-@var): Add to index and document how to come up
159 with pass-specific option and dump file names.
160 (-fdump-passes): Clarify where to look for output.
162 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
165 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
166 statistics of the analyzed path; allow threading for speed when
167 any of BBs along the path are optimized for speed.
169 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
172 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
173 settings of the virtual registers.
176 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
178 * explow.c (get_dynamic_stack_size): Take known alignment of stack
179 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
182 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
184 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
185 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
187 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
189 * config/s390/s390.md: Add missing comments with the expanded
191 * config/s390/vector.md: Likewise.
192 * config/s390/vx-builtins.md: Likewise.
194 2017-02-02 Jakub Jelinek <jakub@redhat.com>
197 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
198 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
199 conditions on a single line.
201 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
203 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
204 __S390_VX__ to __VX__.
206 2017-02-01 Andrew Pinski <apinski@cavium.com>
208 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
209 stmt_info to record_stmt_cost.
210 (vect_get_known_peeling_cost): Pass stmt_info if known to
212 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
213 cpu_vector_cost field into
214 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
215 field into vec_int_stmt_cost and vec_fp_stmt_cost.
216 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
217 splitting of scalar_stmt_cost and vec_stmt_cost.
218 (thunderx_vector_cost): Likewise.
219 (cortexa57_vector_cost): LIkewise.
220 (exynosm1_vector_cost): Likewise.
221 (xgene1_vector_cost): Likewise.
222 (thunderx2t99_vector_cost): Improve after the splitting of the two
224 (aarch64_builtin_vectorization_cost): Update for the splitting of
225 scalar_stmt_cost and vec_stmt_cost.
227 2017-02-01 Torvald Riegel <triegel@redhat.com>
228 Richard Henderson <rth@redhat.com>
230 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
231 conditional on existance of a fast atomic load.
232 * optabs-query.c (can_atomic_load_p): New function.
233 * optabs-query.h (can_atomic_load_p): Declare it.
234 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
235 no fast atomic load is available for the particular size of access.
236 (expand_atomic_compare_and_swap): Likewise.
237 (expand_atomic_load): Likewise.
238 (expand_atomic_store): Likewise.
239 (expand_atomic_fetch_op): Likewise.
240 * testsuite/lib/target-supports.exp
241 (check_effective_target_sync_int_128): Remove x86 because it provides
243 (check_effective_target_sync_int_128_runtime): Likewise.
245 2017-02-01 Richard Biener <rguenther@suse.de>
247 * graphite.c: Include tree-vectorizer.h for find_loop_location.
248 (graphite_transform_loops): Provide opt-info for optimized nests.
249 * tree-parloop.c (parallelize_loops): Provide opt-info for
252 2017-02-01 Richard Biener <rguenther@suse.de>
255 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
258 2017-02-01 Richard Biener <rguenther@suse.de>
260 PR tree-optimization/71824
261 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
262 Verify the loops are valid in the merged SESE region.
263 (scop_detection::can_represent_loop_1): Check analyzing the
264 evolution of the number of iterations in the region succeeds.
266 2017-01-31 Ian Lance Taylor <iant@golang.org>
268 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
269 REG_ARGS_SIZE note to 32-bit push insns and call insn.
271 2017-01-31 David Malcolm <dmalcolm@redhat.com>
273 PR preprocessor/79210
274 * input.c (get_substring_ranges_for_loc): Replace line_width
275 assertion with error-handling.
277 2017-01-31 Richard Biener <rguenther@suse.de>
279 PR tree-optimization/77318
280 * graphite-sese-to-poly.c (extract_affine): Fix assert.
281 (create_pw_aff_from_tree): Take loop parameter.
282 (add_condition_to_pbb): Pass loop of the condition to
283 create_pw_aff_from_tree.
285 2017-01-31 Jakub Jelinek <jakub@redhat.com>
287 * config/s390/s390.c (s390_asan_shadow_offset): New function.
288 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
290 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
294 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
296 (convert_int_to_float128): Likewise.
297 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
298 (convert_int_to_float128): Likewise.
299 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
300 (UNSPEC_IEEE128_CONVERT): Likewise.
301 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
302 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
303 Use local variables for IBM extended format.
304 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
305 (fix_trunc<mode>si2_fprs): Likewise.
306 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
307 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
308 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
309 to know that we can now have integers of all sizes in vector
311 (fix<uns>_<mode>di2_hw): Likewise.
312 (float<uns>_<mode>si2_hw): Likewise.
313 (fix_<mode>si2_hw): Likewise.
314 (fixuns_<mode>si2_hw): Likewise.
315 (float<uns>_<mode>di2_hw): Likewise.
316 (float_<mode>di2_hw): Likewise.
317 (float_<mode>si2_hw): Likewise.
318 (floatuns_<mode>di2_hw): Likewise.
319 (floatuns_<mode>si2_hw): Likewise.
320 (xscvqp<su>wz_<mode>): Delete, no longer used.
321 (xscvqp<su>dz_<mode>): Likewise.
322 (xscv<su>dqp_<mode>): Likewise.
323 (ieee128_mfvsrd_64bit): Likewise.
324 (ieee128_mfvsrd_32bit): Likewise.
325 (ieee128_mfvsrwz): Likewise.
326 (ieee128_mtvsrw): Likewise.
327 (ieee128_mtvsrd_64bit): Likewise.
328 (ieee128_mtvsrd_32bit): Likewise.
330 2017-01-31 Martin Liska <mliska@suse.cz>
333 * ipa-prop.c (ipa_free_all_node_params): Call release method
334 instead of ~sumbol_summary to not to trigger double times
337 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
339 PR tree-optimization/71691
340 * bitmap.h (class auto_bitmap): New.
341 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
342 is_maybe_undefined instead of ssa_undefined_value_p.
344 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
346 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
347 __S390_ARCH_LEVEL__ to __ARCH__.
349 2017-01-31 Jakub Jelinek <jakub@redhat.com>
351 PR tree-optimization/79267
352 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
353 if should_remove_lhs_p is true.
355 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
358 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
359 (add_alignment_attribute): New.
360 (base_type_die): Add alignment attribute.
361 (subrange_type_die): Likewise.
362 (modified_type_die): Likewise.
363 (gen_array_type_die): Likewise.
364 (gen_descr_array_type_die: Likewise.
365 (gen_enumeration_type_die): Likewise.
366 (gen_subprogram_die): Likewise.
367 (gen_variable_die): Likewise.
368 (gen_field_die): Likewise.
369 (gen_ptr_to_mbr_type_die): Likewise.
370 (gen_struct_or_union_type_die): Likewise.
371 (gen_subroutine_type_die): Likewise.
372 (gen_typedef_die): Likewise.
373 (base_type_cmp): Compare alignment attribute.
375 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
378 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
379 (setb_unsigned) New pattern for setb with CCUNS.
380 * config/rs6000/rs6000.c (expand_block_compare): Use a different
381 subfc./subfe sequence to avoid overflow problems. Generate a
382 shorter sequence with cmpld/setb for power9.
383 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
384 for generating subfc. instruction.
385 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
386 now uses this instruction.
388 2017-01-30 Ian Lance Taylor <iant@google.com>
391 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
392 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
394 2017-01-30 Martin Sebor <msebor@redhat.com>
396 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
397 Move constant to the right of a relational operator.
398 (get_mpfr_format_length, format_character, format_string): Ditto.
399 (should_warn_p, maybe_warn): Same.
401 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
403 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
406 * asan.c (get_translation_unit_decl): Remove function.
407 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
409 2017-01-30 Martin Liska <mliska@suse.cz>
411 PR gcov-profile/79259
412 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
415 2017-01-30 Martin Liska <mliska@suse.cz>
418 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
419 Initialize variables with NULL value.
421 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
424 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
426 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
428 2017-01-30 Richard Biener <rguenther@suse.de>
430 PR tree-optimization/79276
431 * tree-vrp.c (process_assert_insertions): Properly adjust common
432 when removing a duplicate.
434 * gcc.dg/torture/pr79276.c: New testcase.
436 2017-01-30 Richard Biener <rguenther@suse.de>
438 PR tree-optimization/79256
439 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
440 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
442 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
444 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
447 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
448 ("*r<noxa>sbg_<mode>_sll_bitmask")
449 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
450 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
451 Use contiguous_bitmask_nowrap_operand.
453 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
456 * config/rs6000/altivec.h (vec_xl): Revise #define.
459 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
461 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
463 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
465 PR rtl-optimization/79194
466 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
467 traps before call to bypass_conditional_jumps.
469 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
471 PR tree-optimization/71374
472 * lra-constraints.c (check_conflict_input_operands): New.
473 (match_reload): Use it.
475 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
478 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
479 account to calculate conflict_set.
481 2017-01-27 Bin Cheng <bin.cheng@arm.com>
483 PR rtl-optimization/78559
484 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
485 other_insn in combine.
487 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
489 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
490 uint16_type_node for BT_UINT16.
492 2017-01-27 David Malcolm <dmalcolm@redhat.com>
494 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
496 (GIMPLE Tests): New node.
497 (RTL Tests): New node.
499 2017-01-27 Richard Biener <rguenther@suse.de>
501 PR tree-optimization/79245
502 * tree-loop-distribution.c (distribute_loop): Apply cost
503 modeling also to detected patterns.
505 2017-01-27 Richard Biener <rguenther@suse.de>
507 PR tree-optimization/71433
508 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
509 (compare_assert_loc): New function.
510 (process_assert_insertions): Sort and optimize assert locations
511 to remove duplicates and push down identical assertions on
512 edges to their destination block.
514 2017-01-27 Richard Biener <rguenther@suse.de>
516 PR tree-optimization/79244
517 * tree-vrp.c (remove_range_assertions): Forcefully propagate
518 out SSA names even if abnormal.
520 2017-01-27 Jakub Jelinek <jakub@redhat.com>
522 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
523 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
524 instead of MPFR_RNDN.
526 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
529 * arm.c (arm_option_override): Don't call build_target_option_node
530 until after doing all option overrides.
531 (arm_valid_target_attribute_tree): Likewise.
533 2017-01-27 Martin Liska <mliska@suse.cz>
535 * doc/invoke.texi (-fprofile-arcs): Document profiling support
536 for {cd}tors and C++ {cd}tors.
538 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
540 * config/s390/s390.md ("*setmem_long_and")
541 ("*setmem_long_and_31z"): Use zero_extend instead of and.
543 2017-01-26 Martin Sebor <msebor@redhat.com>
545 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
548 2017-01-26 Martin Sebor <msebor@redhat.com>
550 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
551 HAVE_DFmode before using XFmode or DFmode.
552 (parse_directive): Avoid using the z length modifier to avoid
553 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
556 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
557 to accept adjustment as an array.
558 (get_int_range): New function.
559 (struct directive): Make width and prec arrays.
560 (directive::set_width, directive::set_precision): Call get_int_range.
561 (format_integer, format_floating): Handle width and precision ranges.
562 (format_string, parse_directive): Same.
564 2017-01-26 Jakub Jelinek <jakub@redhat.com>
567 * dwarf2out.c (generate_skeleton_bottom_up): For children with
568 comdat_type_p set, just clone them, but keep the children in the
572 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
573 which have direct callers with -fvar-tracking-assignments enabled
575 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
576 inside of type units.
