gcc/ChangeLog

   1 2019-01-15  Tom de Vries  <tdevries@suse.de>
   2
   3         PR target/80547
   4         * config/nvptx/nvptx.c (nvptx_goacc_reduction_init): Handle
   5         lhs == NULL_TREE for gang-level reduction.
   6
   7 2019-01-15  Richard Biener  <rguenther@suse.de>
   8             Prathamesh Kulkarni  <prathamesh.kulkarni@linaro.org>
   9
  10         PR ipa/88788
  11         * ipa-pure-const.c (malloc_candidate_p_1): Add parameter visited and
  12         return true if SSA_NAME is already marked in visited bitmap.
  13         (malloc_candidate_p): Pass visited to malloc_candidate_p_1.
  14
  15 2019-01-15  Jakub Jelinek  <jakub@redhat.com>
  16
  17         PR tree-optimization/88775
  18         * match.pd (cmp (convert1?@2 addr@0) (convert2? addr@1)): Optimize
  19         equal == 0 equality pointer comparisons some more if compared in
  20         integral types and either one points to an automatic var and the
  21         other to a global, or we can prove at least one points to the middle
  22         or both point to start or both point to end.
  23
  24 2019-01-14  Andi Kleen  <ak@linux.intel.com>
  25
  26         * Makefile.in: Lower autofdo sampling rate by 10x.
  27         * Makefile.tpl: Dito.
  28
  29 2019-01-14  Tom Honermann  <tom@honermann.net>
  30
  31         * defaults.h: Define CHAR8_TYPE.
  32
  33 2019-01-14  Martin Sebor  <msebor@redhat.com>
  34
  35         PR target/88638
  36         * doc/extend.texi (Darwin Format Checks): Clarify.
  37
  38 2019-01-14  Richard Biener  <rguenther@suse.de>
  39
  40         * genmatch.c (dt_simplify::gen_1): Change dumping dependent on
  41         whether we are in (simplify ...) or (match ...) context.
  42
  43 2019-01-14  Jakub Jelinek  <jakub@redhat.com>
  44
  45         PR rtl-optimization/88796
  46         * emit-rtl.h (struct rtl_data): Add stack_protect_guard_decl field.
  47         * cfgexpand.c (stack_protect_prologue): Initialize
  48         crtl->stack_protect_guard_decl.
  49         * function.c (stack_protect_epilogue): Use it instead of calling
  50         targetm.stack_protect_guard again.
  51         * dse.c (check_mem_read_rtx): Ignore MEM_VOLATILE_P reads from
  52         MEMs with MEM_EXPR equal to crtl->stack_protect_guard or
  53         crtl->stack_protect_guard_decl.
  54         * config/i386/i386.c (ix86_stack_protect_guard): Set TREE_THIS_VOLATILE
  55         on the returned MEM_EXPR.
  56
  57 2019-01-12  Tom de Vries  <tdevries@suse.de>
  58
  59         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Alow setting
  60         vector length using -fopenacc-dim.
  61
  62 2019-01-12  Tom de Vries  <tdevries@suse.de>
  63
  64         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims): Take larger vector
  65         lengths into account.
  66
  67 2019-01-12  Svante Signell  <svante.signell@gmail.com>
  68
  69         * config/i386/gnu.h (TARGET_THREAD_SSP_OFFSET): Define.
  70         (TARGET_CAN_SPLIT_STACK): Define.
  71         (TARGET_THREAD_SPLIT_STACK_OFFSET): Define.
  72
  73 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
  74
  75         * params.def (inline-unit-growth): Set to 40.
  76
  77 2019-01-12  Jakub Jelinek  <jakub@redhat.com>
  78
  79         * tree-ssa-loop-ivopts.c (find_inv_vars): Fix a comment typo.
  80
  81 2019-01-12  Tom de Vries  <tdevries@suse.de>
  82
  83         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): In offloading
  84         region calling vector-partitionable routine, set default_vector_length
  85         to WARP_SIZE.
  86
  87 2019-01-12  Tom de Vries  <tdevries@suse.de>
  88
  89         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Add an use new
  90         variable default_vector_length.
  91
  92 2019-01-12  Tom de Vries  <tdevries@suse.de>
  93
  94         PR middle-end/88703
  95         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Apply defaults
  96         from oacc_default_dims, as oacc_validate_dims would do it, and apply
  97         dimensions limits.
  98
  99 2019-01-12  Tom de Vries  <tdevries@suse.de>
 100
 101         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1)
 102         (nvptx_goacc_validate_dims): Add used parameter.
 103         * doc/tm.texi: Regenerate.
 104         * omp-offload.c (oacc_parse_default_dims, oacc_validate_dims): Add
 105         argument to call to targetm.goacc.validate_dims.
 106         (default_goacc_validate_dims): Add used
 107         parameter.
 108         * target.def (validate_dims): Add used parameter in DEFHOOK.
 109         * targhooks.h (default_goacc_validate_dims): Add used parameter.
 110
 111 2019-01-11  Jakub Jelinek  <jakub@redhat.com>
 112
 113         PR middle-end/85956
 114         PR lto/88733
 115         * tree-inline.h (struct copy_body_data): Add adjust_array_error_bounds
 116         field.
 117         * tree-inline.c (remap_type_1): Formatting fix.  If TYPE_MAX_VALUE of
 118         ARRAY_TYPE's TYPE_DOMAIN is newly error_mark_node, replace it with
 119         a dummy "omp dummy var" variable if id->adjust_array_error_bounds.
 120         * omp-low.c (new_omp_context): Set cb.adjust_array_error_bounds.
 121
 122 2019-01-11  Vladimir Makarov  <vmakarov@redhat.com>
 123
 124         PR rtl-optimization/87305
 125         * lra-assigns.c
 126         (setup_live_pseudos_and_spill_after_risky_transforms): Add code
 127         for little endian pseudos used as paradoxical subreg.
 128
 129 2019-01-11  Jakub Jelinek  <jakub@redhat.com>
 130
 131         PR tree-optimization/88693
 132         * tree-ssa-strlen.c (get_min_string_length): Don't set *full_string_p
 133         for STRING_CSTs that don't contain any NUL characters in the first
 134         TREE_STRING_LENGTH bytes.
 135
 136 2019-01-11  Alan Modra  <amodra@gmail.com>
 137
 138         PR 88777
 139         PR 88614
 140         * genattrtab.c (min_fn): Don't translate values.
 141         (min_attr_value): Return INT_MAX when the value can't be calculated.
 142         Return minimum among any values that can be calculated.
 143         (max_attr_value): Adjust.
 144
 145 2019-01-11  Jakub Jelinek  <jakub@redhat.com>
 146
 147         * Makefile.in (PLUGIN_HEADERS): Add $(INSN_ATTR_H).
 148
 149 2019-01-11  Steve Ellcey  <sellcey@marvell.com>
 150
 151         * config/aarch64/aarch64.c (aarch64_simd_call_p): New function.
 152         (aarch64_hard_regno_call_part_clobbered): Add insn argument.
 153         (aarch64_return_call_with_max_clobbers): New function.
 154         (TARGET_RETURN_CALL_WITH_MAX_CLOBBERS): New macro.
 155         * config/avr/avr.c (avr_hard_regno_call_part_clobbered): Add insn
 156         argument.
 157         * config/i386/i386.c (ix86_hard_regno_call_part_clobbered): Ditto.
 158         * config/mips/mips.c (mips_hard_regno_call_part_clobbered): Ditto.
 159         * config/rs6000/rs6000.c (rs6000_hard_regno_call_part_clobbered): Ditto.
 160         * config/s390/s390.c (s390_hard_regno_call_part_clobbered): Ditto.
 161         * cselib.c (cselib_process_insn): Add argument to
 162         targetm.hard_regno_call_part_clobbered call.
 163         * ira-conflicts.c (ira_build_conflicts): Ditto.
 164         * ira-costs.c (ira_tune_allocno_costs): Ditto.
 165         * lra-constraints.c (inherit_reload_reg): Ditto.
 166         * lra-int.h (struct lra_reg): Add call_insn field, remove call_p field.
 167         * lra-lives.c (check_pseudos_live_through_calls): Add call_insn
 168         argument.  Call targetm.return_call_with_max_clobbers.
 169         Add argument to targetm.hard_regno_call_part_clobbered call.
 170         (calls_have_same_clobbers_p): New function.
 171         (process_bb_lives): Add call_insn and last_call_insn variables.
 172         Pass call_insn to check_pseudos_live_through_calls.
 173         Modify if stmt to check targetm.return_call_with_max_clobbers.
 174         Update setting of flush variable.
 175         (lra_create_live_ranges_1): Set call_insn to NULL instead of call_p
 176         to false.
 177         * lra.c (initialize_lra_reg_info_element): Set call_insn to NULL.
 178         * regcprop.c (copyprop_hardreg_forward_1): Add argument to
 179         targetm.hard_regno_call_part_clobbered call.
 180         * reginfo.c (choose_hard_reg_mode): Ditto.
 181         * regrename.c (check_new_reg_p): Ditto.
 182         * reload.c (find_equiv_reg): Ditto.
 183         * reload1.c (emit_reload_insns): Ditto.
 184         * sched-deps.c (deps_analyze_insn): Ditto.
 185         * sel-sched.c (init_regs_for_mode): Ditto.
 186         (mark_unavailable_hard_regs): Ditto.
 187         * targhooks.c (default_dwarf_frame_reg_mode): Ditto.
 188         * target.def (hard_regno_call_part_clobbered): Add insn argument.
 189         (return_call_with_max_clobbers): New target function.
 190         * doc/tm.texi: Regenerate.
 191         * doc/tm.texi.in (TARGET_RETURN_CALL_WITH_MAX_CLOBBERS): New hook.
 192         * hooks.c (hook_bool_uint_mode_false): Change to
 193         hook_bool_insn_uint_mode_false.
 194         * hooks.h (hook_bool_uint_mode_false): Ditto.
 195
 196 2019-01-11  Steve Ellcey  <sellcey@marvell.com>
 197
 198         * config/aarch64/aarch64.c (aarch64_simd_call_p): New function.
 199         (aarch64_remove_extra_call_preserved_regs): New function.
 200         (TARGET_REMOVE_EXTRA_CALL_PRESERVED_REGS): New macro.
 201         * doc/tm.texi.in (TARGET_REMOVE_EXTRA_CALL_PRESERVED_REGS): New hook.
 202         * doc/tm.texi: Regenerate.
 203         * final.c (get_call_reg_set_usage): Call new hook.
 204         * target.def (remove_extra_call_preserved_regs): New hook.
 205         * targhooks.c (default_remove_extra_call_preserved_regs): New function.
 206         * targhooks.h (default_remove_extra_call_preserved_regs): New function.
 207
 208 2019-01-11  Jakub Jelinek  <jakub@redhat.com>
 209
 210         PR bootstrap/88714
 211         * passes.c (finish_optimization_passes): Call print_combine_total_stats
 212         inside of pass_combine_1 dump rather than pass_profile_1.
 213
 214 2019-01-11  Tom de Vries  <tdevries@suse.de>
 215
 216         * config/nvptx/nvptx.c (PTX_CTA_NUM_BARRIERS, PTX_PER_CTA_BARRIER)
 217         (PTX_NUM_PER_CTA_BARRIER, PTX_FIRST_PER_WORKER_BARRIER)
 218         (PTX_NUM_PER_WORKER_BARRIERS): Define.
 219         (nvptx_apply_dim_limits): Prevent vector_length 64 and
 220         num_workers 16.
 221
 222 2019-01-11  Tom de Vries  <tdevries@suse.de>
 223
 224         * config/nvptx/nvptx.c (PTX_CTA_SIZE): Move up.
 225
 226 2019-01-11  Jan Beulich  <jbeulich@suse.com>
 227
 228         * config/i386/i386.md (rex64suffix): Add L suffix for SI.
 229         * config/i386/sse.md (cvtusi2<ssescalarmodesuffix>32<round_name>,
 230         sse2_cvtsi2sd): Add {l}.
 231         (sse2_cvtsi2sdq<round_name>): Make q conditional upon AT&T
 232         syntax.
 233
 234 2019-01-10  Jakub Jelinek  <jakub@redhat.com>
 235
 236         PR target/88785
 237         * config/i386/sse.md (float<floatunssuffix>v2div2sf2): Turn into
 238         define_expand.
 239         (*float<floatunssuffix>v2div2sf2): New define_insn.
 240         (float<floatunssuffix>v2div2sf2_mask): Turn into define_expand.
 241         (*float<floatunssuffix>v2div2sf2_mask): New define_insn.
 242         (*float<floatunssuffix>v2div2sf2_mask_1): Replace
 243         subrtxes (const_vector:V2SF [(const_int 0) (const_int 0)]) with
 244         match_operands with "const0_operand" "C".
 245
 246 2019-01-10  Tamar Christina  <tamar.christina@arm.com>
 247
 248         * config/aarch64/aarch64-builtins.c
 249         (aarch64_init_builtins): Move aarch64_init_fcmla_laneq_builtins...
 250         (aarch64_init_simd_builtins): ...Here
 251
 252 2019-01-10  Vladimir Makarov  <vmakarov@redhat.com>
 253
 254         PR rtl-optimization/87305
 255         * lra-assigns.c
 256         (setup_live_pseudos_and_spill_after_risky_transforms): Check
 257         allocation for big endian pseudos used as paradoxical subregs and
 258         spill them if it is wrong.
 259         * lra-constraints.c (lra_constraints): Add a comment.
 260
 261 2019-01-10  Richard Biener  <rguenther@suse.de>
 262
 263         PR tree-optimization/88792
 264         * tree-ssa-pre.c (get_representative_for): Do not return a
 265         value-number here.
 266
 267 2019-01-10  Jakub Jelinek  <jakub@redhat.com>
 268
 269         PR middle-end/84877
 270         PR bootstrap/88450
 271         * function.c (assign_stack_local_1): Revert the 2018-11-21 changes.
 272         (assign_parm_setup_block): Do the argument slot realignment here
 273         instead.
 274
 275 2019-01-10  Stefan Agner  <stefan@agner.ch>
 276
 277         PR target/88648
 278         * config/arm/arm.c (arm_option_override_internal): Force
 279         opts->x_inline_asm_unified to true only if TARGET_THUMB2_P.
 280
 281 2019-01-10  Jakub Jelinek  <jakub@redhat.com>
 282
 283         PR c/88568
 284         * attribs.c (handle_dll_attribute): Clear TREE_STATIC after setting
 285         DECL_EXTERNAL.
 286
 287 2019-01-10  Tamar Christina  <tamar.christina@arm.com>
 288
 289         * config/arm/arm-builtins.c
 290         (enum arm_type_qualifiers): Add qualifier_lane_pair_index.
 291         (MAC_LANE_PAIR_QUALIFIERS): New.
 292         (arm_expand_builtin_args): Use it.
 293         (arm_expand_builtin_1): Likewise.
 294         * config/arm/arm-protos.h (neon_vcmla_lane_prepare_operands): New.
 295         * config/arm/arm.c (neon_vcmla_lane_prepare_operands): New.
 296         * config/arm/arm-c.c (arm_cpu_builtins): Add __ARM_FEATURE_COMPLEX.
 297         * config/arm/arm_neon.h:
 298         (vcadd_rot90_f16): New.
 299         (vcaddq_rot90_f16): New.
 300         (vcadd_rot270_f16): New.
 301         (vcaddq_rot270_f16): New.
 302         (vcmla_f16): New.
 303         (vcmlaq_f16): New.
 304         (vcmla_lane_f16): New.
 305         (vcmla_laneq_f16): New.
 306         (vcmlaq_lane_f16): New.
 307         (vcmlaq_laneq_f16): New.
 308         (vcmla_rot90_f16): New.
 309         (vcmlaq_rot90_f16): New.
 310         (vcmla_rot90_lane_f16): New.
 311         (vcmla_rot90_laneq_f16): New.
 312         (vcmlaq_rot90_lane_f16): New.
 313         (vcmlaq_rot90_laneq_f16): New.
 314         (vcmla_rot180_f16): New.
 315         (vcmlaq_rot180_f16): New.
 316         (vcmla_rot180_lane_f16): New.
 317         (vcmla_rot180_laneq_f16): New.
 318         (vcmlaq_rot180_lane_f16): New.
 319         (vcmlaq_rot180_laneq_f16): New.
 320         (vcmla_rot270_f16): New.
 321         (vcmlaq_rot270_f16): New.
 322         (vcmla_rot270_lane_f16): New.
 323         (vcmla_rot270_laneq_f16): New.
 324         (vcmlaq_rot270_lane_f16): New.
 325         (vcmlaq_rot270_laneq_f16): New.
 326         (vcadd_rot90_f32): New.
 327         (vcaddq_rot90_f32): New.
 328         (vcadd_rot270_f32): New.
 329         (vcaddq_rot270_f32): New.
 330         (vcmla_f32): New.
 331         (vcmlaq_f32): New.
 332         (vcmla_lane_f32): New.
 333         (vcmla_laneq_f32): New.
 334         (vcmlaq_lane_f32): New.
 335         (vcmlaq_laneq_f32): New.
 336         (vcmla_rot90_f32): New.
 337         (vcmlaq_rot90_f32): New.
 338         (vcmla_rot90_lane_f32): New.
 339         (vcmla_rot90_laneq_f32): New.
 340         (vcmlaq_rot90_lane_f32): New.
 341         (vcmlaq_rot90_laneq_f32): New.
 342         (vcmla_rot180_f32): New.
 343         (vcmlaq_rot180_f32): New.
 344         (vcmla_rot180_lane_f32): New.
 345         (vcmla_rot180_laneq_f32): New.
 346         (vcmlaq_rot180_lane_f32): New.
 347         (vcmlaq_rot180_laneq_f32): New.
 348         (vcmla_rot270_f32): New.
 349         (vcmlaq_rot270_f32): New.
 350         (vcmla_rot270_lane_f32): New.
 351         (vcmla_rot270_laneq_f32): New.
 352         (vcmlaq_rot270_lane_f32): New.
 353         (vcmlaq_rot270_laneq_f32): New.
 354         * config/arm/arm_neon_builtins.def (vcadd90, vcadd270, vcmla0, vcmla90,
 355         vcmla180, vcmla270, vcmla_lane0, vcmla_lane90, vcmla_lane180, vcmla_lane270,
 356         vcmla_laneq0, vcmla_laneq90, vcmla_laneq180, vcmla_laneq270,
 357         vcmlaq_lane0, vcmlaq_lane90, vcmlaq_lane180, vcmlaq_lane270): New.
 358         * config/arm/neon.md (neon_vcmla_lane<rot><mode>,
 359         neon_vcmla_laneq<rot><mode>, neon_vcmlaq_lane<rot><mode>): New.
 360         * config/arm/arm.c (arm_arch8_3, arm_arch8_4): New.
 361         * config/arm/arm.h (TARGET_COMPLEX, arm_arch8_3, arm_arch8_4): New.
 362         (arm_option_reconfigure_globals): Use them.
 363         * config/arm/iterators.md (VDF, VQ_HSF): New.
 364         (VCADD, VCMLA): New.