578 2017-01-26 Martin Sebor <msebor@redhat.com>
581 * gimple-ssa-sprintf.c (struct result_range): Add likely and
583 (struct format_result): Replace number_chars, number_chars_min,
584 and number_chars_max with a single member of struct result_range.
586 (format_result::operator+=): Adjust.
587 (struct fmtresult): Remove bounded. Handle likely and unlikely
589 (fmtresult::adjust_for_width_or_precision): New function.
590 (fmtresult:type_max_digits): New function.
591 (bytes_remaining): Handle likely and unlikely counters.
592 (min_bytes_remaining): Remove.
593 (format_percent): Simplify.
594 (format_integer, format_floating): Set likely and unlikely counters.
595 (get_string_length, format_character, format_string): Same.
596 (format_plain, should_warn_p): New function.
597 (maybe_warn): Call should_warn_p. Update diagnostic messages
598 and handle those for all directives, including plain strings.
599 (format_directive): Handle likely and unlikely counters.
600 Remove unnecessary quoting from diagnostics. Add an informational
603 (pass_sprintf_length::compute_format_length): Simplify.
604 (try_substitute_return_value): Handle likely and unlikely counters.
606 2017-01-26 Carl Love <cel@us.ibm.com>
608 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
609 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
611 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
614 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
615 endianess for subregs into account.
616 * lra-constraints.c (lra_constraints): Do risky transformations
617 always on the first iteration.
618 * lra-lives.c (check_pseudos_live_through_calls): Add arg
619 last_call_used_reg_set.
620 (process_bb_lives): Define and use last_call_used_reg_set.
621 * lra.c (lra): Always continue after lra_constraints on the first
624 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
626 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
628 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
630 2017-01-26 Jakub Jelinek <jakub@redhat.com>
632 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
633 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
634 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
635 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
636 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
637 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
638 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
639 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
640 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
642 2017-01-26 Marek Polacek <polacek@redhat.com>
645 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
646 for the third operand.
648 2017-01-26 Jakub Jelinek <jakub@redhat.com>
651 * omp-low.c (struct omp_context): Add simt_stmt field.
652 (scan_omp_for): Return omp_context *.
653 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
654 context to the _simt_ SIMD stmt.
655 (lower_omp_for): For combined SIMD with sibling _simt_
656 SIMD, make sure to use the same decls in _looptemp_
657 clauses as in the sibling.
659 2017-01-26 David Sherwood <david.sherwood@arm.com>
662 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
665 2017-01-26 Jakub Jelinek <jakub@redhat.com>
668 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
669 emit fld b; fld a; if possible.
671 * brig-builtins.def: Update copyright years.
672 * config/arm/arm_acle_builtins.def: Update copyright years.
674 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
677 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
678 constraint instead of o for the stxsd instruction.
680 2017-01-25 Carl Love <cel@us.ibm.com>
682 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
683 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
685 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
687 * doc/invoke.texi (C++ Dialect Options): Fix typo.
689 2017-01-25 Richard Biener <rguenther@suse.de>
691 PR tree-optimization/69264
692 * target.def (vector_alignment_reachable): Improve documentation.
693 * doc/tm.texi: Regenerate.
694 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
696 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
697 earlier changes with respect to TYPE_USER_ALIGN.
698 (vector_alignment_reachable_p): Likewise. Improve dumping.
700 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
703 * config/arm/arm.md (xordi3): Force constant operand into a register
706 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
708 * doc/invoke.texi (-fstore-merging): Correct default optimization
709 levels at which it is enabled.
710 (-O): Move -fstore-merging from list to...
713 2017-01-25 Richard Biener <rguenther@suse.de>
716 * omp-expand.c: Include debug.h.
717 (expand_omp_taskreg): Make sure to generate early debug before
718 outlining anything from a function.
719 (expand_omp_target): Likewise.
720 (grid_expand_target_grid_body): Likewise.
722 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
725 * asan.c (get_translation_unit_decl): New function.
726 (asan_add_global): Extract modules file name from globals
727 TRANSLATION_UNIT_DECL name.
729 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
732 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
733 for long calls with APCS frame and VFP.
735 2017-01-24 David Malcolm <dmalcolm@redhat.com>
737 * cfg.c (original_copy_tables_initialized_p): New function.
738 * cfg.h (original_copy_tables_initialized_p): New decl.
739 * cfgrtl.c (relink_block_chain): Guard the call to
740 free_original_copy_tables with a call to
741 original_copy_tables_initialized_p.
742 * cgraph.h (symtab_node::native_rtl_p): New decl.
743 * cgraphunit.c (symtab_node::native_rtl_p): New function.
744 (symtab_node::needed_p): Don't assert for early assembly output
746 (cgraph_node::finalize_function): Set "force_output" for __RTL
748 (cgraph_node::analyze): Bail out early for __RTL functions.
749 (analyze_functions): Update assertion to support __RTL functions.
750 (cgraph_node::expand): Bail out early for __RTL functions.
751 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
753 * function.h (struct function): Update comment for field
755 * gimple-expr.c: Include "tree-pass.h".
756 (gimple_has_body_p): Return false for __RTL functions.
757 * Makefile.in (OBJS): Add run-rtl-passes.o.
758 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
760 (gcc::pass_manager::get_clean_slate): New accessor.
761 * passes.c: Include "insn-addr.h".
762 (should_skip_pass_p): Add logging. Update logic for running
763 "expand" to be compatible with both __GIMPLE and __RTL. Guard
764 property-provider override so it is only done for gimple passes.
766 (skip_pass): New function.
767 (execute_one_pass): Call skip_pass when skipping passes.
768 * read-md.c (md_reader::read_char): Support filtering
769 the input to a subset of line numbers.
770 (md_reader::md_reader): Initialize fields
771 m_first_line and m_last_line.
772 (md_reader::read_file_fragment): New function.
773 * read-md.h (md_reader::read_file_fragment): New decl.
774 (md_reader::m_first_line): New field.
775 (md_reader::m_last_line): New field.
776 * read-rtl-function.c (function_reader::create_function): Only
777 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
778 curr_properties. Set DECL_INITIAL to a dummy block.
779 (read_rtl_function_body_from_file_range): New function.
780 * read-rtl-function.h (read_rtl_function_body_from_file_range):
782 * run-rtl-passes.c: New file.
783 * run-rtl-passes.h: New file.
785 2017-01-24 Jeff Law <law@redhat.com>
787 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
790 2017-01-24 Bin Cheng <bin.cheng@arm.com>
792 PR tree-optimization/79159
793 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
794 (record_nonwrapping_iv): Improve boundary using above function if no
795 value range information.
797 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
798 Martin Jambor <mjambor@suse.cz>
800 * brig-builtins.def: New file.
801 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
802 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
803 (DEF_HSAIL_SAT_BUILTIN): Likewise.
804 (DEF_HSAIL_INTR_BUILTIN): Likewise.
805 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
806 * builtin-types.def (BT_INT8): New.
807 (BT_INT16): Likewise.
808 (BT_UINT8): Likewise.
809 (BT_UINT16): Likewise.
810 (BT_FN_ULONG): Likewise.
811 (BT_FN_UINT_INT): Likewise.
812 (BT_FN_UINT_ULONG): Likewise.
813 (BT_FN_UINT_LONG): Likewise.
814 (BT_FN_UINT_PTR): Likewise.
815 (BT_FN_ULONG_PTR): Likewise.
816 (BT_FN_INT8_FLOAT): Likewise.
817 (BT_FN_INT16_FLOAT): Likewise.
818 (BT_FN_UINT32_FLOAT): Likewise.
819 (BT_FN_UINT16_FLOAT): Likewise.
820 (BT_FN_UINT8_FLOAT): Likewise.
821 (BT_FN_UINT64_FLOAT): Likewise.
822 (BT_FN_UINT16_UINT32): Likewise.
823 (BT_FN_UINT32_UINT16): Likewise.
824 (BT_FN_UINT16_UINT16_UINT16): Likewise.
825 (BT_FN_INT_PTR_INT): Likewise.
826 (BT_FN_UINT_PTR_UINT): Likewise.
827 (BT_FN_LONG_PTR_LONG): Likewise.
828 (BT_FN_ULONG_PTR_ULONG): Likewise.
829 (BT_FN_VOID_UINT64_UINT64): Likewise.
830 (BT_FN_UINT8_UINT8_UINT8): Likewise.
831 (BT_FN_INT8_INT8_INT8): Likewise.
832 (BT_FN_INT16_INT16_INT16): Likewise.
833 (BT_FN_INT_INT_INT): Likewise.
834 (BT_FN_UINT_FLOAT_UINT): Likewise.
835 (BT_FN_FLOAT_UINT_UINT): Likewise.
836 (BT_FN_ULONG_UINT_UINT): Likewise.
837 (BT_FN_ULONG_UINT_PTR): Likewise.
838 (BT_FN_ULONG_ULONG_ULONG): Likewise.
839 (BT_FN_UINT_UINT_UINT): Likewise.
840 (BT_FN_VOID_UINT_PTR): Likewise.
841 (BT_FN_UINT_UINT_PTR: Likewise.
842 (BT_FN_UINT32_UINT64_PTR): Likewise.
843 (BT_FN_INT_INT_UINT_UINT): Likewise.
844 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
845 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
846 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
847 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
848 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
849 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
850 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
851 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
852 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
853 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
854 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
855 * doc/frontends.texi: List BRIG FE.
856 * doc/install.texi (Testing): Add BRIG tesring requirements.
857 * doc/invoke.texi (Overall Options): Mention BRIG.
858 * doc/standards.texi (Standards): Doucment BRIG HSA version.
860 2017-01-24 Richard Biener <rguenther@suse.de>
863 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
865 2017-01-24 Martin Jambor <mjambor@suse.cz>
868 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
869 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
872 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
875 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
876 casts from signed to unsigned really don't have a range.
878 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
880 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
881 GMP_RNDx for compatiblity.
883 2017-01-24 Martin Liska <mliska@suse.cz>
886 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
887 that would prevent us to call alloca with -1 as argument.
889 2017-01-24 Jakub Jelinek <jakub@redhat.com>
891 * dwarf2out.c (output_compilation_unit_header, output_file_names):
892 Avoid -Wformat-security warning.
894 2017-01-23 Andrew Pinski <apinski@cavium.com>
896 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
899 2017-01-23 Martin Sebor <msebor@redhat.com>
902 * gimple-ssa-sprintf.c (warn_level): New global.
903 (format_integer): Use it here and throughout the rest of the file.
904 Use the same switch to compute sign as base.
905 (maybe_warn): New function.
906 (format_directive): Factor out warnings into maybe_warn.
907 Add debugging output. Use warn_level.
908 (add_bytes): Use warn_level.
909 (pass_sprintf_length::compute_format_length): Add debugging output.
910 (try_substitute_return_value): Same.
911 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
914 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
915 (struct fmtresult, format_integer, format_floating): Adjust.
916 (fmtresult::fmtresult): Set max correctly in two argument ctor.
917 (get_string_length, format_string,format_directive): Same.
918 (pass_sprintf_length::compute_format_length): Same.
919 (try_substitute_return_value): Simplify slightly.
922 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
923 (fmtresult::operator+=): Outlined.
924 (struct fmtresult): Add ctors.
925 (struct conversion_spec): Rename...
926 (struct directive): ...to this. Add and remove data members.
927 (directive::set_width, directive::set_precision): New functions.