 365         (VF_constraint, rot, rotsplit1, rotsplit2): Add V4HF and V8HF.
 366         * config/arm/neon.md (neon_vcadd<rot><mode>, neon_vcmla<rot><mode>): New.
 367         * config/arm/unspecs.md (UNSPEC_VCADD90, UNSPEC_VCADD270,
 368         UNSPEC_VCMLA, UNSPEC_VCMLA90, UNSPEC_VCMLA180, UNSPEC_VCMLA270): New.
 369
 370 2019-01-10  Tamar Christina  <tamar.christina@arm.com>
 371
 372         * config/aarch64/aarch64-builtins.c (enum aarch64_type_qualifiers): Add qualifier_lane_pair_index.
 373         (emit-rtl.h): Include.
 374         (TYPES_QUADOP_LANE_PAIR): New.
 375         (aarch64_simd_expand_args): Use it.
 376         (aarch64_simd_expand_builtin): Likewise.
 377         (AARCH64_SIMD_FCMLA_LANEQ_BUILTINS, aarch64_fcmla_laneq_builtin_datum): New.
 378         (FCMLA_LANEQ_BUILTIN, AARCH64_SIMD_FCMLA_LANEQ_BUILTIN_BASE,
 379         AARCH64_SIMD_FCMLA_LANEQ_BUILTINS, aarch64_fcmla_lane_builtin_data,
 380         aarch64_init_fcmla_laneq_builtins, aarch64_expand_fcmla_builtin): New.
 381         (aarch64_init_builtins): Add aarch64_init_fcmla_laneq_builtins.
 382         (aarch64_expand_buildin): Add AARCH64_SIMD_BUILTIN_FCMLA_LANEQ0_V2SF,
 383         AARCH64_SIMD_BUILTIN_FCMLA_LANEQ90_V2SF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ180_V2SF,
 384         AARCH64_SIMD_BUILTIN_FCMLA_LANEQ2700_V2SF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ0_V4HF,
 385         AARCH64_SIMD_BUILTIN_FCMLA_LANEQ90_V4HF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ180_V4HF,
 386         AARCH64_SIMD_BUILTIN_FCMLA_LANEQ270_V4HF.
 387         * config/aarch64/iterators.md (FCMLA_maybe_lane): New.
 388         * config/aarch64/aarch64-c.c (aarch64_update_cpp_builtins): Add __ARM_FEATURE_COMPLEX.
 389         * config/aarch64/aarch64-simd-builtins.def (fcadd90, fcadd270, fcmla0, fcmla90,
 390         fcmla180, fcmla270, fcmla_lane0, fcmla_lane90, fcmla_lane180, fcmla_lane270,
 391         fcmla_laneq0, fcmla_laneq90, fcmla_laneq180, fcmla_laneq270,
 392         fcmlaq_lane0, fcmlaq_lane90, fcmlaq_lane180, fcmlaq_lane270): New.
 393         * config/aarch64/aarch64-simd.md (aarch64_fcmla_lane<rot><mode>,
 394         aarch64_fcmla_laneq<rot>v4hf, aarch64_fcmlaq_lane<rot><mode>,aarch64_fcadd<rot><mode>,
 395         aarch64_fcmla<rot><mode>): New.
 396         * config/aarch64/arm_neon.h:
 397         (vcadd_rot90_f16): New.
 398         (vcaddq_rot90_f16): New.
 399         (vcadd_rot270_f16): New.
 400         (vcaddq_rot270_f16): New.
 401         (vcmla_f16): New.
 402         (vcmlaq_f16): New.
 403         (vcmla_lane_f16): New.
 404         (vcmla_laneq_f16): New.
 405         (vcmlaq_lane_f16): New.
 406         (vcmlaq_rot90_lane_f16): New.
 407         (vcmla_rot90_laneq_f16): New.
 408         (vcmla_rot90_lane_f16): New.
 409         (vcmlaq_rot90_f16): New.
 410         (vcmla_rot90_f16): New.
 411         (vcmlaq_laneq_f16): New.
 412         (vcmla_rot180_laneq_f16): New.
 413         (vcmla_rot180_lane_f16): New.
 414         (vcmlaq_rot180_f16): New.
 415         (vcmla_rot180_f16): New.
 416         (vcmlaq_rot90_laneq_f16): New.
 417         (vcmlaq_rot270_laneq_f16): New.
 418         (vcmlaq_rot270_lane_f16): New.
 419         (vcmla_rot270_laneq_f16): New.
 420         (vcmlaq_rot270_f16): New.
 421         (vcmla_rot270_f16): New.
 422         (vcmlaq_rot180_laneq_f16): New.
 423         (vcmlaq_rot180_lane_f16): New.
 424         (vcmla_rot270_lane_f16): New.
 425         (vcadd_rot90_f32): New.
 426         (vcaddq_rot90_f32): New.
 427         (vcaddq_rot90_f64): New.
 428         (vcadd_rot270_f32): New.
 429         (vcaddq_rot270_f32): New.
 430         (vcaddq_rot270_f64): New.
 431         (vcmla_f32): New.
 432         (vcmlaq_f32): New.
 433         (vcmlaq_f64): New.
 434         (vcmla_lane_f32): New.
 435         (vcmla_laneq_f32): New.
 436         (vcmlaq_lane_f32): New.
 437         (vcmlaq_laneq_f32): New.
 438         (vcmla_rot90_f32): New.
 439         (vcmlaq_rot90_f32): New.
 440         (vcmlaq_rot90_f64): New.
 441         (vcmla_rot90_lane_f32): New.
 442         (vcmla_rot90_laneq_f32): New.
 443         (vcmlaq_rot90_lane_f32): New.
 444         (vcmlaq_rot90_laneq_f32): New.
 445         (vcmla_rot180_f32): New.
 446         (vcmlaq_rot180_f32): New.
 447         (vcmlaq_rot180_f64): New.
 448         (vcmla_rot180_lane_f32): New.
 449         (vcmla_rot180_laneq_f32): New.
 450         (vcmlaq_rot180_lane_f32): New.
 451         (vcmlaq_rot180_laneq_f32): New.
 452         (vcmla_rot270_f32): New.
 453         (vcmlaq_rot270_f32): New.
 454         (vcmlaq_rot270_f64): New.
 455         (vcmla_rot270_lane_f32): New.
 456         (vcmla_rot270_laneq_f32): New.
 457         (vcmlaq_rot270_lane_f32): New.
 458         (vcmlaq_rot270_laneq_f32): New.
 459         * config/aarch64/aarch64.h (TARGET_COMPLEX): New.
 460         * config/aarch64/iterators.md (UNSPEC_FCADD90, UNSPEC_FCADD270,
 461         UNSPEC_FCMLA, UNSPEC_FCMLA90, UNSPEC_FCMLA180, UNSPEC_FCMLA270): New.
 462         (FCADD, FCMLA): New.
 463         (rot): New.
 464         * config/arm/types.md (neon_fcadd, neon_fcmla): New.
 465
 466 2019-01-09  Sandra Loosemore  <sandra@codesourcery.com>
 467
 468         PR other/16615
 469
 470         * config/pa/pa.c: Change "can not" to "cannot".
 471         * gimple-ssa-evrp-analyze.c: Likewise.
 472         * ipa-icf.c: Likewise.
 473         * ipa-polymorphic-call.c: Likewise.
 474         * ipa-pure-const.c: Likewise.
 475         * lra-constraints.c: Likewise.
 476         * lra-remat.c: Likewise.
 477         * reload1.c: Likewise.
 478         * reorg.c: Likewise.
 479         * tree-ssa-uninit.c: Likewise.
 480
 481 2019-01-09  Sandra Loosemore  <sandra@codesourcery.com>
 482
 483         PR other/16615
 484
 485         * Makefile.in: Mechanically replace "can not" with "cannot".
 486         * alias.c: Likewise.
 487         * builtins.c: Likewise.
 488         * calls.c: Likewise.
 489         * cgraph.c: Likewise.
 490         * cgraph.h: Likewise.
 491         * cgraphclones.c: Likewise.
 492         * cgraphunit.c: Likewise.
 493         * combine-stack-adj.c: Likewise.
 494         * combine.c: Likewise.
 495         * common/config/i386/i386-common.c: Likewise.
 496         * config/aarch64/aarch64.c: Likewise.
 497         * config/alpha/sync.md: Likewise.
 498         * config/arc/arc.c: Likewise.
 499         * config/arc/predicates.md: Likewise.
 500         * config/arm/arm-c.c: Likewise.
 501         * config/arm/arm.c: Likewise.
 502         * config/arm/arm.h: Likewise.
 503         * config/arm/arm.md: Likewise.
 504         * config/arm/cortex-r4f.md: Likewise.
 505         * config/csky/csky.c: Likewise.
 506         * config/csky/csky.h: Likewise.
 507         * config/darwin-f.c: Likewise.
 508         * config/epiphany/epiphany.md: Likewise.
 509         * config/i386/i386.c: Likewise.
 510         * config/i386/sol2.h: Likewise.
 511         * config/m68k/m68k.c: Likewise.
 512         * config/mcore/mcore.h: Likewise.
 513         * config/microblaze/microblaze.md: Likewise.
 514         * config/mips/20kc.md: Likewise.
 515         * config/mips/sb1.md: Likewise.
 516         * config/nds32/nds32.c: Likewise.
 517         * config/nds32/predicates.md: Likewise.
 518         * config/pa/pa.c: Likewise.
 519         * config/rs6000/e300c2c3.md: Likewise.
 520         * config/rs6000/rs6000.c: Likewise.
 521         * config/s390/s390.h: Likewise.
 522         * config/sh/sh.c: Likewise.
 523         * config/sh/sh.md: Likewise.