928 (format_percent): Use fmtresult ctor.
929 (get_width_and_precision): Remove.
930 (format_integer): Make naming changes. Avoid computing width and
932 (format_floating): Same. Adjust indentation.
933 (format_character, format_none): New functions.
934 (format_string): Moved character handling to format_character.
935 (format_directive): Remove arguments, change return type.
936 (parse_directive): New function.
937 (pass_sprintf_length::compute_format_length): Move directive
938 parsing to parse_directive.
940 2017-01-23 Jakub Jelinek <jakub@redhat.com>
942 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
943 (assign_assembler_name_if_needed): ... this.
944 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
945 (assign_assembler_name_if_needed): ... this.
946 (free_lang_data_in_cgraph): Adjust callers.
947 * cgraphunit.c (cgraph_node::analyze): Likewise.
948 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
951 2017-01-23 Richard Biener <rguenther@suse.de>
953 PR tree-optimization/79088
954 PR tree-optimization/79188
955 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
956 resetting loop bounds after last path deletion. Reset loop
957 bounds of the target loop, make code match the comments.
958 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
959 Make sure loops need no fixups.
961 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
963 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
964 exponent support with double type for first argument.
965 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
966 type returned by __builtin_vec_extract_sig,
967 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
968 functions from "vector int" to "vector unsigned int" or from
969 "vector long long int" to "vector unsigned long long int".
970 Changed type returned by __builtin_vec_extract_exp,
971 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
972 functions from "vector int" to "vector unsigned int" or from
973 "vector long long int" to "vector unsigned long long int".
974 Changed return type of __builtin_vec_test_data_class,
975 __builtin_vec_test_data_class_sp, and
976 __builtin_vec_test_data_class_dp from "vector int" to
977 "vector bool int" or from "vector long long int" to "vector bool
978 long long int" and changed second argument type from "unsigned
979 int" to "int". Added new overloaded function forms "vector float
980 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
981 "vector float __builtin_vec_insert_exp_sp (vector float, vector
982 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
983 double, vector unsigned long long int)" and "vector double
984 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
985 long int)". Changed return type of
986 __builtin_scalar_test_data_class and
987 __builtin_scalar_test_data_class_sp and
988 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
989 int" and changed second argument from "unsigned int" to "int".
990 Changed type returned by __builtin_scalar_test_neg,
991 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
992 from "int" to "bool int". Added new overloaded function form
993 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
994 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
995 exponent double-precision with floating point first argument.
996 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
997 documentation of scalar_test_data_class, scalar_test_neg,
998 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
999 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1000 vec_test_data_class built-in functions to reflect refinements in
1001 their type signatures.
1003 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1005 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1007 (aarch64_elf_asm_destructor): Likewise.
1009 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1011 PR rtl-optimization/78634
1012 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1013 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1014 * ifcvt.c (noce_try_cmove): Add missing cost check.
1016 PR rtl-optimization/71724
1017 * combine.c (if_then_else_cond): Look for situations where it is
1018 beneficial to undo the work of one of the recursive calls.
1020 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1022 PR tree-optimization/70754
1023 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1024 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1025 combined stmt before it if not NULL.
1026 (combine_chains): Process refs reversely and compute dominance point
1029 2017-01-23 Martin Liska <mliska@suse.cz>
1031 PR tree-optimization/79196
1032 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1033 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1035 (strlen_optimize_stmt): Call the renamed function.
1037 2017-01-23 Michael Matz <matz@suse.de>
1039 PR tree-optimization/78384
1040 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
1042 2017-01-23 Richard Biener <rguenther@suse.de>
1044 PR tree-optimization/79186
1045 * tree-vrp.c (register_new_assert_for): Make sure we've seen
1046 both incoming edges before moving an assert.
1048 2017-01-23 Martin Jambor <mjambor@suse.cz>
1050 * ipa-prop.c (load_from_param_1): Removed.
1051 (load_from_unmodified_param): Bits from load_from_param_1 put back
1053 (load_from_param): Removed.
1054 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
1055 with stmt. Reverted back to use of load_from_unmodified_param.
1057 2017-01-23 Martin Jambor <mjambor@suse.cz>
1060 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
1061 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
1062 field a pointer to garbage collected vector, mark lattices and
1063 ipcp_orig_node with GTY((skip)).
1064 (ipa_get_param_count): Adjust to descriptors being a pointer.
1065 (ipa_get_param): Likewise.
1066 (ipa_get_type): Likewise.
1067 (ipa_get_param_move_cost): Likewise.
1068 (ipa_set_param_used): Likewise.
1069 (ipa_get_controlled_uses): Likewise.
1070 (ipa_set_controlled_uses): Likewise.
1071 (ipa_is_param_used): Likewise.
1072 (ipa_node_params_t): Move into garbage collector. New methods insert
1074 (ipa_node_params_sum): Annotate wth GTY(()).
1075 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
1077 (ipa_load_from_parm_agg): Adjust declaration.
1078 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
1079 * ipa-profile.c (ipa_profile): Likewise.
1080 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
1081 (ipa_populate_param_decls): Make descriptors parameter garbage
1083 (ipa_dump_param): Adjust to descriptors being a pointer.
1084 (ipa_alloc_node_params): Likewise.
1085 (ipa_initialize_node_params): Likewise.
1086 (load_from_param_1): Make descriptors parameter garbage collected.
1087 (load_from_unmodified_param): Likewise.
1088 (load_from_param): Likewise.
1089 (ipa_load_from_parm_agg): Likewise.
1090 (ipa_node_params::~ipa_node_params): Removed.
1091 (ipa_free_all_node_params): Remove call to delete operator.
1092 (ipa_node_params_t::insert): New.
1093 (ipa_node_params_t::remove): Likewise.
1094 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
1095 copy known_csts and known_contexts vectors.
1096 (ipa_read_node_info): Adjust to descriptors being a pointer.
1097 (ipcp_modif_dom_walker): Make m_descriptors field garbage
1099 (ipcp_transform_function): Make descriptors variable garbage
1102 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1104 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1105 * config/i386/avx512dqintrin.h: Ditto.
1106 * config/i386/avx512fintrin.h: Ditto.
1107 * gcc/config/i386/i386.c: Handle new builtins.
1108 * config/i386/i386-builtin.def: Add new builtins.
1109 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1110 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1112 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1113 Martin Liska <mliska@suse.cz>
1115 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1116 * asan.c (asan_expand_poison_ifn): Support stores and use
1117 appropriate ASAN report function.
1118 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1119 * internal-fn.def (ASAN_POISON_USE): Declare.
1120 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1121 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1122 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1123 ASAN_POISON calls w/o LHS.
1124 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1125 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1126 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1127 * gimplify.c (asan_poison_variables): Add attribute
1128 use_after_scope_memory to variables that really needs to live
1130 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1131 having the attribute.
1133 2017-01-23 Martin Liska <mliska@suse.cz>
1135 * asan.c (create_asan_shadow_var): New function.
1136 (asan_expand_poison_ifn): Likewise.
1137 * asan.h (asan_expand_poison_ifn): New declaration.
1138 * internal-fn.c (expand_ASAN_POISON): Likewise.
1139 * internal-fn.def (ASAN_POISON): New builtin.
1140 * sanopt.c (pass_sanopt::execute): Expand
1141 asan_expand_poison_ifn.
1142 * tree-inline.c (copy_decl_for_dup_finish): Make function
1144 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1145 * tree-ssa.c (is_asan_mark_p): New function.
1146 (execute_update_addresses_taken): Rewrite local variables
1147 (identified just by use-after-scope as addressable) into SSA.
1149 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1151 * doc/install.texi (Specific): opensource.apple.com uses https
1152 now. Remove trailing slash.
1154 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1156 * README.Portability: Remove note on an Irix compatibility issue.
1158 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1160 * gcov.c (INCLUDE_ALGORITHM): Define.
1161 (INCLUDE_VECTOR): Define.
1162 No longer include <vector> and <algorithm> directly.
1164 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1166 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1168 * doc/invoke.texi (Code Gen Options): Ditto.
1170 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1173 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1175 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1177 rtl-optimization/79125
1178 * cprop.c (local_cprop_pass): Handle cases where we make an
1181 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1185 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1186 read from, for big endian.
1188 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1190 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1191 register pauth builtins for LP64 only.
1193 2017-01-20 Marek Polacek <polacek@redhat.com>
1196 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1199 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1201 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1203 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1204 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1205 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1207 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1210 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1211 in big-endian mode when they are not a single duplicated value.
1213 2017-01-20 Richard Biener <rguenther@suse.de>
1215 * BASE-VER: Bump to 7.0.1.
1217 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1219 * omp-low.c (omplow_simd_context): New struct. Use it...
1220 (lower_rec_simd_input_clauses): ...here and...
1221 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1222 references to idx, lane, max_vf, is_simt.
1224 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1226 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1229 2017-01-20 Martin Jambor <mjambor@suse.cz>
1231 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1232 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1234 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1235 (GTFILES): Rename hsa.c to hsa-common.c.
1236 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1237 * hsa-dump.c: Likewise.
1238 * hsa-gen.c: Likewise.
1239 * hsa-regalloc.c: Likewise.
1240 * ipa-hsa.c: Likewise.
1241 * omp-expand.c: Likewise.
1242 * omp-low.c: Likewise.
1243 * toplev.c: Likewise.
1245 2017-01-20 Marek Polacek <polacek@redhat.com>
1248 * doc/invoke.texi: Document -Wduplicated-branches.
1249 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1250 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1251 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1252 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1253 return 0 only when not OEP_LEXICOGRAPHIC.
1254 (fold_build_cleanup_point_expr): Use the expression
1255 location when building CLEANUP_POINT_EXPR.
1256 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1257 * tree.c (add_expr): Handle error_mark_node.
1259 2017-01-20 Martin Liska <mliska@suse.cz>
1262 * tree-profile.c (init_ic_make_global_vars): Do not call
1264 (gimple_init_gcov_profiler): Likewise.
1266 2017-01-20 Martin Liska <mliska@suse.cz>
1269 * cgraph.h (maybe_create_reference): Remove argument and
1271 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1273 * ipa-cp.c (create_specialized_node): Likewise.
1274 * symtab.c (symtab_node::maybe_create_reference): Handle
1275 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1277 2017-01-20 Martin Liska <mliska@suse.cz>
1279 * read-rtl-function.c (function_reader::create_function): Use
1280 build_decl instread of build_decl_stat.
1282 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1284 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1285 * config/i386/avx512dqintrin.h: Ditto.
1286 * config/i386/avx512fintrin.h: Ditto.
1287 * config/i386/i386-builtin-types.def: Add new types.
1288 * gcc/config/i386/i386.c: Handle new types.
1289 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1290 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1291 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1292 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1293 (__builtin_ia32_kshiftridi): New.
1294 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1296 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1300 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1301 define to rs6000_init_stack_protect_guard.
1302 (rs6000_init_stack_protect_guard): New function.
1304 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1305 Yunqiang Su <yunqiang.su@imgtec.com>
1307 * config.gcc (supported_defaults): Add madd4.
1308 (with_madd4): Add validation.
1309 (all_defaults): Add madd4.
1310 * config/mips/mips.opt (mmadd4): New option.
1311 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1313 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1315 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1316 (ISA_HAS_FUSED_MADD4): Likewise.