 524         * config/spu/vmx2spu.h: Likewise.
 525         * cprop.c: Likewise.
 526         * dbxout.c: Likewise.
 527         * df-scan.c: Likewise.
 528         * doc/cfg.texi: Likewise.
 529         * doc/extend.texi: Likewise.
 530         * doc/fragments.texi: Likewise.
 531         * doc/gty.texi: Likewise.
 532         * doc/invoke.texi: Likewise.
 533         * doc/lto.texi: Likewise.
 534         * doc/md.texi: Likewise.
 535         * doc/objc.texi: Likewise.
 536         * doc/rtl.texi: Likewise.
 537         * doc/tm.texi: Likewise.
 538         * dse.c: Likewise.
 539         * emit-rtl.c: Likewise.
 540         * emit-rtl.h: Likewise.
 541         * except.c: Likewise.
 542         * expmed.c: Likewise.
 543         * expr.c: Likewise.
 544         * fold-const.c: Likewise.
 545         * genautomata.c: Likewise.
 546         * gimple-fold.c: Likewise.
 547         * hard-reg-set.h: Likewise.
 548         * ifcvt.c: Likewise.
 549         * ipa-comdats.c: Likewise.
 550         * ipa-cp.c: Likewise.
 551         * ipa-devirt.c: Likewise.
 552         * ipa-fnsummary.c: Likewise.
 553         * ipa-icf.c: Likewise.
 554         * ipa-inline-transform.c: Likewise.
 555         * ipa-inline.c: Likewise.
 556         * ipa-polymorphic-call.c: Likewise.
 557         * ipa-profile.c: Likewise.
 558         * ipa-prop.c: Likewise.
 559         * ipa-pure-const.c: Likewise.
 560         * ipa-reference.c: Likewise.
 561         * ipa-split.c: Likewise.
 562         * ipa-visibility.c: Likewise.
 563         * ipa.c: Likewise.
 564         * ira-build.c: Likewise.
 565         * ira-color.c: Likewise.
 566         * ira-conflicts.c: Likewise.
 567         * ira-costs.c: Likewise.
 568         * ira-int.h: Likewise.
 569         * ira-lives.c: Likewise.
 570         * ira.c: Likewise.
 571         * ira.h: Likewise.
 572         * loop-invariant.c: Likewise.
 573         * loop-unroll.c: Likewise.
 574         * lower-subreg.c: Likewise.
 575         * lra-assigns.c: Likewise.
 576         * lra-constraints.c: Likewise.
 577         * lra-eliminations.c: Likewise.
 578         * lra-lives.c: Likewise.
 579         * lra-remat.c: Likewise.
 580         * lra-spills.c: Likewise.
 581         * lra.c: Likewise.
 582         * lto-cgraph.c: Likewise.
 583         * lto-streamer-out.c: Likewise.
 584         * postreload-gcse.c: Likewise.
 585         * predict.c: Likewise.
 586         * profile-count.h: Likewise.
 587         * profile.c: Likewise.
 588         * recog.c: Likewise.
 589         * ree.c: Likewise.
 590         * reload.c: Likewise.
 591         * reload1.c: Likewise.
 592         * reorg.c: Likewise.
 593         * resource.c: Likewise.
 594         * rtl.def: Likewise.
 595         * rtl.h: Likewise.
 596         * rtlanal.c: Likewise.
 597         * sched-deps.c: Likewise.
 598         * sched-ebb.c: Likewise.
 599         * sched-rgn.c: Likewise.
 600         * sel-sched-ir.c: Likewise.
 601         * sel-sched.c: Likewise.
 602         * shrink-wrap.c: Likewise.
 603         * simplify-rtx.c: Likewise.
 604         * symtab.c: Likewise.
 605         * target.def: Likewise.
 606         * toplev.c: Likewise.
 607         * tree-call-cdce.c: Likewise.
 608         * tree-cfg.c: Likewise.
 609         * tree-complex.c: Likewise.
 610         * tree-core.h: Likewise.
 611         * tree-eh.c: Likewise.
 612         * tree-inline.c: Likewise.
 613         * tree-loop-distribution.c: Likewise.
 614         * tree-nrv.c: Likewise.
 615         * tree-profile.c: Likewise.
 616         * tree-sra.c: Likewise.
 617         * tree-ssa-alias.c: Likewise.
 618         * tree-ssa-dce.c: Likewise.
 619         * tree-ssa-dom.c: Likewise.
 620         * tree-ssa-forwprop.c: Likewise.
 621         * tree-ssa-loop-im.c: Likewise.
 622         * tree-ssa-loop-ivcanon.c: Likewise.
 623         * tree-ssa-loop-ivopts.c: Likewise.
 624         * tree-ssa-loop-niter.c: Likewise.
 625         * tree-ssa-phionlycprop.c: Likewise.
 626         * tree-ssa-phiopt.c: Likewise.
 627         * tree-ssa-propagate.c: Likewise.
 628         * tree-ssa-threadedge.c: Likewise.
 629         * tree-ssa-threadupdate.c: Likewise.
 630         * tree-ssa-uninit.c: Likewise.
 631         * tree-ssanames.c: Likewise.
 632         * tree-streamer-out.c: Likewise.
 633         * tree.c: Likewise.
 634         * tree.h: Likewise.
 635         * vr-values.c: Likewise.
 636
 637 2019-01-09  Uroš Bizjak  <ubizjak@gmail.com>
 638
 639         * config/i386/i386-protos.h (ix86_expand_xorsign): New prototype.
 640         (ix86_split_xorsign): Ditto.
 641         * config/i386/i386.c (ix86_expand_xorsign): New function.
 642         (ix86_split_xorsign): Ditto.
 643         * config/i386/i386.md (UNSPEC_XORSIGN): New unspec.
 644         (xorsign<mode>3): New expander.
 645         (xorsign<mode>3_1): New insn_and_split pattern.
 646         * config/i386/sse.md (xorsign<mode>3): New expander.
 647
 648 2019-01-09  Eric Botcazou  <ebotcazou@adacore.com>
 649
 650         * config/sparc/sparc.md (*tablejump_sp32): Merge into...
 651         (*tablejump_sp64): Likewise.
 652         (*tablejump<P:mode>): ...this.
 653         (*call_address_sp32): Merge into...
 654         (*call_address_sp64): Likewise.
 655         (*call_address<P:mode>): ...this.
 656         (*call_symbolic_sp32): Merge into...
 657         (*call_symbolic_sp64): Likewise.
 658         (*call_symbolic<P:mode>): ...this.
 659         (call_value): Remove constraint and add predicate.
 660         (*call_value_address_sp32): Merge into...
 661         (*call_value_address_sp64): Likewise.
 662         (*call_value_address<P:mode>): ...this.
 663         (*call_value_symbolic_sp32): Merge into...
 664         (*call_value_symbolic_sp64): Likewise.
 665         (*call_value_symbolic<P:mode>): ...this.
 666         (*sibcall_symbolic_sp32): Merge into...
 667         (*sibcall_symbolic_sp64): Likewise.
 668         (*sibcall_symbolic<P:mode>): ...this.
 669         (sibcall_value): Remove constraint and add predicate.
 670         (*sibcall_value_symbolic_sp32): Merge into...
 671         (*sibcall_value_symbolic_sp64): Likewise.
 672         (*sibcall_value_symbolic<P:mode>): ...this.
 673         (window_save): Minor tweak.
 674         (*branch_sp32): Merge into...
 675         (*branch_sp64): Likewise.
 676         (*branch<P:mode>): ...this.
 677
 678 2019-01-09  Eric Botcazou  <ebotcazou@adacore.com>
 679             James Clarke  <jrtc27@jrtc27.com>
 680
 681         PR target/84010
 682         * config/sparc/sparc.c (sparc_legitimize_tls_address): Only use Pmode
 683         consistently in TLS address generation and adjust code to the renaming
 684         of patterns.  Mark calls to __tls_get_addr as const.
 685         * config/sparc/sparc.md (tgd_hi22): Turn into...
 686         (tgd_hi22<P:mode>): ...this and use Pmode throughout.
 687         (tgd_lo10): Turn into...
 688         (tgd_lo10<P:mode>): ...this and use Pmode throughout.
 689         (tgd_add32): Merge into...
 690         (tgd_add64): Likewise.
 691         (tgd_add<P:mode>): ...this and use Pmode throughout.
 692         (tldm_hi22): Turn into...
 693         (tldm_hi22<P:mode>): ...this and use Pmode throughout.
 694         (tldm_lo10): Turn into...
 695         (tldm_lo10<P:mode>): ...this and use Pmode throughout.
 696         (tldm_add32): Merge into...
 697         (tldm_add64): Likewise.
 698         (tldm_add<P:mode>): ...this and use Pmode throughout.
 699         (tldm_call32): Merge into...
 700         (tldm_call64): Likewise.
 701         (tldm_call<P:mode>): ...this and use Pmode throughout.
 702         (tldo_hix22): Turn into...
 703         (tldo_hix22<P:mode>): ...this and use Pmode throughout.
 704         (tldo_lox10): Turn into...
 705         (tldo_lox10<P:mode>): ...this and use Pmode throughout.
 706         (tldo_add32): Merge into...
 707         (tldo_add64): Likewise.
 708         (tldo_add<P:mode>): ...this and use Pmode throughout.
 709         (tie_hi22): Turn into...
 710         (tie_hi22<P:mode>): ...this and use Pmode throughout.
 711         (tie_lo10): Turn into...
 712         (tie_lo10<P:mode>): ...this and use Pmode throughout.
 713         (tie_ld64): Use DImode throughout.
 714         (tie_add32): Merge into...
 715         (tie_add64): Likewise.
 716         (tie_add<P:mode>): ...this and use Pmode throughout.
 717         (tle_hix22_sp32): Merge into...
 718         (tle_hix22_sp64): Likewise.
 719         (tle_hix22<P:mode>): ...this and use Pmode throughout.
 720         (tle_lox22_sp32): Merge into...
 721         (tle_lox22_sp64): Likewise.
 722         (tle_lox22<P:mode>): ...this and use Pmode throughout.
 723         (*tldo_ldub_sp32): Merge into...
 724         (*tldo_ldub_sp64): Likewise.
 725         (*tldo_ldub<P:mode>): ...this and use Pmode throughout.
 726         (*tldo_ldub1_sp32): Merge into...
 727         (*tldo_ldub1_sp64): Likewise.
 728         (*tldo_ldub1<P:mode>): ...this and use Pmode throughout.
 729         (*tldo_ldub2_sp32): Merge into...
 730         (*tldo_ldub2_sp64): Likewise.
 731         (*tldo_ldub2<P:mode>): ...this and use Pmode throughout.
 732         (*tldo_ldsb1_sp32): Merge into...
 733         (*tldo_ldsb1_sp64): Likewise.
 734         (*tldo_ldsb1<P:mode>): ...this and use Pmode throughout.
 735         (*tldo_ldsb2_sp32): Merge into...
 736         (*tldo_ldsb2_sp64): Likewise.
 737         (*tldo_ldsb2<P:mode>): ...this and use Pmode throughout.
 738         (*tldo_ldub3_sp64): Use DImode throughout.
 739         (*tldo_ldsb3_sp64): Likewise.
 740         (*tldo_lduh_sp32): Merge into...
 741         (*tldo_lduh_sp64): Likewise.
 742         (*tldo_lduh<P:mode>): ...this and use Pmode throughout.
 743         (*tldo_lduh1_sp32): Merge into...
 744         (*tldo_lduh1_sp64): Likewise.
 745         (*tldo_lduh1<P:mode>): ...this and use Pmode throughout.
 746         (*tldo_ldsh1_sp32): Merge into...
 747         (*tldo_ldsh1_sp64): Likewise.
 748         (*tldo_ldsh1<P:mode>): ...this and use Pmode throughout.
 749         (*tldo_lduh2_sp64): Use DImode throughout.
 750         (*tldo_ldsh2_sp64): Likewise.
 751         (*tldo_lduw_sp32): Merge into...
 752         (*tldo_lduw_sp64): Likewise.
 753         (*tldo_lduw<P:mode>): ...this and use Pmode throughout.
 754         (*tldo_lduw1_sp64): Use DImode throughout.
 755         (*tldo_ldsw1_sp64): Likewise.
 756         (*tldo_ldx_sp64): Likewise.
 757         (*tldo_stb_sp32): Merge into...
 758         (*tldo_stb_sp64): Likewise.
 759         (*tldo_stb<P:mode>): ...this and use Pmode throughout.
 760         (*tldo_sth_sp32): Merge into...
 761         (*tldo_sth_sp64): Likewise.
 762         (*tldo_sth<P:mode>): ...this and use Pmode throughout.
 763         (*tldo_stw_sp32): Merge into...
 764         (*tldo_stw_sp64): Likewise.
 765         (*tldo_stw<P:mode>): ...this and use Pmode throughout.
 766         (*tldo_stx_sp64): Use DImode throughout.
 767
 768 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 769
 770         * config/aarch64/aarch64.c (aarch64_override_options): Add case to
 771         check configure option to set BTI and Return Address Signing.
 772         * configure.ac: Add --enable-standard-branch-protection and
 773         --disable-standard-branch-protection.
 774         * configure: Regenerated.
 775         * doc/install.texi: Document the same.
 776
 777 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 778             Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
 779
 780         * config.gcc (aarch64*-*-*): Add aarch64-bti-insert.o.
 781         * config/aarch64/aarch64.h: Update comment for TRAMPOLINE_SIZE.
 782         * config/aarch64/aarch64.c (aarch64_asm_trampoline_template): Update
 783         if bti is enabled.
 784         * config/aarch64/aarch64-bti-insert.c: New file.
 785         * config/aarch64/aarch64-passes.def (INSERT_PASS_BEFORE): Insert bti
 786         pass.
 787         * config/aarch64/aarch64-protos.h (make_pass_insert_bti): Declare the
 788         new bti pass.
 789         * config/aarch64/aarch64.md (unspecv): Add UNSPECV_BTI_NOARG,
 790         UNSPECV_BTI_C, UNSPECV_BTI_J and UNSPECV_BTI_JC.
 791         (bti_noarg, bti_j, bti_c, bti_jc): New define_insns.
 792         * config/aarch64/t-aarch64: Add rule for aarch64-bti-insert.o.
 793
 794 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 795
 796         * config/aarch64/aarch64-protos.h (aarch64_bti_enabled): Declare.
 797         * config/aarch64/aarch64.c (aarch64_handle_no_branch_protection):
 798         Disable bti for -mbranch-protection=none.
 799         (aarch64_handle_standard_branch_protection): Enable bti for
 800         -mbranch-protection=standard.
 801         (aarch64_handle_bti_protection): Enable bti for "bti" in the string to
 802         -mbranch-protection.
 803         (aarch64_bti_enabled): Check if bti is enabled.
 804         * config/aarch64/aarch64.opt: Declare target variable.
 805         * doc/invoke.texi: Add bti to the -mbranch-protection documentation.
 806
 807 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 808
 809         * config/aarch64/aarch64.c (aarch64_expand_prologue): Use new
 810         epilogue/prologue scratch registers EP0_REGNUM and EP1_REGNUM.
 811         (aarch64_expand_epilogue): Likewise.
 812         (aarch64_output_mi_thunk): Likewise
 813         * config/aarch64/aarch64.h (REG_CLASS_CONTENTS): Change
 814         TAILCALL_ADDR_REGS to x16 and x17.
 815         * config/aarch64/aarch64.md: Define EP0_REGNUM and EP1_REGNUM.
 816
 817 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 818
 819         * config/aarch64/aarch64-option-extensions.def: Define
 820         AARCH64_OPT_EXTENSION for memtag, rng, sb, ssbs and predres.
 821         * gcc/config/aarch64/aarch64.h (AARCH64_FL_RNG): New.
 822         (AARCH64_FL_MEMTAG, ARCH64_FL_SB, AARCH64_FL_SSBS): New.
 823         (AARCH64_FL_PREDRES): New.
 824         (AARCH64_FL_FOR_ARCH8_5): Add AARCH64_FL_SB, AARCH64_FL_SSBS and
 825         AARCH64_FL_PREDRES by default.
 826         * gcc/doc/invoke.texi: Document rng, memtag, sb, ssbs and predres.
 827
 828 2018-01-09  Sudakshina Das  <sudi.das@arm.com>
 829
 830         * config/aarch64/aarch64-arches.def: Define AARCH64_ARCH for
 831         ARMv8.5-A.
 832         * gcc/config/aarch64/aarch64.h (AARCH64_FL_V8_5): New.
 833         (AARCH64_FL_FOR_ARCH8_5, AARCH64_ISA_V8_5): New.
 834         * gcc/doc/invoke.texi: Document ARMv8.5-A.
 835
 836 2019-01-09  Alejandro Martinez  <alejandro.martinezvicente@arm.com>
 837
 838         * config/aarch64/aarch64-sve.md (copysign<mode>3): New define_expand.
 839         (xorsign<mode>3): Likewise.
 840
 841 2019-01-09  Jelinek  <jakub@redhat.com>
 842
 843         PR middle-end/88758
 844         * tree.c (initializer_each_zero_or_onep) <case VECTOR_CST>: Use
 845         vector_cst_elt instead of VECTOR_CST_ENCODED_ELT.
 846
 847         PR rtl-optimization/88331
 848         * function.c (assign_stack_local_1): Don't set dynamic_align_addr if
 849         not currently_expanding_to_rtl.
 850
 851 2019-01-09  Eric Botcazou  <ebotcazou@adacore.com>
 852
 853         * doc/invoke.texi (-Os): Remove trailing spaces.
 854         (-finline-functions): Remove reference to -O2.
 855
 856 2019-01-08  Jakub Jelinek  <jakub@redhat.com>
 857
 858         PR rtl-optimization/79593
 859         * config/i386/i386.md (reg = mem; mem = reg): New define_peephole2.
 860
 861         * config/rs6000/rs6000.c (rs6000_delegitimize_address): Delegitimize
 862         UNSPEC_FUSION_GPR to its argument.  Formatting fixes.
 863
 864 2019-01-08  Eric Botcazou  <ebotcazou@adacore.com>
 865
 866         PR bootstrap/88721
 867         * config/sparc/sparc.c (function_arg_slotno): Set *PPREGNO & *PPADDING
 868         to -1 on entry.
 869
 870         PR debug/88723
 871         * config/sparc/sparc.c (sparc_delegitimize_address): Deal with naked
 872         UNSPECs and UNSPEC_MOVE_GOTDATA specifically.
 873
 874 2019-01-08  H.J. Lu  <hongjiu.lu@intel.com>
 875
 876         PR target/88717
 877         * config/i386/i386.c (ix86_avx_u128_mode_exit): Call
 878         ix86_avx_u128_mode_entry.
 879
 880 2019-01-08  Martin Liska  <mliska@suse.cz>
 881
 882         PR tree-optimization/88753
 883         * tree-switch-conversion.c (switch_conversion::build_one_array):
 884         Come up with local variable constructor.  Convert first to
 885         type of constructor values.
 886
 887 2019-01-08  Richard Biener  <rguenther@suse.de>
 888
 889         PR tree-optimization/86554
 890         * tree-ssa-sccvn.c (eliminate_dom_walker, rpo_elim,
 891         rpo_avail): Move earlier.