1317 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1318 * gcc/doc/install.texi (--with-madd4): Document the new option.
1320 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1322 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1323 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1324 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1325 (aarch64_init_pauth_hint_builtins): New.
1326 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1327 (aarch64_expand_builtin): Expand new builtins.
1329 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1331 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1332 * combine-stack-adj.c (no_unhandled_cfa): Handle
1333 REG_CFA_TOGGLE_RA_MANGLE.
1334 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1335 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1336 info for return address signing.
1337 (aarch64_expand_epilogue): Likewise.
1339 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1341 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1342 * config/aarch64/aarch64-protos.h
1343 (aarch64_return_address_signing_enabled): New declaration.
1344 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1346 (aarch64_expand_prologue): Sign return address before it's pushed onto
1348 (aarch64_expand_epilogue): Authenticate return address fetched from
1350 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1351 (aarch64_attributes): New function attributes for "sign-return-address".
1352 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1353 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1354 ("*do_return"): Generate combined instructions according to key index.
1355 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1356 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1358 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1359 * config/aarch64/aarch64.opt (msign-return-address=): New.
1360 * doc/extend.texi (AArch64 Function Attributes): Documents
1361 "sign-return-address=".
1362 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1364 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1366 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1367 overall option summary.
1369 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1371 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1372 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1373 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1374 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1376 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1378 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1379 -mpower9-minmax by default for -mcpu=power9.
1380 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1381 128-bit floating point.
1383 2017-01-20 Alan Modra <amodra@gmail.com>
1385 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1386 optimizing for size.
1388 2017-01-20 Alan Modra <amodra@gmail.com>
1391 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1392 for strcmp and strncmp from corresponding builtin decl.
1394 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1396 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1397 instead of i386/rtems-64.h.
1398 * config/i386/rtems-64.h: Remove.
1400 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1404 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1406 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1408 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1410 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1411 Change int to HOST_WIDE_INT.
1412 * config/aarch64/aarch64-protos.h
1413 (aarch64_simd_gen_const_vector_dup): Likewise.
1414 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1416 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1418 * langhooks-def.h (lhd_type_for_size): New decl.
1419 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1420 * langhooks.c (lhd_type_for_size): New function, taken from
1423 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1425 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1426 define_bypass for CR latency.
1427 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1428 (power9-alu2): Add define_bypass for CR latency.
1430 (power9-mul): Update insn latency.
1431 (power9-mul-compare): Update insn latency, bypass latency and remove
1434 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1436 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1438 * config/aarch64/aarch64.md
1439 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1440 aarch64_nopcrelative_literal_loads.
1441 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1443 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1445 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1447 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1449 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1452 * config.gcc (supported_defaults): Add lxc1-sxc1.
1453 (with_lxc1_sxc1): Add validation.
1454 (all_defaults): Add lxc1-sxc1.
1455 * config/mips/mips.opt (mlxc1-sxc1): New option.
1456 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1458 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1459 __mips_no_lxc1_sxc1.
1460 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1461 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1462 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1464 2017-01-19 Richard Biener <rguenther@suse.de>
1466 PR tree-optimization/72488
1467 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1468 sure to restore SSA info.
1469 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1471 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1473 PR rtl-optimization/79121
1474 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1475 of the inner type when shifting an extended value.
1477 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1480 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1481 interposable aliases.
1483 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1486 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1487 Use the evmergelohi instruction.
1488 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1489 (mov_sitf_e500_subreg8_2_be): Likewise.
1490 (mov_sitf_e500_subreg12_2_le): Likewise.
1491 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1492 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1493 (mov_sitf_e500_subreg8_2_le): Likewise.
1494 (mov_sitf_e500_subreg12_2_be): Likewise.
1496 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1498 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1499 attribute from vecsimple to vecperm.
1500 (altivec_vbpermq2): Likewise.
1502 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1505 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1507 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1508 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1509 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1510 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1511 case where N arg is SIZE_MAX.
1512 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1513 (cmpstrsi): Add pattern.
1515 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1517 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1518 __builtin_vec_revb builtins.
1519 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1520 built-in functions to support generation of the ISA 3.0 XXBR<x>
1521 vector byte reverse instructions.
1522 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1523 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1524 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1525 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1526 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1527 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1528 (P9V_BUILTIN_VEC_REVB): Likewise.
1529 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1530 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1531 (p9_xxbrq_v16qi): Likewise.
1532 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1533 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1534 (p9_xxbrh_v8hi): Likewise.
1535 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1536 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1537 vec_revb built-in functions.
1539 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1541 PR rtl-optimization/78952
1542 * config/i386/i386.md (any_extract): New code iterator.
1543 (*insvqi_2): Use any_extract for source operand.
1544 (*insvqi_3): Use any_shiftrt for source operand.
1546 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1548 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1550 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1552 2017-01-18 Matthias Klose <doko@ubuntu.com>
1554 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1556 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1558 * config/rs6000/altivec.h (vec_bperm): Change #define.
1559 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1560 (altivec_vbpermq2): New define_insn.
1561 (altivec_vbpermd): Likewise.
1562 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1564 (VBPERMD): Likewise.
1565 (VBPERM): New polymorphic function interface.
1566 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1567 Add entries for P9V_BUILTIN_VEC_VBPERM.
1568 * doc/extend.texi: Add interfaces for vec_bperm.
1570 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1572 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1573 first letter of error messages.
1574 (s390_resolve_overloaded_builtin): Likewise.
1575 * config/s390/s390.c (s390_expand_builtin): Likewise.
1576 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1577 (s390_valid_target_attribute_inner_p): Likewise.
1578 * config/s390/s390.md ("tabort"): Likewise.
1580 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1582 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1583 (ISA_AVOID_DIV_HILO): New macro.
1584 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1585 (ISA_HAS_DDIV): Likewise.
1587 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1589 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1591 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1593 * doc/invoke.texi (fabi-version): Spelling fix.
1595 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1598 * doc/invoke.texi (fabi-version): Mention mangling fix for
1601 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1604 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1606 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1609 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1610 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1612 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1614 (stack_protect_test): Ditto.
1615 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1616 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1618 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1619 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1620 -mstack-protector-guard-offset=.
1621 (RS/6000 and PowerPC Options): Ditto.
1623 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1625 * config/i386/i386.h (MASK_CLASS_P): New define.
1626 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1627 there are no registers from different register sets also when
1628 mask registers are used. Update function comment.
1629 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1630 to (*k/*r) and (*k/*km) alternatives.
1632 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1634 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1635 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1636 (EH_RETURN_HANDLER_RTX): New define.
1637 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1638 Force frame pointer in EH return functions.
1639 (aarch64_expand_epilogue): Add barrier for eh_return.
1640 (aarch64_final_eh_return_addr): Remove.
1641 (aarch64_eh_return_handler_rtx): New function.
1642 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1644 (aarch64_eh_return_handler_rtx): New prototype.
1646 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1648 * config/rs6000/altivec.h (vec_rlmi): New #define.
1649 (vec_vrlnm): Likewise.
1650 (vec_rlnm): Likewise.
1651 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1652 (UNSPEC_VRLNM): Likewise.
1653 (VIlong): New mode iterator.
1654 (altivec_vrl<VI_char>mi): New define_insn.
1655 (altivec_vrl<VI_char>nm): Likewise.
1656 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1659 (RLNM): New polymorphic function entry.
1660 (VRLWMI): New monomorphic function entry.
1662 (RLMI): New polymorphic function entry.
1663 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1664 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1665 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1668 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1671 * dwarf2out.c (field_byte_offset): Restore the
1672 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1673 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1674 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1677 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1680 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1682 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1685 * dwarf2out.c (add_data_member_location_attribute): For constant
1686 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1687 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1688 DW_AT_byte_size attributes.
1690 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1692 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1693 after forcing to constant memory when the code model is medium.
1695 2017-01-17 Julia Koval <julia.koval@intel.com>
1698 * config/i386/avx512fintrin.h
1699 (_mm512_i32gather_ps): Change __addr type to void const*.
1700 (_mm512_mask_i32gather_ps): Ditto.
1701 (_mm512_i32gather_pd): Ditto.
1702 (_mm512_mask_i32gather_pd): Ditto.
1703 (_mm512_i64gather_ps): Ditto.
1704 (_mm512_mask_i64gather_ps): Ditto.
1705 (_mm512_i64gather_pd): Ditto.
1706 (_mm512_mask_i64gather_pd): Ditto.
1707 (_mm512_i32gather_epi32): Ditto.
1708 (_mm512_mask_i32gather_epi32): Ditto.
1709 (_mm512_i32gather_epi64): Ditto.
1710 (_mm512_mask_i32gather_epi64): Ditto.
1711 (_mm512_i64gather_epi32): Ditto.
1712 (_mm512_mask_i64gather_epi32): Ditto.
1713 (_mm512_i64gather_epi64): Ditto.
1714 (_mm512_mask_i64gather_epi64): Ditto.
1715 (_mm512_i32scatter_ps): Change __addr type to void*.
1716 (_mm512_mask_i32scatter_ps): Ditto.
1717 (_mm512_i32scatter_pd): Ditto.
1718 (_mm512_mask_i32scatter_pd): Ditto.
1719 (_mm512_i64scatter_ps): Ditto.
1720 (_mm512_mask_i64scatter_ps): Ditto.
1721 (_mm512_i64scatter_pd): Ditto.
1722 (_mm512_mask_i64scatter_pd): Ditto.
1723 (_mm512_i32scatter_epi32): Ditto.
1724 (_mm512_mask_i32scatter_epi32): Ditto.
1725 (_mm512_i32scatter_epi64): Ditto.
1726 (_mm512_mask_i32scatter_epi64): Ditto.
1727 (_mm512_i64scatter_epi32): Ditto.
1728 (_mm512_mask_i64scatter_epi32): Ditto.
1729 (_mm512_i64scatter_epi64): Ditto.
1730 (_mm512_mask_i64scatter_epi64): Ditto.
1731 * config/i386/avx512pfintrin.h
1732 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1733 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1734 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1735 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1736 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1737 (_mm512_prefetch_i32scatter_ps): Ditto.
1738 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1739 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1740 (_mm512_prefetch_i64scatter_pd): Ditto.
1741 (_mm512_prefetch_i64scatter_ps): Ditto.
1742 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1743 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1744 * config/i386/avx512vlintrin.h
1745 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1746 (_mm_mmask_i32gather_ps): Ditto.
1747 (_mm256_mmask_i32gather_pd): Ditto.
1748 (_mm_mmask_i32gather_pd): Ditto.
1749 (_mm256_mmask_i64gather_ps): Ditto.
1750 (_mm_mmask_i64gather_ps): Ditto.
1751 (_mm256_mmask_i64gather_pd): Ditto.
1752 (_mm_mmask_i64gather_pd): Ditto.
1753 (_mm256_mmask_i32gather_epi32): Ditto.
1754 (_mm_mmask_i32gather_epi32): Ditto.
1755 (_mm256_mmask_i32gather_epi64): Ditto.
1756 (_mm_mmask_i32gather_epi64): Ditto.
1757 (_mm256_mmask_i64gather_epi32): Ditto.
1758 (_mm_mmask_i64gather_epi32): Ditto.
1759 (_mm256_mmask_i64gather_epi64): Ditto.
1760 (_mm_mmask_i64gather_epi64): Ditto.