 892         (visit_nary_op): When value-numbering to expressions
 893         with different overflow behavior make sure there's an
 894         available expression on the path.
 895
 896 2019-01-08  Sam Tebbs  <sam.tebbs@arm.com>
 897
 898         * config/aarch64/aarch64.c (BRANCH_PROTECT_STR_MAX,
 899         aarch64_parse_branch_protection,
 900         struct aarch64_branch_protect_type,
 901         aarch64_handle_no_branch_protection,
 902         aarch64_handle_standard_branch_protection,
 903         aarch64_validate_mbranch_protection,
 904         aarch64_handle_pac_ret_protection,
 905         aarch64_handle_attr_branch_protection,
 906         accepted_branch_protection_string,
 907         aarch64_pac_ret_subtypes,
 908         aarch64_branch_protect_types,
 909         aarch64_handle_pac_ret_leaf): Define.
 910         (aarch64_override_options_after_change_1, aarch64_override_options):
 911         Add check for accepted_branch_protection_string.
 912         (aarch64_option_save): Save accepted_branch_protection_string.
 913         (aarch64_option_restore): Save accepted_branch_protection_string.
 914         * config/aarch64/aarch64.c (aarch64_attributes): Add branch-protection.
 915         * config/aarch64/aarch64.opt: Add mbranch-protection. Deprecate
 916         msign-return-address.
 917         * doc/invoke.texi: Add mbranch-protection.
 918
 919 2019-01-08  Alan Modra  <amodra@gmail.com>
 920
 921         PR target/88614
 922         * genattrtab.c (max_attr_value, min_attr_value, or_attr_value):
 923         Delete "unknownp" parameter.  Adjust callers.  Handle
 924         CONST_INT, PLUS, MINUS, and MULT.
 925         (attr_value_aligned): Renamed from or_attr_value.
 926         (min_attr_value): Return INT_MIN for unhandled rtl case..
 927         (min_fn): ..and translate to INT_MAX here.
 928         (write_length_unit_log): Modify to cope without "unknown".
 929         (write_attr_value): Handle IF_THEN_ELSE.
 930
 931 2019-01-07  Richard Sandiford  <richard.sandiford@arm.com>
 932
 933         * tree-vect-stmts.c (vectorizable_store): Don't use the dataref_offset
 934         optimization for masked stores.
 935
 936 2019-01-07  Richard Sandiford  <richard.sandiford@arm.com>
 937
 938         PR middle-end/88567
 939         * tree-vect-loop.c (get_initial_defs_for_reduction): Pass the
 940         output vector directly to duplicate_and_interleave instead of
 941         going through a temporary.  Postpone insertion of ctor_seq to
 942         the end of the loop.
 943
 944 2019-01-07  Richard Earnshaw  <rearnsha@arm.com>
 945
 946         PR target/86891
 947         * config/aarch64/aarch64.c (aarch64_expand_subvti): New parameter
 948         unsigned_p.  Handle signed and unsigned overflow correction as
 949         required.
 950         * config/aarch64/aarch64-protos.h (aarch64_expand_subvti): Update
 951         prototype.
 952         * config/aarch64/aarch64.md (addv<mode>4): Use aarch64_plus_operand
 953         for operand 2.
 954         (add<mode>3_compareV_imm): Make this callable for expanding.
 955         (subv<GPI:mode>4): Use register_operand for operand 1.  Use
 956         aarch64_plus_operand for operand 2.
 957         (subv<GPI:mode>_insn): New insn pattern.
 958         (subv<GPI:mode>_imm): Likewise.
 959         (negv<GPI:mode>3): New expand pattern.
 960         (negv<GPI:mode>_insn): New insn pattern.
 961         (negv<GPI:mode>_cmp_only): Likewise.
 962         (cmpv<GPI:mode>_insn): Likewise.
 963         (subvti4): Use register_operand for operand 1.  Update call to
 964         aarch64_expand_subvti.
 965         (usubvti4): Likewise.
 966         (negvti3): New expand pattern.
 967         (negdi_carryout): New insn pattern.
 968         (negvdi_carryinV): New insn pattern.
 969         (sub<mode3>_compare1_imm): Delete named insn pattern, make anonymous
 970         version the named version.
 971         (peepholes to convert to sub<mode3>_compare1_imm): Adjust order of
 972         operands.
 973         (usub<GPI:mode>3_carryinC, usub<GPI:mode>3_carryinC_z1): New insn
 974         patterns.
 975         (usub<GPI:mode>3_carryinC_z2, usub<GPI:mode>3_carryinC): New insn
 976         patterns.
 977         (sub<mode>3_carryinCV, sub<mode>3_carryinCV_z1_z2): Delete.
 978         (sub<mode>3_carryinCV_z1, sub<mode>3_carryinCV_z2): Delete.
 979         (sub<mode>3_carryinCV): Delete.
 980         (sub<GPI:mode>3_carryinV): New expand pattern.
 981         sub<mode>3_carryinV, sub<mode>3_carryinV_z2): New insn patterns.
 982
 983 2019-01-07  Richard Biener  <rguenther@suse.de>
 984
 985         * tree-ssa-uncprop.c (ssa_equip_hash_traits): Remove in favor
 986         of tree_operand_hash.
 987
 988 2019-01-07  Richard Sandiford  <richard.sandiford@arm.com>
 989
 990         PR tree-optimization/88598
 991         * tree.h (single_nonzero_element): Declare.
 992         * tree.c (single_nonzero_element): New function.
 993         * match.pd: Fold certain reductions of X & CST to X[I] & CST[I]
 994         if I is the only nonzero element of CST.
 995
 996 2019-01-07  Richard Sandiford  <richard.sandiford@arm.com>
 997
 998         PR tree-optimization/88598
 999         * tree.h (initializer_each_zero_or_onep): Declare.
1000         * tree.c (initializer_each_zero_or_onep): New function.
1001         (signed_or_unsigned_type_for): Handle float types too.
1002         (unsigned_type_for, signed_type_for): Update comments accordingly.
1003         * match.pd: Fold x * { 0 or 1, 0 or 1, ...} to
1004         x & { 0 or -1, 0 or -1, ... }.
1005
1006 2019-01-07  Jonathan Wakely  <jwakely@redhat.com>
1007
1008         * doc/install.texi: Replace references to x86_64-unknown-linux-gnu
1009         with x86_64-pc-linux-gnu.
1010
1011 2019-01-07  Tom de Vries  <tdevries@suse.de>
1012
1013         PR target/85486
1014         * config/nvptx/nvptx.c (has_vector_partitionable_routine_calls_p): New
1015         function.
1016         (nvptx_goacc_validate_dims): Force vl32 if calling vector-partitionable
1017         routines.
1018
1019 2019-01-07  Jakub Jelinek  <jakub@redhat.com>
1020
1021         * config/i386/sse.md (vec_extract<mode><ssehalfvecmodelower>): Use
1022         V_256_512 iterator instead of V_512 and TARGET_AVX instead of
1023         TARGET_AVX512F as condition.
1024
1025         PR debug/88723
1026         * dwarf2out.c (const_ok_for_output_1): Remove redundant call to
1027         const_not_ok_for_debug_p target hook.
1028         (mem_loc_descriptor) <case UNSPEC>: Only call const_ok_for_output_1
1029         on UNSPEC and subexpressions thereof if all subexpressions of the
1030         UNSPEC are CONSTANT_P.
1031
1032         PR tree-optimization/88676
1033         * tree-ssa-phiopt.c (two_value_replacement): New function.
1034         (tree_ssa_phiopt_worker): Call it.
1035
1036         PR sanitizer/88619
1037         * cfgexpand.c (expand_stack_vars): Only align prev_offset to
1038         ASAN_MIN_RED_ZONE_SIZE, not to maximum of that and alignb.
1039
1040         PR c++/85052
1041         * tree-vect-generic.c: Include insn-config.h and recog.h.
1042         (expand_vector_piecewise): Add defaulted ret_type argument,
1043         if non-NULL, use that in preference to type for the result type.
1044         (expand_vector_parallel): Formatting fix.
1045         (do_vec_conversion, do_vec_narrowing_conversion,
1046         expand_vector_conversion): New functions.
1047         (expand_vector_operations_1): Call expand_vector_conversion
1048         for VEC_CONVERT ifn calls.
1049         * internal-fn.def (VEC_CONVERT): New internal function.
1050         * internal-fn.c (expand_VEC_CONVERT): New function.
1051         * fold-const-call.c (fold_const_vec_convert): New function.
1052         (fold_const_call): Use it for CFN_VEC_CONVERT.
1053         * doc/extend.texi (__builtin_convertvector): Document.
1054
1055 2019-01-07  Tom de Vries  <tdevries@suse.de>
1056
1057         * config/nvptx/nvptx-protos.h (nvptx_output_red_partition): Declare.
1058         * config/nvptx/nvptx.c (vector_red_size, vector_red_align,
1059         vector_red_partition, vector_red_sym): New global variables.
1060         (nvptx_option_override): Initialize vector_red_sym.
1061         (nvptx_declare_function_name): Restore red_partition register.
1062         (nvptx_file_end): Emit code to declare the vector reduction variables.
1063         (nvptx_output_red_partition): New function.
1064         (nvptx_expand_shared_addr): Add vector argument. Use it to handle
1065         large vector reductions.
1066         (enum nvptx_builtins): Add NVPTX_BUILTIN_VECTOR_ADDR.
1067         (nvptx_init_builtins): Add VECTOR_ADDR.
1068         (nvptx_expand_builtin): Update call to nvptx_expand_shared_addr.