1761 (_mm256_i32scatter_ps): Change __addr type to void*.
1762 (_mm256_mask_i32scatter_ps): Ditto.
1763 (_mm_i32scatter_ps): Ditto.
1764 (_mm_mask_i32scatter_ps): Ditto.
1765 (_mm256_i32scatter_pd): Ditto.
1766 (_mm256_mask_i32scatter_pd): Ditto.
1767 (_mm_i32scatter_pd): Ditto.
1768 (_mm_mask_i32scatter_pd): Ditto.
1769 (_mm256_i64scatter_ps): Ditto.
1770 (_mm256_mask_i64scatter_ps): Ditto.
1771 (_mm_i64scatter_ps): Ditto.
1772 (_mm_mask_i64scatter_ps): Ditto.
1773 (_mm256_i64scatter_pd): Ditto.
1774 (_mm256_mask_i64scatter_pd): Ditto.
1775 (_mm_i64scatter_pd): Ditto.
1776 (_mm_mask_i64scatter_pd): Ditto.
1777 (_mm256_i32scatter_epi32): Ditto.
1778 (_mm256_mask_i32scatter_epi32): Ditto.
1779 (_mm_i32scatter_epi32): Ditto.
1780 (_mm_mask_i32scatter_epi32): Ditto.
1781 (_mm256_i32scatter_epi64): Ditto.
1782 (_mm256_mask_i32scatter_epi64): Ditto.
1783 (_mm_i32scatter_epi64): Ditto.
1784 (_mm_mask_i32scatter_epi64): Ditto.
1785 (_mm256_i64scatter_epi32): Ditto.
1786 (_mm256_mask_i64scatter_epi32): Ditto.
1787 (_mm_i64scatter_epi32): Ditto.
1788 (_mm_mask_i64scatter_epi32): Ditto.
1789 (_mm256_i64scatter_epi64): Ditto.
1790 (_mm256_mask_i64scatter_epi64): Ditto.
1791 (_mm_i64scatter_epi64): Ditto.
1792 (_mm_mask_i64scatter_epi64): Ditto.
1793 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1794 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1795 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1796 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1797 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1798 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1799 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1800 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1801 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1802 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1803 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1804 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1805 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1806 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1807 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1808 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1809 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1810 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1811 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1812 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1813 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1814 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1815 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1816 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1817 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1818 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1819 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1820 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1821 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1822 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1823 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1824 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1825 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1826 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1827 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1828 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1829 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1830 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1831 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1832 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1833 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1834 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1835 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1836 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1837 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1838 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1839 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1840 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1841 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1842 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1843 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1844 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1845 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1846 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1847 definitions accordingly.
1849 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1850 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1853 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1856 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1859 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1860 hard regs for inner regno.
1862 2017-01-17 Martin Liska <mliska@suse.cz>
1865 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1866 assumption and add comment.
1868 2017-01-17 Nathan Sidwell <nathan@acm.org>
1870 * ipa-visibility.c (localize_node): New function, broken out of ...
1871 (function_and_variable_visibility): ... here. Call it.
1873 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1876 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1877 correctly set frequency of oudgoing edge.
1878 (duplicate_thread_path): Fix profile updating.
1880 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1883 * configure.ac: Add GCC_BASE_VER.
1884 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1885 version from BASE-VER file.
1886 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1887 (gcc.o): Depend on $(BASEVER).
1888 * common.opt (dumpfullversion): New option.
1889 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1890 * doc/invoke.texi: Document -dumpfullversion.
1891 * doc/install.texi: Document --with-gcc-major-version-only.
1892 * configure: Regenerated.
1894 2017-01-17 Richard Biener <rguenther@suse.de>
1896 PR tree-optimization/71433
1897 * tree-vrp.c (register_new_assert_for): Merge same asserts
1898 on all incoming edges.
1899 (process_assert_insertions_for): Handle insertions at the
1902 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1904 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1905 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1907 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1910 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1913 2017-01-17 Alan Modra <amodra@gmail.com>
1916 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1917 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1918 symbolic stack limit when pic.
1920 2017-01-16 Martin Sebor <msebor@redhat.com>
1922 PR tree-optimization/78608
1923 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1925 2017-01-16 Jeff Law <law@redhat.com>
1928 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1929 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1930 for several include directories that may be relative to sysroot.
1931 * config/i386/x-mingw32 (gplus_includedir): Define.
1932 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1933 (native_system_includedir): Likewise.
1934 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1935 override if TARGET_SYSTEM_ROOT is defined.
1936 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1938 PR tree-optimization/79090
1939 PR tree-optimization/33562
1940 PR tree-optimization/61912
1941 PR tree-optimization/77485
1942 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1943 and computed trims into the dump file.
1945 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1947 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1949 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1952 * gimplify.c (gimplify_init_constructor): If want_value and
1953 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1957 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1958 sequence. Formatting fixes.
1959 (doloop_optimize): Formatting fixes.
1962 * gcc.c (debug_level_greater_than_spec_func): New function.
1963 (static_spec_functions): Add debug-level-gt spec function.
1964 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1966 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1967 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1968 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1969 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1970 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1971 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1973 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1975 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1976 QImode fixups to general and mask registers only.
1978 2017-01-16 Carl Love <cel@us.ibm.com>
1980 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1981 for built-in functions
1982 vector signed char vec_nabs (vector signed char)
1983 vector signed short vec_nabs (vector signed short)
1984 vector signed int vec_nabs (vector signed int)
1985 vector signed long long vec_nabs (vector signed long long)
1986 vector float vec_nabs (vector float)
1987 vector double vec_nabs (vector double)
1988 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1990 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1991 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1992 * doc/extend.texi: Update the documentation file for the new built-in
1995 2017-01-16 Martin Sebor <msebor@redhat.com>
1997 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2000 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2002 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2003 UNSPEC_VSX__XXSPLTD to require special splat handling.
2005 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2008 * system.h: Poison strndup.
2010 2017-01-16 Alan Modra <amodra@gmail.com>
2013 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2016 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2018 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2020 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2022 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2023 call recog here. Assert that INSN_CODE (insn) is non-negative.
2025 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2028 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2030 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2031 in the currently scheduled RTL fragment.
2033 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2035 PR rtl-optimization/78751
2036 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2039 2017-01-14 Jeff Law <law@redhat.com>
2041 PR tree-optimization/79090
2042 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
2043 variable length stores.
2044 (compute_trims): Delete dead assignment to *trim_tail.
2045 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
2048 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
2050 PR rtl-optimization/78626
2051 PR rtl-optimization/78727
2052 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
2053 of a block, and split such blocks after everything else is finished.
2055 2017-01-14 Alan Modra <amodra@gmail.com>
2058 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
2059 target legitimate_combined_insn.
2060 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
2061 (rs6000_legitimate_combined_insn): New function.
2062 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
2064 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
2065 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
2066 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
2068 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
2070 * doc/frontends.texi (G++ and GCC): Remove references to Java.
2072 2017-01-13 Jeff Law <law@redhat.com>
2074 PR tree-optimization/33562
2075 PR tree-optimization/61912
2076 PR tree-optimization/77485
2077 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
2079 (delete_dead_assignment): Likewise.
2080 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
2081 statement to delete_dead_call and delete_dead_assignment.
2083 2017-01-13 David Malcolm <dmalcolm@redhat.com>
2086 * substring-locations.c (format_warning_va): Strengthen case 1 so
2087 that both endpoints of the substring must be within the format
2088 range for just the substring to be printed.
2090 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
2092 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
2093 * config/i386/i386.c (ix86_target_string): Add missing options
2094 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
2095 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
2096 flags_other and ix86_target_other to flags2_other. Display unknown
2098 (ix86_valid_target_attribute_inner_p): Add missing options and
2099 reorder options by implied ISAs, as in ix86_target_string.
2101 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2103 * hash-table.h (hash_table::too_empty_p): New function.
2104 (hash_table::expand): Use it.
2105 (hash_table::traverse): Likewise.
2106 (hash_table::empty_slot): Use sizeof (value_type) instead of
2107 sizeof (PTR) to convert bytes to elements. Shrink the table
2108 if the current size is excessive for the current number of
2111 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2113 * ira-costs.c (record_reg_classes): Break from the inner loop
2114 early once alt_fail is known to be true. Update outer loop
2115 handling accordingly.
2117 2017-01-13 Jeff Law <law@redhat.com>
2119 * tree-ssa-dse.c (decrement_count): New function.
2120 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2121 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2122 when we know the partially dead statement is a mem* function.
2124 PR tree-optimization/61912
2125 PR tree-optimization/77485
2126 * tree-ssa-dse.c: Include expr.h.
2127 (maybe_trim_constructor_store): New function.
2128 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2130 PR tree-optimization/33562
2131 PR tree-optimization/61912
2132 PR tree-optimization/77485
2133 * doc/invoke.texi: Document new dse-max-object-size param.
2134 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2135 * tree-ssa-dse.c: Include params.h.
2136 (dse_store_status): New enum.
2137 (initialize_ao_ref_for_dse): New, partially extracted from
2139 (valid_ao_ref_for_dse, normalize_ref): New.
2140 (setup_live_bytes_from_ref, compute_trims): Likewise.
2141 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2142 (maybe_trim_partially_dead_store): Likewise.
2143 (maybe_trim_complex_store): Likewise.
2144 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2145 Track what bytes live from the original store. Return tri-state
2146 for dead, partially dead or live.
2147 (dse_dom_walker): Add constructor, destructor and new private members.
2148 (delete_dead_call, delete_dead_assignment): New extracted from
2150 (dse_optimize_stmt): Make a member of dse_dom_walker.
2151 Use initialize_ao_ref_for_dse.
2153 PR tree-optimization/33562
2154 PR tree-optimization/61912
2155 PR tree-optimization/77485
2156 * sbitmap.h (bitmap_count_bits): Prototype.
2157 (bitmap_clear_range, bitmap_set_range): Likewise.
2158 * sbitmap.c (bitmap_clear_range): New function.
2159 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2161 2017-01-13 Martin Liska <mliska@suse.cz>
2164 * function.c (set_cfun): Add new argument force.
2165 * function.h (set_cfun): Likewise.
2166 * ipa-inline-transform.c (inline_call): Use the function when
2167 strict alising from is dropped for function we inline to.
2169 2017-01-13 Richard Biener <rguenther@suse.de>
2171 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2172 for dumping GIMPLE INTEGER_CSTs.
2174 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2176 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2177 to 201112L since C++17.
2179 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2182 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2183 if -fsanitize=kernel-address is present.
2185 2017-01-13 Richard Biener <rguenther@suse.de>
2187 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2188 as _Literal ( type ) number in case usual suffixes do not
2189 preserve all information.
2191 2017-01-13 Richard Biener <rguenther@suse.de>
2193 PR tree-optimization/77283
2194 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2195 and ssa-iterators.h.
2196 (is_feasible_trace): Implement a cost model based on joiner
2199 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2202 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2203 char or short to __float128/_Float128 directly.
2205 2017-01-12 Martin Sebor <msebor@redhat.com>
2207 to -Wformat-overflow.
2208 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2209 (min_bytes_remaining): Same.
2210 (get_string_length): Same.
2211 (format_string): Same.
2212 (format_directive): Same.
2214 (pass_sprintf_length::handle_gimple_call): Same.
2216 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2218 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2219 info.nowrite calls with no lhs that can't throw. Return bool
2220 whether gsi_remove has been called or not.
2221 (pass_sprintf_length::handle_gimple_call): Return bool whether
2222 try_substitute_return_value called gsi_remove. Formatting fix.
2223 (pass_sprintf_length::execute): Don't use gsi_remove if
2224 handle_gimple_call returned true.
2227 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2228 be removed due to side-effects, don't remove following barrier nor
2229 turn the successor edge into fallthru edge.
2231 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2234 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2235 element-reversing loads and stores as not swappable.
2237 2017-01-12 Nathan Sidwell <nathan@acm.org>
2238 Nicolai Stange <nicstange@gmail.com>
2240 * combine.c (try_combine): Don't ignore result of overlap checking
2241 loop. Combine overlap & asm check into single loop.
2243 2017-01-12 Richard Biener <rguenther@suse.de>
2245 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2246 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2248 2017-01-12 Richard Biener <rguenther@suse.de>
2250 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2251 and TS_TARGET_OPTION directly derive from TS_BASE.
2252 * tree-core.h (tree_optimization_option): Derive from tree_base.
2253 (tree_target_option): Likewise.
2255 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2257 * config/i386/i386.c (memory_address_length): Increase len
2258 only when rip_relative_addr_p returns false.
2260 2017-01-11 Julia Koval <julia.koval@intel.com>
2262 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2263 (OPTION_MASK_ISA_SGX_SET): New.
2264 (ix86_handle_option): Handle OPT_msgx.
2265 * config.gcc: Added sgxintrin.h.
2266 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2267 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2268 * config/i386/i386.c (ix86_target_string): Add -msgx.
2270 (ix86_option_override_internal): Handle new options.
2271 (ix86_valid_target_attribute_inner_p): Add sgx.
2272 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2273 * config/i386/i386.opt: Add msgx.
2274 * config/i386/sgxintrin.h: New file.
2275 * config/i386/x86intrin.h: Add sgxintrin.h.
2277 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2280 * fold-const.c (maybe_nonzero_address): Return 1 for function
2282 (tree_single_nonzero_warnv_p): Don't handle function local objects
2286 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2289 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2292 * opts.c: Include "spellcheck.h"
2293 (struct string_fragment): New struct.
2294 (struct edit_distance_traits<const string_fragment &>): New
2296 (get_closest_sanitizer_option): New function.
2297 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2299 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2301 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2303 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2304 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2305 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2306 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2307 for initial die_offset if dwarf_split_debug_info.
2308 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2309 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2311 (output_skeleton_debug_sections): Formatting fix. Use
2312 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2313 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2315 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2317 * config/arm/cortex-a53.md: Add bypasses for
2319 (cortex_a53_r2f): Only use for transfers.
2320 (cortex_a53_f2r): Likewise.
2321 (cortex_a53_r2f_cvt): Add reservation for conversions.
2322 (cortex_a53_f2r_cvt): Likewise.
2324 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2326 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2327 to all inlined functions, change static to extern.
2329 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2332 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2334 (arm_assemble_integer): Likewise.
2336 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2338 * config.gcc: Use new awk script to check CPU, FPU and architecture
2339 parameters for --with-... options.
2340 * config/arm/parsecpu.awk: New file
2341 * config/arm/arm-cpus.in: New file.
2342 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2344 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2346 * config/arm/t-arm: Update dependency rules.
2347 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2348 of processing .def files.
2349 * config/arm/genopt.sh: Deleted.
2350 * config/arm/gentune.sh: Deleted.
2351 * config/arm/arm-cores.def: Deleted.
2352 * config/arm/arm-arches.def: Deleted.
2353 * config/arm/arm-fpus.def: Deleted.
2354 * config/arm/arm-tune.md: Regenerated.
2355 * config/arm/arm-tables.opt: Regenerated.
2356 * config/arm/arm-cpu.h: New generated file.
2357 * config/arm/arm-cpu-data.h: New generated file.
2358 * config/arm/arm-cpu-cdata.h: New generated file.
2360 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2363 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2365 (input_varpool_node): Unpack dynamically_initialized bit.
2367 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2369 PR rtl-optimization/79032
2370 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2371 the alignment of the adjusted memory reference against that of MODE,
2372 instead of the alignment of the original memory reference.
2374 2017-01-11 Martin Jambor <mjambor@suse.cz>
2376 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2378 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2379 decorated functions.
2381 2017-01-11 Richard Biener <rguenther@suse.de>
2383 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2384 set range/nonnull info for PHI results. Do not set it on
2385 stmts marked for removal.
2387 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2389 * expr.c (store_field): In the bitfield case, fetch the return value
2390 from the registers before applying a single big-endian adjustment.
2391 Always do a final load for a BLKmode value not larger than a word.
2393 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2396 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2397 that we correctly handle column numbers greater than
2398 LINE_MAP_MAX_COLUMN_NUMBER.
2400 2017-01-10 Martin Sebor <msebor@redhat.com>
2403 * gimple-ssa-sprintf.c (get_destination_size): Call
2404 {init,fini}object_sizes.
2405 * tree-object-size.c (addr_object_size): Adjust.
2406 (pass_through_call): Adjust.
2407 (pass_object_sizes::execute): Adjust.
2408 * tree-object-size.h (fini_object_sizes): Declare.
2410 2017-01-10 Martin Sebor <msebor@redhat.com>
2412 PR tree-optimization/78775
2413 * builtins.c (get_size_range): Move...
2414 * calls.c: ...to here.
2415 (alloc_max_size): Accept zero argument.
2416 (operand_signed_p): Remove.
2417 (maybe_warn_alloc_args_overflow): Call get_size_range.
2418 * calls.h (get_size_range): Declare.
2420 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2422 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2423 from TI's devices.csv file as of September 2016.
2424 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2426 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2428 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2429 * doc/invoke.texi: Likewise.
2430 * doc/md.texi: Likewise.
2431 * doc/objc.texi: Likewise.
2433 2017-01-10 Joshua Conner <joshconner@google.com>
2435 * config/arm/fuchsia-elf.h: New file.
2436 * config/fuchsia.h: New file.
2437 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2438 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2440 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2442 2016-01-10 Richard Biener <rguenther@suse.de>
2444 PR tree-optimization/79034
2445 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2446 Propagate out degenerate PHIs in the joiner.
2448 2017-01-10 Martin Liska <mliska@suse.cz>
2450 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2451 (sort_congruence_classes_by_decl_uid): Likewise.
2452 (sort_congruence_class_groups_by_decl_uid): Likewise.
2453 (sem_item_optimizer::merge_classes): Sort class, groups in these
2454 classes and members in the groups by DECL_UID of declarations.
2455 This would make merge operations stable.
2457 2017-01-10 Martin Liska <mliska@suse.cz>
2459 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2460 usage of m_classes_vec.
2461 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2462 (sem_item_optimizer::get_group_by_hash): Likewise.
2463 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2464 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2465 (sem_item_optimizer::verify_classes): Likewise.
2466 (sem_item_optimizer::process_cong_reduction): Likewise.
2467 (sem_item_optimizer::dump_cong_classes): Likewise.
2468 (sem_item_optimizer::merge_classes): Likewise.
2469 * ipa-icf.h (congruence_class_hash): Rename from
2470 congruence_class_group_hash. Remove declaration of m_classes_vec.
2472 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2474 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2475 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2476 * config.gcc: Add avx512vpopcntdqintrin.h.
2477 * config/i386/avx512vpopcntdqintrin.h: New.
2478 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2479 * config/i386/i386-builtin-types.def: Add new types.
2480 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2481 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2482 __builtin_ia32_vpopcountq_v8di_mask): New.
2483 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2484 __AVX512VPOPCNTDQ__.
2485 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2486 (PTA_AVX512VPOPCNTDQ): Define.
2487 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2488 TARGET_AVX512VPOPCNTDQ_P): Define.
2489 * config/i386/i386.opt: Add mavx512vpopcntdq.
2490 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2491 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2493 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2496 * predict.def (PRED_CALL): Set to 67.
2498 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2500 * expr.c (store_field): In the bitfield case, if the value comes from
2501 a function call and is of an aggregate type returned in registers, do
2502 not modify the field mode; extract the value in all cases if the mode
2503 is BLKmode and the size is not larger than a word.
2505 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2508 * config/i386/cpuid.h: Fix undefined behavior.
2510 2017-01-04 Jeff Law <law@redhat.com>
2512 PR tree-optimization/79007
2513 PR tree-optimization/67955
2514 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2515 conservative for pt.null when flag_non_call_exceptions is on.
2517 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2519 PR translation/79019
2520 PR translation/79020
2521 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2522 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2523 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2525 * config/avr/avr.opt (maccumulate-args): Likewise.
2526 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2527 * common.opt (freport-bug): Likewise.
2528 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2529 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2530 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2531 translatable string.
2532 * config/i386/i386.c (function_value_32): Likewise.
2533 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2534 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2536 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2537 * common/config/msp430/msp430-common.c (msp430_handle_option):
2539 * symtab.c (symtab_node::verify_base): Likewise.
2540 * opts.c (set_debug_level): Likewise.
2541 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2542 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2543 missing whitespace to translatable strings.
2544 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2545 * config/sh/superh.h: Likewise.
2546 * config/i386/xopintrin.h: Likewise.
2547 * config/i386/znver1.md: Likewise.
2548 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2549 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2550 * double-int.h (struct double_int): Likewise.
2551 * double-int.c (div_and_round_double): Likewise.
2552 * wide-int.cc: Likewise.
2553 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2554 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2555 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2556 (crossjumps_occurred): ... this.
2557 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2560 PR tree-optimization/78899
2561 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2562 returning bool return struct loop *, NULL for failure and the new
2564 (versionable_outer_loop_p): Don't version outer loop if it has
2565 dont_vectorized bit set.
2566 (tree_if_conversion): When versioning outer loop, ensure
2567 tree_if_conversion is performed also on the inner loop of the
2568 non-vectorizable outer loop copy.
2569 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2570 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2571 prevent vectorization of it.
2572 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2573 the outer loop vectorization of the non-scalar version is attempted
2574 before vectorization of the inner loop in scalar version. If
2575 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2576 vectorization of its inner loop.
2577 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2578 has 2 inner loops, rename also on edges from bb whose single pred
2579 is outer_loop->header. Fix typo in function comment.
2581 2017-01-09 Martin Sebor <msebor@redhat.com>
2584 * asan.c (asan_emit_stack_protection): Increase local buffer size
2585 to avoid snprintf truncation warning.
2587 2017-01-09 Andrew Pinski <apinski@cavium.com>
2589 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2590 to reference thunderx2t99 for the tuning structure
2591 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2593 (thunderx2t99_extra_costs): This.
2594 * config/aarch64/aarch64-tune.md: Regenerate.
2595 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2596 (vulcan_addrcost_table): This.
2597 (vulcan_regmove_cost): Rename to ...
2598 (thunderx2t99_regmove_cost): This.
2599 (vulcan_vector_cost): Rename to ...
2600 (thunderx2t99_vector_cost): this.
2601 (vulcan_branch_cost): Rename to ...
2602 (thunderx2t99_branch_cost): This.
2603 (vulcan_tunings): Rename to ...
2604 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2605 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2607 2017-01-09 Martin Jambor <mjambor@suse.cz>
2611 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2612 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2613 (propagate_vr_accross_jump_function): Use the above function for all
2614 value range computations for pass-through jump functions and type
2615 converasion from explicit value range values.