1069         Handle nvptx_expand_shared_addr.
1070         (nvptx_get_shared_red_addr): Add vector argument and handle large
1071         vectors.
1072         (nvptx_goacc_reduction_setup): Add offload_attrs argument and handle
1073         large vectors.
1074         (nvptx_goacc_reduction_init): Likewise.
1075         (nvptx_goacc_reduction_fini): Likewise.
1076         (nvptx_goacc_reduction_teardown): Likewise.
1077         (nvptx_goacc_reduction): Update calls to nvptx_goacc_reduction_{setup,
1078         init,fini,teardown}.
1079         (nvptx_init_axis_predicate): Initialize vector_red_partition.
1080         (nvptx_set_current_function): Init vector_red_partition.
1081         * config/nvptx/nvptx.md (UNSPECV_RED_PART): New unspecv.
1082         (nvptx_red_partition): New insn.
1083         * config/nvptx/nvptx.h (struct machine_function): Add red_partition.
1084
1085 2019-01-07  Tom de Vries  <tdevries@suse.de>
1086
1087         PR target/85381
1088         * config/nvptx/nvptx.c (nvptx_process_pars): Don't emit barriers for
1089         empty loops.
1090
1091 2019-01-07  Tom de Vries  <tdevries@suse.de>
1092
1093         * config/nvptx/nvptx.c (oacc_bcast_partition): Declare.
1094         (nvptx_option_override): Init oacc_bcast_partition.
1095         (nvptx_init_oacc_workers): New function.
1096         (nvptx_declare_function_name): Call nvptx_init_oacc_workers.
1097         (nvptx_needs_shared_bcast): New function.
1098         (nvptx_find_par): Generalize to enable vectors to use shared-memory
1099         to propagate state.
1100         (nvptx_shared_propagate): Initialize vector bcast partition and
1101         synchronization state.
1102         (nvptx_single):  Generalize to enable vectors to use shared-memory
1103         to propagate state.
1104         (nvptx_process_pars): Likewise.
1105         (nvptx_set_current_function): Initialize oacc_broadcast_partition.
1106         * config/nvptx/nvptx.h (struct machine_function): Add
1107         bcast_partition and sync_bar members.
1108
1109 2019-01-07  Tom de Vries  <tdevries@suse.de>
1110
1111         * config/nvptx/nvptx.c (nvptx_welformed_vector_length_p)
1112         (nvptx_apply_dim_limits): New function.
1113         (nvptx_goacc_validate_dims_1): Allow PTX_MAX_VECTOR_LENGTH larger than
1114         PTX_WARP_SIZE.
1115
1116 2019-01-07  Tom de Vries  <tdevries@suse.de>
1117
1118         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Move warnings to
1119         as late as possible.
1120
1121 2019-01-07  Tom de Vries  <tdevries@suse.de>
1122
1123         * config/nvptx/nvptx.c (PTX_VECTOR_LENGTH): Remove.
1124         (PTX_DEFAULT_VECTOR_LENGTH, PTX_MAX_VECTOR_LENGTH): Define.
1125         (nvptx_goacc_validate_dims_1, nvptx_dim_limit)
1126         (nvptx_goacc_reduction_fini): Use PTX_DEFAULT_VECTOR_LENGTH,
1127         PTX_MAX_VECTOR_LENGTH and PTX_WARP_SIZE instead of PTX_VECTOR_LENGTH.
1128
1129 2019-01-07  Tom de Vries  <tdevries@suse.de>
1130
1131         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims): Add asserts.
1132
1133 2019-01-07  Tom de Vries  <tdevries@suse.de>
1134
1135         * omp-offload.c (oacc_get_min_dim): New function.
1136         * omp-offload.h (oacc_get_min_dim): Declare.
1137
1138 2018-12-26  Mateusz B  <mateuszb@poczta.onet.pl>
1139
1140         PR target/88521
1141         * config/i386/i386.c (function_value_ms_64): Return small sturct in
1142         AX_REG and float/double in FIRST_SSE_REG for 4 or 8 byte modes.
1143
1144 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
1145
1146         PR tree-opt/86020
1147         Revert:
1148         2017-05-22  Jan Hubicka  <hubicka@ucw.cz>
1149
1150         * ipa-inline.c (edge_badness): Use inlined_time instead of
1151         inline_summaries->get.
1152
1153 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
1154
1155         * opts.c (enable_fdo_optimizations): Enable
1156         version-loops-for-strides, loop-interchange, unrol-and-jam
1157         and tree-loop-distribution.
1158         * invoke.texi: Document newly enabled options.
1159
1160 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
1161
1162         * doc/invoke.texi (max-inline-insns-small): New parameters.
1163         * ipa-inline.c (want_early_inline_function_p): simplify.
1164         (want_inline_small_function_p): Fix pasto from previous patch;
1165         use max-inline-insns-small bound.
1166         * params.def (max-inline-insns-small): New param.
1167         * ipa-fnsummary.c (analyze_function_body): Initialize time/size
1168         variables correctly.
1169
1170 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
1171
1172         * doc/invoke.texi: Document max-inline-insns-size,
1173         uninlined-function-insns, uninlined-function-time,
1174         uninlined-thunk-insns and uninlined-thunk-time.
1175         * params.def: Add max-inline-insns-size,
1176         uninlined-function-insns, uninlined-function-time,
1177         uninlined-thunk-insns and uninlined-thunk-time.
1178         * ipa-fnsummary.c (compute_fn_summary, analyze_function_body): Use
1179         new parameters.
1180         * ipa-inline.c (can_inline_edge_by_limits_p,
1181         want_inline_small_function_p): Use new parameters.
1182
1183 2019-01-05  Jan Hubicka  <hubicka@ucw.cz>
1184
1185         * ipa-fnsummary.c (analyze_function_body): Fix accounting of time.
1186
1187 2019-01-05  Jakub Jelinek  <jakub@redhat.com>
1188
1189         PR middle-end/82564
1190         PR target/88620
1191         * expr.c (expand_assignment): For calls returning VLA structures
1192         if to_rtx is not a MEM, force it into a stack temporary.
1193
1194         PR debug/88635
1195         * dwarf2out.c (const_ok_for_output_1): Reject MINUS that contains
1196         SYMBOL_REF, CODE_LABEL or UNSPEC in subexpressions of second argument.
1197         Reject PLUS that contains SYMBOL_REF, CODE_LABEL or UNSPEC in
1198         subexpressions of both operands.
1199         (mem_loc_descriptor): Handle UNSPEC if target hook acks it and all the
1200         subrtxes are CONSTANT_P.
1201         * config/i386/i386.c (ix86_const_not_ok_for_debug_p): Revert
1202         2018-11-09 changes.
1203
1204 2019-01-04  Jan Hubicka  <hubicka@ucw.cz>
1205
1206         * params.def (hot-bb-count-ws-permille): Set to 990.
1207
1208 2019-01-04  Martin Sebor  <msebor@redhat.com>
1209
1210         PR c/88546
1211         * attribs.c (decls_mismatched_attributes): Avoid warning for attribute
1212         leaf.
1213
1214 2019-01-04  Martin Sebor  <msebor@redhat.com>
1215
1216         PR c/88363
1217         * doc/extend.texi (attribute alloc_align, alloc_size): Update.
1218
1219 2019-01-04  Jakub Jelinek  <jakub@redhat.com>
1220
1221         * gdbinit.in: Turn off pagination for the skip commands, restore
1222         it to previous state afterwards.
1223
1224 2019-01-04  Sam Tebbs  <sam.tebbs@arm.com>
1225
1226         PR gcc/87763
1227         * gcc.target/aarch64/combine_bfxil.c: Change scan-assembler-times bfxil
1228         count to 18.
1229
1230 2019-01-04  Jakub Jelinek  <jakub@redhat.com>
1231
1232         PR target/88594
1233         * config/i386/i386.c (ix86_expand_divmod_libfunc): Use mode instead
1234         of GET_MODE (opN) as modes of the libcall arguments.
1235
1236 2019-01-04  Jan Beulich  <jbeulich@suse.com>
1237
1238         * config/i386/sse.md
1239         (<avx512>_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>,
1240         <avx512>_cmp<mode>3<mask_scalar_merge_name>,
1241         <avx512>_ucmp<mode>3<mask_scalar_merge_name>,
1242         <avx512>_ucmp<mode>3<mask_scalar_merge_name>,
1243         avx512f_vmcmp<mode>3<round_saeonly_name>,
1244         avx512f_vmcmp<mode>3_mask<round_saeonly_name>,
1245         avx512f_maskcmp<mode>3,
1246         <avx512>_cvt<ssemodesuffix>2mask<mode>,
1247         <avx512>_cvt<ssemodesuffix>2mask<mode>,
1248         *<avx512>_cvtmask2<ssemodesuffix><mode>,
1249         *<avx512>_cvtmask2<ssemodesuffix><mode>,
1250         <avx512>_eq<mode>3<mask_scalar_merge_name>_1,
1251         <avx512>_eq<mode>3<mask_scalar_merge_name>_1,
1252         <avx512>_gt<mode>3<mask_scalar_merge_name>,
1253         <avx512>_gt<mode>3<mask_scalar_merge_name>,
1254         <avx512>_testm<mode>3<mask_scalar_merge_name>,
1255         <avx512>_testnm<mode>3<mask_scalar_merge_name>,
1256         *<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext,
1257         *<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext_mask,
1258         *<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext,
1259         *<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext_mask,
1260         avx512cd_maskb_vec_dup<mode>,
1261         avx512cd_maskw_vec_dup<mode>,
1262         avx512dq_fpclass<mode><mask_scalar_merge_name>,
1263         avx512dq_vmfpclass<mode>,
1264         avx512vl_vpshufbitqmb<mode><mask_scalar_merge_name>): Use =k
1265         instead of =Yk.