2616 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2617 parameters from TYPE_ARG_TYPES.
2618 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2619 (ipa_write_node_info): Stream type of the actual argument.
2620 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2622 2017-01-09 Martin Liska <mliska@suse.cz>
2625 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2626 (lookup_compiler): Do not show error message with have_E.
2628 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2630 PR tree-optimization/78938
2631 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2632 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2633 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2634 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2637 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2639 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2642 2017-01-09 Richard Biener <rguenther@suse.de>
2644 PR tree-optimization/78997
2645 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2646 name condition properly.
2648 2017-01-09 Richard Biener <rguenther@suse.de>
2651 * dwarf2out.c (is_cxx): New overload with context.
2652 (is_naming_typedef_decl): Use it.
2654 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2656 * invoke.texi (Option Summary): Correct spacing in option lists
2657 and add line breaks to fix over-long lines.
2659 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2663 * extend.texi (Common Variable Attributes): Add xref to GCC
2664 Internals manual to explain mode attribute keywords.
2666 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2669 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2670 and Preprocessor Options.
2671 (Options for Linking): Document -pthread here....
2672 (RS/6000 and PowerPC Options): ...not here.
2673 (Solaris 2 Options): ...or here.
2674 * doc/cppopts.texi: Document -pthread.
2676 2017-01-08 Martin Sebor <msebor@redhat.com>
2679 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2680 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2681 New member functions.
2682 (format_directive): Used them.
2684 (pass_sprintf_length::handle_gimple_call): Same.
2685 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2686 to avoid truncation for any argument.
2687 (extract_affine_mul): Same.
2688 * tree.c (get_file_function_name): Same.
2690 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2693 * predict.def (PRED_INDIR_CALL): Set to 86.
2695 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2697 PR preprocessor/54124
2698 * doc/cppopts.texi: Reformat -d subtable to list the full name
2699 of the options. Add cross-reference to the docs for the general
2700 compiler -d options.
2701 * doc/invoke.texi (Developer Options): Add cross-reference to the
2702 preprocessor-specific -d option documentation.
2704 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2706 PR preprocessor/13498
2707 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2708 redudant material, and reflect new command-line options.
2709 (System Headers): Likewise.
2711 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2713 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2714 -isystem, and -idirafter. Copy-edit.
2715 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2716 default for -ftrack-macro-expansion. Delete obsolete and
2717 badly-formatted implementation details about -fdebug-cpp output.
2718 * doc/cppwarnopts.texi: Copy-edit.
2720 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2723 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2724 that the transition from a max line width >= 1<<10 to narrower
2725 lines works correctly.
2727 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2729 * doc/options.texi (PerFunction): New.
2730 * opt-functions.awk (switch_flags): Map both Optimization and
2731 PerFunction to CL_OPTIMIZATION.
2732 * opth-gen.awk: Test for PerFunction flag along with
2734 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2735 it only when the latter is present. Skip those that don't in
2736 the hash function generator.
2737 * common.opt (fvar-tracking): Mark as PerFunction instead of
2739 (fvar-tracking-assignments): Likewise.
2740 (fvar-tracking-assignments-toggle): Likewise.
2741 (fvar-tracking-uninit): Likewise.
2743 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2745 PR translation/79018
2746 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2749 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2752 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2753 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2754 TARGET_LONG_JUMP_TABLE_OFFSETS.
2755 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2756 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2757 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2758 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2759 * config/m68k/m68k.md (tablejump expander): Likewise.
2760 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2761 TARGET_LONG_JUMP_TABLE_OFFSETS.
2762 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2763 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2765 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2766 David Holsgrove <david.holsgrove@xilinx.com>
2768 * common/config/microblaze/microblaze-common.c
2769 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2770 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2772 * config/microblaze/microblaze.c (microblaze_must_save_register)
2773 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2775 (microblaze_eh_return): New function.
2776 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2777 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2778 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2779 * config/microblaze/microblaze.md (eh_return): New pattern.
2781 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2783 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2784 GCC_DIAGNOSTIC_STRINGIFY): Define.
2786 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2788 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2790 * config/arm/arm.md (<mcrr>): New.
2792 * config/arm/arm.c (arm_arch5te): New.
2793 (arm_option_override): Set arm_arch5te.
2794 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2796 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2797 (arm_mcrr_qualifiers): ... this. New.
2798 (MRRC_QUALIFIERS): Define to...
2799 (arm_mrrc_qualifiers): ... this. New.
2800 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2802 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2803 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2804 (MRRCI, mrrc, MRRC): New.
2805 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2806 VUNSPEC_MRRC2): New.
2808 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2810 * config/arm/arm.md (<mcr>): New.
2812 * config/arm/arm.c (arm_coproc_builtin_available): Add
2813 support for mcr, mrc, mcr2 and mrc2.
2814 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2815 (arm_mcr_qualifiers): ... this. New.
2816 (MRC_QUALIFIERS): Define to ...
2817 (arm_mrc_qualifiers): ... this. New.
2818 (MCR_QUALIFIERS): Define to ...
2819 (arm_mcr_qualifiers): ... this. New.
2820 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2822 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2823 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2824 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2827 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2829 * config/arm/arm.md (*ldc): New.
2833 * config/arm/arm.c (arm_coproc_builtin_available): Add
2834 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2835 (arm_coproc_ldc_stc_legitimate_address): New.
2836 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2837 'qualifier_const_pointer'.
2838 (LDC_QUALIFIERS): Define to...
2839 (arm_ldc_qualifiers): ... this. New.
2840 (STC_QUALIFIERS): Define to...
2841 (arm_stc_qualifiers): ... this. New.
2842 * config/arm/arm-protos.h
2843 (arm_coproc_ldc_stc_legitimate_address): New.
2844 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2845 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2846 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2847 stc2, stcl, stc2l): New.
2848 * config/arm/constraints.md (Uz): New.
2849 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2850 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2851 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2852 VUNSPEC_STC2L): New.
2854 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2856 * config/arm/arm.md (<cdp>): New.
2857 * config/arm/arm.c (neon_const_bounds): Rename this ...
2858 (arm_const_bounds): ... this.
2859 (arm_coproc_builtin_available): New.
2860 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2861 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2862 (CDP_QUALIFIERS): Define to...
2863 (arm_cdp_qualifiers): ... this. New.
2865 (arm_expand_builtin_args): Add case for 6 arguments.
2866 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2867 (arm_const_bounds): ... this.
2868 (arm_coproc_builtin_available): New.
2869 * config/arm/arm_acle.h (__arm_cdp): New.
2871 * config/arm/arm_acle_builtins.def (cdp): New.
2873 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2874 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2876 * config/arm/types.md (coproc): New.
2877 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2878 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2879 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2880 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2882 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2884 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2885 (UBINOP_QUALIFIERS): New.
2887 (acle_builtin_data): New. Change comment.
2888 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2889 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2890 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2891 arm_acle_builtins.def.
2892 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2893 (arm_init_acle_builtins): New.
2894 (CRC32_BUILTIN): Remove.
2895 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2896 crc32cb, crc32ch and crc32cw.
2897 (arm_init_crc32_builtins): Remove.
2898 (arm_init_builtins): Use arm_init_acle_builtins rather
2899 than arm_init_crc32_builtins.
2900 (arm_expand_acle_builtin): New.
2901 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2902 * config/arm/arm_acle_builtins.def: New.
2904 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2906 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2907 (arm_builtin_datum): ... this.
2908 (arm_init_neon_builtin): Rename to ...
2909 (arm_init_builtin): ... this. Add a new parameters PREFIX
2910 and USE_SIG_IN_NAME.
2911 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2912 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2913 'arm_builtin_datum'.
2914 (arm_init_vfp_builtins): Likewise.
2915 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2916 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2917 (arm_expand_neon_args): Rename to ...
2918 (arm_expand_builtin_args): ... this. Rename builtin_arg
2919 enum values and differentiate between ARG_BUILTIN_MEMORY
2920 and ARG_BUILTIN_NEON_MEMORY.
2921 (arm_expand_neon_builtin_1): Rename to ...
2922 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2923 values, arm_expand_builtin_args and add bool parameter NEON.
2924 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2925 (arm_expand_vfp_builtin): Likewise.
2926 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2928 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2931 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2932 * predict.c (tree_estimate_probability_bb): Reverse direction of
2933 polymorphic call predictor.
2935 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2937 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2938 (determine_pass_name_match): ...this new function and...
2939 (should_skip_pass_p): ...this new function.
2941 2017-01-06 Nathan Sidwell <nathan@acm.org>
2943 * ipa-visibility.c (function_and_variable_visibility): Reformat
2944 comments and long lines. Remove extrneous if.
2945 * symtab.c (symtab_node::make_decl_local): Fix code format.
2946 (symtab_node::set_section_for_node): Fix comment typo.
2948 2017-01-06 Martin Liska <mliska@suse.cz>
2951 * lra-constraints.c: Rename invariant to lra_invariant.
2952 * predict.c (set_even_probabilities): Initialize e to NULL.
2954 2017-01-05 Martin Sebor <msebor@redhat.com>
2956 PR tree-optimization/78910
2957 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2958 (format_integer): Correct off-by-one error in the handling
2959 of precision with negative numbers in signed conversions..
2961 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2963 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2965 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2967 PR tree-optimization/71016
2968 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2969 factor_out_conditional_conversion. Formatting fix.
2970 (factor_out_conditional_conversion): Add cond_stmt argument.
2971 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2972 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2975 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2977 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2978 read-rtl-function.o, and selftest-rtl.o.
2979 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2980 (selftest::aarch64_test_loading_full_dump): New function.
2981 (selftest::aarch64_run_selftests): New function.
2982 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2983 selftest::aarch64_run_selftests.
2984 * config/i386/i386.c
2985 (selftest::ix86_test_loading_dump_fragment_1): New function.
2986 (selftest::ix86_test_loading_call_insn): New function.
2987 (selftest::ix86_test_loading_full_dump): New function.
2988 (selftest::ix86_test_loading_unspec): New function.
2989 (selftest::ix86_run_selftests): Call the new functions.
2990 * emit-rtl.c (maybe_set_max_label_num): New function.
2991 * emit-rtl.h (maybe_set_max_label_num): New decl.
2992 * function.c (instantiate_decls): Guard call to
2993 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2994 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2996 * gensupport.c (gen_reader::gen_reader): Pass "false"
2997 for new "compact" param of rtx_reader.
2998 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2999 rather than an empty string for NULL strings.
3000 * read-md.c: Potentially include config.h rather than bconfig.h.
3001 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3002 (have_error): New global, copied from errors.c.
3003 (md_reader::read_name): Rename to...
3004 (md_reader::read_name_1): ...this, adding "out_loc" param,
3005 and converting "missing name or number" to returning false, rather
3007 (md_reader::read_name): Reimplement in terms of read_name_1.
3008 (md_reader::read_name_or_nil): New function.
3009 (md_reader::read_string): Handle "(nil)" by returning NULL.
3010 (md_reader::md_reader): Add new param "compact".
3011 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3012 (md_reader::read_file): New method.
3013 * read-md.h (md_reader::md_reader): Add new param "compact".
3014 (md_reader::read_file): New method.
3015 (md_reader::is_compact): New accessor.
3016 (md_reader::read_name): Convert return type from void to file_location.
3017 (md_reader::read_name_or_nil): New decl.
3018 (md_reader::read_name_1): New decl.
3019 (md_reader::m_compact): New field.
3020 (noop_reader::noop_reader): Pass "false" for new "compact" param
3022 (rtx_reader::rtx_reader): Add new "compact" param.
3023 (rtx_reader::read_rtx_operand): Make virtual and convert return
3024 type from void to rtx.
3025 (rtx_reader::read_until): New decl.
3026 (rtx_reader::handle_any_trailing_information): New virtual function.
3027 (rtx_reader::postprocess): New virtual function.
3028 (rtx_reader::finalize_string): New virtual function.
3029 (rtx_reader::m_in_call_function_usage): New field.
3030 (rtx_reader::m_reuse_rtx_by_id): New field.
3031 * read-rtl-function.c: New file.
3032 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3033 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3034 (selftest::verify_three_block_rtl_cfg): New decl.
3035 * read-rtl-function.h: New file.
3036 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3037 For host, include function.h, memmodel.h, and emit-rtl.h.
3038 (one_time_initialization): New function.
3039 (struct compact_insn_name): New struct.
3040 (compact_insn_names): New array.
3041 (find_code): Handle insn codes in compact dumps.
3042 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
3043 (bind_subst_iter_and_attr): Likewise.
3044 (add_condition_to_string): Likewise.
3045 (add_condition_to_rtx): Likewise.
3046 (apply_attribute_uses): Likewise.
3047 (add_current_iterators): Likewise.
3048 (apply_iterators): Likewise.
3049 (initialize_iterators): Guard usage of apply_subst_iterator with
3050 #ifdef GENERATOR_FILE.
3051 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
3052 (md_reader::read_mapping): Likewise.
3053 (add_define_attr_for_define_subst): Likewise.
3054 (add_define_subst_attr): Likewise.
3055 (read_subst_mapping): Likewise.
3056 (check_code_iterator): Likewise.
3057 (rtx_reader::read_rtx): Likewise. Move one-time initialization
3059 (one_time_initialization): New function.
3060 (rtx_reader::read_until): New method.
3061 (read_flags): New function.
3062 (parse_reg_note_name): New function.
3063 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
3064 Handle reuse_rtx ids.
3065 Wrap iterator lookup within #ifdef GENERATOR_FILE.
3066 Add parsing support for RTL dumps, mirroring the special-cases in
3067 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
3068 values, and calling handle_any_trailing_information.
3069 (rtx_reader::read_rtx_operand): Convert return type from void
3070 to rtx, returning return_rtx. Handle case 'e'. Call
3071 finalize_string on XSTR and XTMPL fields.
3072 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
3073 "(nil)" values were omitted. Call the postprocess vfunc on the
3075 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
3076 class ctor. Initialize m_in_call_function_usage. Call
3077 one_time_initialization.
3078 * rtl-tests.c (selftest::test_uncond_jump): Call
3079 set_new_first_and_last_insn.
3080 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
3081 * selftest-rtl.c: New file.
3082 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
3083 (selftest::get_insn_by_uid): New decl.
3084 * selftest-run-tests.c (selftest::run_tests): Call
3085 read_rtl_function_c_tests.
3086 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
3087 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
3090 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
3092 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
3093 operands in a special way. Assert that pos+len <= mode precision.
3095 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3097 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
3098 3 argument Alias with unlimited for the negative form.
3099 (fno-vect-cost-model): Removed.
3101 2017-01-05 Martin Liska <mliska@suse.cz>
3103 * hsa-gen.c (gen_hsa_divmod): New function.
3104 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3106 2017-01-05 Martin Liska <mliska@suse.cz>
3109 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3112 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3114 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3115 small constant length operands.
3117 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3119 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3120 between loop iterations.
3122 2017-01-05 Martin Liska <mliska@suse.cz>
3125 * gimplify.c (gimplify_decl_expr): Compare to
3126 asan_poisoned_variables instread of checking flags.
3127 (gimplify_target_expr): Likewise.
3128 (gimplify_expr): Likewise.
3129 (gimplify_function_tree): Conditionally initialize
3130 asan_poisoned_variables.
3132 2017-01-04 Jeff Law <law@redhat.com>
3134 PR tree-optimizatin/78812
3135 * rtl.h (contains_mem_rtx_p): Prototype.
3136 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3137 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3138 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3139 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3140 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3142 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3144 * input.c (assert_char_at_range): Default-initialize actual_range.
3146 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3148 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3149 to match the caller.
3151 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3153 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3154 insns after final jump in test to emit dummy move.
3156 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3158 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3159 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3161 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3163 * multiple_target.c (create_dispatcher_calls): Init e_next.
3164 * tree-ssa-loop-split.c (split_loop): Init border.
3165 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3168 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3173 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3174 (altivec_register_operand): Do not return true if the operand
3175 contains a SUBREG mixing SImode and SFmode.
3176 (vsx_register_operand): Likewise.
3177 (vsx_reg_sfsubreg_ok): New predicate.
3178 (vfloat_operand): Do not return true if the operand contains a
3179 SUBREG mixing SImode and SFmode.
3180 (vint_operand): Likewise.
3181 (vlogical_operand): Likewise.
3182 (gpc_reg_operand): Likewise.
3183 (int_reg_operand): Likewise.
3184 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3185 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3186 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3188 (rs6000_emit_move_si_sf_subreg): New helper function.
3189 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3190 fixup SUBREGs involving SImode and SFmode.
3191 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3192 numbers for the new peephole2 optimization.
3193 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3194 the GLIBC math library that do AND/IOR/XOR operations on single
3195 precision floating point.
3196 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3197 target macros to say whether we need to avoid SUBREGs mixing
3199 (TARGET_ALLOW_SF_SUBREG): Likewise.
3200 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3201 (UNSPEC_SI_FROM_SF): Likewise.
3202 (iorxor): Change spacing.
3203 (and_ior_xor): New iterator for AND, IOR, and XOR.
3204 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3205 (movdi_from_sf_zero_ext): Likewise.
3206 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3207 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3208 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3209 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3210 (fms<mode>4): Likewise.
3211 (fnma<mode>4): Likewise.
3212 (fnms<mode>4): Likewise.
3213 (nfma<mode>4): Likewise.
3214 (nfms<mode>4): Likewise.
3216 2017-01-04 Marek Polacek <polacek@redhat.com>
3219 * doc/invoke.texi: Document -Wpointer-compare.
3221 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3223 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3226 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3227 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3230 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3233 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3234 documentation of the powerpc_popcntb_ok attribute.
3235 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3236 code to issue warning messages if a requested CPU configuration is
3237 not supported by the binary (assembler and loader) toolchain.
3238 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3239 made to define a built-in function that has been disabled.
3240 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3241 made to define a built-in function that has been disabled.
3242 (altivec_init_builtins): Add comment explaining why definition
3243 of the DST built-in functions is not preceded by an assertion
3244 check. Add assertions to prevent ICE if attempts are made to
3245 define an altivec predicate or an abs* built-in function that has
3247 (htm_init_builtins): Add comment explaining why definition of the
3248 htm built-in functions is not preceded by an assertion check.
3250 2017-01-04 Jeff Law <law@redhat.com>
3252 PR tree-optimizatin/67955
3253 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3254 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3255 the points-to solution does not include pt_null. Use DECL_PT_UID
3258 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3260 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3261 Use gen_int_mode instead of gen_lopwart for const_int operands.
3263 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3265 PR tree-optimization/71563
3266 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3267 out of range value - has low bits known to be zero.
3269 2017-01-04 Alan Modra <amodra@gmail.com>
3271 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3272 * configure: Regenerate.
3273 * config.in: Regenerate.
3275 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3278 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3279 a substring of the message, but strcmp with the whole message. Ifdef
3280 ENABLE_NLS, translate the message first using dgettext.
3282 2017-01-03 Jeff Law <law@redhat.com>
3284 PR tree-optimizatin/78856
3285 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3286 (mark_threaded_blocks): Remove code to truncate thread paths that
3287 cross multiple loop headers. Instead invalidate the cached loop
3288 iteration information and handle case of a thread path walking
3289 into an irreducible region.
3291 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3294 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3295 assertions. Add support for doing the signbit if the IEEE 128-bit
3296 floating point value is in a GPR.
3297 * config/rs6000/rs6000.md (Fsignbit): Delete.
3298 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3299 Update the length attribute if the value is in a GPR.
3300 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3301 the sign or zero extension instruction, since the value is always 0/1.
3302 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3305 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3306 extracting SImode to a GPR register so that we can generate a
3307 store, limit the vector to be in a traditional Altivec register
3308 for the vextuwrx instruction.
3310 2017-01-03 Ian Lance Taylor <iant@google.com>
3312 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3314 2017-01-03 Martin Sebor <msebor@redhat.com>
3316 PR tree-optimization/78696
3317 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3318 precision. Use MPFR for %f for greater fidelity. Correct handling
3320 (pass_sprintf_length::compute_format_length): Set width and precision
3321 specified by asrerisk to void_node for vararg functions.
3322 (try_substitute_return_value): Adjust dump output.
3324 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3326 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3328 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3330 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3331 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3332 -mlra/-mno-lra was passed to the compiler.
3334 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3336 PR rtl-optimization/65618
3337 * emit-rtl.c (try_split): Move initialization of "before" and
3338 "after" to just before the call to emit_insn_after_setloc.
3340 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3342 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3344 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3346 * dwarf2out.c (gen_enumeration_type_die): When
3347 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3349 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3351 PR tree-optimization/78965
3352 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3353 Change first argument from const call_info & to call_info &. For %n
3354 set info.nowrite to false.
3357 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3358 possibly throwing calls.
3360 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3361 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3362 and fns handling, rather than in a separate case SSA_NAME.
3364 2017-01-02 Jeff Law <law@redhat.com>
3366 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3367 fixes for first_period and second_period variables.
3369 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3372 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3373 (*insvqi_1): New insn pattern.
3374 (*insvqi_1_mem_rex64): Ditto.
3376 (*insvqi_3): Rename from *insvqi.
3378 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3380 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3382 * doc/cfg.texi (Edges): Remove reference to Java.
3383 (Maintaining the CFG): Ditto.
3385 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3388 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3389 transparent aliases.
3391 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3394 * predict.def (PRED_CALL): Update hitrate.
3395 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3396 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3397 into direct/indirect/polymorphic variants.
3399 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3401 Update copyright years.
3403 * gcc.c (process_command): Update copyright notice dates.
3404 * gcov-dump.c (print_version): Ditto.
3405 * gcov.c (print_version): Ditto.
3406 * gcov-tool.c (print_version): Ditto.
3407 * gengtype.c (create_file): Ditto.
3408 * doc/cpp.texi: Bump @copying's copyright year.
3409 * doc/cppinternals.texi: Ditto.
3410 * doc/gcc.texi: Ditto.
3411 * doc/gccint.texi: Ditto.
3412 * doc/gcov.texi: Ditto.
3413 * doc/install.texi: Ditto.
3414 * doc/invoke.texi: Ditto.
3416 Copyright (C) 2017 Free Software Foundation, Inc.
3418 Copying and distribution of this file, with or without modification,
3419 are permitted in any medium without royalty provided the copyright
3420 notice and this notice are preserved.