1266
1267 2019-01-03  Martin Sebor  <msebor@redhat.com>
1268
1269         PR tree-optimization/88659
1270         * calls.c (maybe_warn_nonstring_arg): Avoid assuming maxlen is set.
1271
1272 2019-01-03  Aaron Sawdey  <acsawdey@linux.ibm.com>
1273
1274         * config/rs6000/rs6000-string.c (expand_block_move): Don't use
1275         unaligned vsx and avoid lxvd2x/stxvd2x.
1276         (gen_lvx_v4si_move): New function.
1277
1278 2019-01-03  Tom de Vries  <tdevries@suse.de>
1279
1280         * config/nvptx/nvptx.c (MACH_VECTOR_LENGTH, MACH_MAX_WORKERS): Define.
1281         (init_axis_dim, nvptx_mach_max_workers, nvptx_mach_vector_length): New
1282         function.
1283         * config/nvptx/nvptx.h (struct machine_function): Add axis_dims.
1284
1285 2019-01-03  Tom de Vries  <tdevries@suse.de>
1286
1287         * config/nvptx/nvptx.c (struct offload_attrs): New.
1288         (populate_offload_attrs): New function.  Factor mask extraction out of
1289         nvptx_reorg.  Add extraction of dimensions.
1290         (nvptx_reorg): Use populate_offload_attrs.
1291
1292 2019-01-03  Tom de Vries  <tdevries@suse.de>
1293
1294         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Add early-out
1295         cases for oacc_min_dims_p and routine_p.  Add asserts for
1296         oacc_default_dims_p and offload_region_p.
1297
1298 2019-01-03  Tom de Vries  <tdevries@suse.de>
1299
1300         * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): New function,
1301         factored out of ...
1302         (nvptx_goacc_validate_dims): ... here.
1303
1304 2019-01-03  Jan Hubicka  <hubicka@ucw.cz>
1305
1306         PR tree-optimization/85574
1307         * tree-ssa-uncprop.c (struct equiv_hash_elt): Remove unused
1308         structure.
1309         (struct ssa_equip_hash_traits): Declare.
1310         (val_ssa_equiv): Use custom hash traits using operand_equal_p.
1311
1312 2019-01-03  Jakub Jelinek  <jakub@redhat.com>
1313
1314         PR debug/88644
1315         * dwarf2out.c (modified_type_die): If type is equal to sizetype,
1316         change it to qualified_type.
1317
1318 2019-01-03  Jan Hubicka  <hubicka@ucw.cz>
1319
1320         * ipa-utils.c (scale_ipa_profile_for_fn): Break out from ...
1321         (ipa_merge_profiles): ... here; do not ICE on thunks and aliases.
1322
1323 2019-01-02  Martin Sebor  <msebor@redhat.com>
1324             Jeff Law  <law@redhat.com>
1325
1326         * gimple-fold.c (strlen_range_kind): Remove SRK_LENRANGE_2.
1327         (get_range_strlen_tree): Update appropriately.
1328         (get_range_strlen)
1329         * gimple-fold.h (get_range_strlen): Drop unused last argument.
1330
1331         * gimple-fold.c (gimple_fold_builtin_strlen): Use set_strlen_range
1332         rather than set_range_info.
1333         * tree-ssa-strlen.c (set_strlen_range): Extracted from
1334         maybe_set_strlen_range.  Handle potentially boundary crossing
1335         cases more conservatively.
1336         (maybe_set_strlen_range): Parts refactored into set_strlen_range.
1337         Call set_strlen_range.
1338         * tree-ssa-strlen.h (set_strlen_range): Add prototype.
1339
1340         PR middle-end/88663
1341         * gimple-fold.c (get_range_strlen): Update prototype to no longer
1342         need the flexp argument.
1343         (get_range_strlen_tree): Drop flexp argument.  Drop flexp argument
1344         from calls to get_range_strlen.  Update comments.  Just update
1345         VAL for an unterminated const char array and let the reset of the
1346         code handle it normally.  No longer try to set *flexp.  Adjust
1347         return value.
1348         (get_range_strlen): Update for the new get_range_strlen API.
1349         (get_maxval_strlen): Similarly.
1350         (gimple_fold_builtin_strlen): Handle update meaning of return value
1351         from get_range_strlen.
1352         * gimple-ssa-sprintf.c (get_string_length): Update for the new
1353         get_range_strlen API.
1354
1355 2019-01-02  Jan Hubicka  <hubicka@ucw.cz>
1356
1357         PR lto/88130
1358         * varpool.c (varpool_node::ctor_useable_for_folding_p): Also return
1359         false at WPA time when body was removed.
1360
1361 2019-01-02  Martin Liska  <mliska@suse.cz>
1362
1363         PR tree-optimization/88650
1364         * predict.c (set_even_probabilities): Calculate probability
1365         remainer only when really used.
1366
1367 2019-01-02  Richard Biener  <rguenther@suse.de>
1368
1369         PR middle-end/88651
1370         * tree-data-ref.c (analyze_subscript_affine_affine): Use
1371         widest_ints when mangling max_stmt_execution results.
1372
1373 2019-01-02  Richard Biener  <rguenther@suse.de>
1374
1375         PR tree-optimization/88621
1376         * tree-ssa-loop-im.c (gather_mem_refs_stmt): Fix pastos, avoid
1377         bitfields when canoncalizing.
1378
1379 2019-01-02  Richard Biener  <rguenther@suse.de>
1380
1381         PR target/87545
1382         * config/i386/x86-tune-costs.h (intel_cost): Adjust
1383         cost of cheap SSE instruction.
1384
1385 2019-01-02  Richard Biener  <rguenther@suse.de>
1386
1387         PR ipa/85574
1388         * ipa-icf.h (sem_item_optimizer::sort_congruence_split): Declare.
1389         * ipa-icf.c (sem_item_optimizer::sort_congruence_split): New
1390         function.
1391         (sem_item_optimizer::do_congruence_step_f): Sort the congruence
1392         set after UIDs before splitting them.
1393
1394 2019-01-01  Martin Sebor  <msebor@redhat.com>
1395             Jeff Law  <law@redhat.com>
1396
1397         * gimple-fold.c (get_range_strlen_tree): Record if the computed
1398         length is optimistic.  If it is, then arrange to compute the
1399         conservative length as well.
1400
1401         * gimple-fold.h (get_range_strlen): Update prototype.
1402         * builtins.c (check_access): Update call to get_range_strlen to use
1403         c_strlen_data pointer.   Change various variable accesses to instead
1404         pull data from the c_strlen_data structure.
1405         (check_strncat_sizes, expand_builtin_strncat): Likewise.
1406         * calls.c (maybe_warn_nonstring_arg): Likewise.
1407         * tree-ssa-strlen.c (maybe_diag_stxncpy_trunc): Likewise.  Reset
1408         minimum length if maximum lengh is unknown.
1409         * gimple-ssa-sprintf.c (get_string_length): Likewise.  Drop code
1410         that used c_strlen, it's no longer needed.  Restructure slightly.
1411         (format_string): Set unlikely range appropriately.
1412         * gimple-fold.c (get_range_strlen): Update comments.  Fix minor
1413         formatting issues.
1414         (get_range_strlen):  Accept c_strlen_data pointer for external
1415         call sites as well.  Pass through to call to internal get_range_strlen.
1416         Adjust minlen, maxlen and maxbound as needed.
1417         (get_maxval_strlen): Update comments.
1418         (gimple_fold_builtin_strlen):  Update call to get_range_strlen
1419         to use c_strlen_data pointer.  Change variable accesses to instead
1420         use c_strlen_data data members.
1421
1422         * gimple-fold.c (get_range_strlen): Update prototype.
1423         (get_range_strlen_tree): Update prototype.  Drop minlen/maxlen
1424         local variables.  Use pdata to return information to caller.
1425         Update calls to get_range_strlen.  Update pdata->maxbound.
1426         (get_range_strlen -- static version): Similarly.
1427         (get_range_strlen -- extern version): Update for internal
1428         get_range_strlen API change.  Convert to external data format.
1429         (get_maxval_strlen): Similarly.
1430
1431 2019-01-01  Jan Hubicka  <hubicka@ucw.cz>
1432
1433         * coverage.c (get_coverage_counts): Use current_function_decl.
1434         * profile.c (read_thunk_profile): New function.
1435         (branch_prob): Add THUNK parameter.
1436         * tree-profile.c (tree_profiling): Handle thunks.
1437         * value-prof.c (init_node_map): Handle thunks.
1438         * value-prof.h (branch_prob): Upate prototype.
1439         (read_thunk_profile): Declare.
1440
1441 2019-01-01  Jakub Jelinek  <jakub@redhat.com>
1442
1443         Update copyright years.
1444
1445         * gcc.c (process_command): Update copyright notice dates.
1446         * gcov-dump.c (print_version): Ditto.
1447         * gcov.c (print_version): Ditto.
1448         * gcov-tool.c (print_version): Ditto.
1449         * gengtype.c (create_file): Ditto.
1450         * doc/cpp.texi: Bump @copying's copyright year.
1451         * doc/cppinternals.texi: Ditto.
1452         * doc/gcc.texi: Ditto.
1453         * doc/gccint.texi: Ditto.
1454         * doc/gcov.texi: Ditto.
1455         * doc/install.texi: Ditto.
1456         * doc/invoke.texi: Ditto.
1457 \f
1458 Copyright (C) 2019 Free Software Foundation, Inc.
1459
1460 Copying and distribution of this file, with or without modification,
1461 are permitted in any medium without royalty provided the copyright
1462 notice and this notice are preserved.