gcc/ChangeLog

   1 2014-02-21  Jakub Jelinek  <jakub@redhat.com>
   2
   3         PR tree-optimization/56490
   4         * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
   5         * tree-ssa-uninit.c: Include params.h.
   6         (compute_control_dep_chain): Add num_calls argument, return false
   7         if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
   8         num_calls to recursive call.
   9         (find_predicates): Change dep_chain into normal array,
  10         cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
  11         variable and adjust compute_control_dep_chain caller.
  12         (find_def_preds): Likewise.
  13
  14 2014-02-21  Thomas Schwinge  <thomas@codesourcery.com>
  15
  16         * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
  17         <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
  18
  19 2014-02-21  Nick Clifton  <nickc@redhat.com>
  20
  21         * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
  22         (pushhi1): Likewise.
  23         (popqi1): Add mode to pre_dec.
  24         (pophi1): Likewise.
  25
  26 2014-02-21  Jakub Jelinek  <jakub@redhat.com>
  27
  28         * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
  29         mode for mask of V8SFmode permutation.
  30
  31 2014-02-20  Richard Henderson  <rth@redhat.com>
  32
  33         PR c++/60272
  34         * builtins.c (expand_builtin_atomic_compare_exchange): Always make
  35         a new pseudo for OLDVAL.
  36
  37 2014-02-20  Jakub Jelinek  <jakub@redhat.com>
  38
  39         PR target/57896
  40         * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
  41         gen_reg_rtx if d->testing_p.
  42         (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
  43         if d->testing_p and we will certainly return true.
  44         (expand_vec_perm_even_odd_1): Likewise.  Don't call gen_reg_rtx
  45         if d->testing_p.
  46
  47 2014-02-20  Uros Bizjak  <ubizjak@gmail.com>
  48
  49         * emit-rtl.c (gen_reg_rtx): Assert that
  50         crtl->emit.regno_pointer_align_length is non-zero.
  51
  52 2014-02-20  Richard Henderson  <rth@redhat.com>
  53
  54         PR c++/60272
  55         * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
  56         on failure the store back into EXPECT.
  57
  58 2014-02-20  Chung-Lin Tang  <cltang@codesourcery.com>
  59             Sandra Loosemore  <sandra@codesourcery.com>
  60
  61         * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
  62         * config/nios2/nios2.c (nios2_function_profiler):
  63         Add -fPIC (flag_pic == 2) support.
  64         (nios2_handle_custom_fpu_cfg): Fix warning parameter.
  65         (nios2_large_offset_p): New function.
  66         (nios2_unspec_reloc_p): Move up position, update to use
  67         nios2_large_offset_p.
  68         (nios2_unspec_address): Remove function.
  69         (nios2_unspec_offset): New function.
  70         (nios2_large_got_address): New function.
  71         (nios2_got_address): Add large offset support.
  72         (nios2_legitimize_tls_address): Update usage of removed and new
  73         functions.
  74         (nios2_symbol_binds_local_p): New function.
  75         (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
  76         (nios2_legitimize_address): Update to use nios2_large_offset_p.
  77         (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
  78         (nios2_print_operand): Merge H/L processing, add hiadj/lo
  79         processing for (const (unspec ...)).
  80         (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
  81
  82 2014-02-20  Richard Biener  <rguenther@suse.de>
  83
  84         * tree-cfg.c (replace_uses_by): Mark altered BBs before
  85         doing the substitution.
  86         (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
  87
  88 2014-02-20  Martin Jambor  <mjambor@suse.cz>
  89
  90         PR ipa/55260
  91         * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
  92         info when checking whether lattices are bottom.
  93
  94 2014-02-20  Richard Biener  <rguenther@suse.de>
  95
  96         PR middle-end/60221
  97         * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
  98         regions at -O0.
  99
 100 2014-02-20  Jan Hubicka  <hubicka@ucw.cz>
 101
 102         PR ipa/58555
 103         * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
 104         parameter specifying the scaling.
 105         (inline_call): Update.
 106         (want_inline_recursively): Guard division by zero.
 107         (recursive_inlining): Update.
 108         * ipa-inline.h (clone_inlined_nodes): Update.
 109
 110 2014-02-20  Ilya Tocar  <ilya.tocar@intel.com>
 111
 112         PR target/60204
 113         * config/i386/i386.c (classify_argument): Pass structures of size
 114         64 bytes or less in register.
 115
 116 2014-02-20  Ilya Tocar  <ilya.tocar@intel.com>
 117             Kirill Yukhin  <kirill.yukhin@intel.com>
 118
 119         * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
 120         (_mm_rcp28_round_ss): Ditto.
 121         (_mm_rsqrt28_round_sd): Ditto.
 122         (_mm_rsqrt28_round_ss): Ditto.
 123         * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
 124         (_mm_rcp14_round_ss): Ditto.
 125         (_mm_rsqrt14_round_sd): Ditto.
 126         (_mm_rsqrt14_round_ss): Ditto.
 127         * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
 128         the first input operand, get rid of match_dup.
 129         (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
 130         attribute to sse.
 131         (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
 132         Ditto.
 133         (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
 134         operand as the first input operand, set type attribute.
 135         (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
 136         Set type attribute.
 137         (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
 138         operand as the first input operand, set type attribute.
 139
 140 2014-02-19  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
 141
 142         * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
 143         bit of zero.
 144
 145 2014-02-19  H.J. Lu  <hongjiu.lu@intel.com>
 146
 147         PR target/60207
 148         * config/i386/i386.c (construct_container): Remove TFmode check
 149         for X86_64_INTEGER_CLASS.
 150
 151 2014-02-19  Uros Bizjak  <ubizjak@gmail.com>
 152
 153         PR target/59794
 154         * config/i386/i386.c (type_natural_mode): Warn for ABI changes
 155         only when -Wpsabi is enabled.
 156
 157 2014-02-19  Michael Hudson-Doyle  <michael.hudson@linaro.org>
 158
 159          PR target/59799
 160         * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
 161         passing arrays in registers are the same as for structs, so remove the
 162         special case for them.
 163
 164 2014-02-19  Eric Botcazou  <ebotcazou@adacore.com>
 165
 166         * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
 167         destination type, extract only the valid bits if the source type is not
 168         integral and has a different mode.
 169
 170 2014-02-19  Richard Biener  <rguenther@suse.de>
 171
 172         PR ipa/60243
 173         * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
 174         for all calls.
 175
 176 2014-02-19  Richard Biener  <rguenther@suse.de>
 177
 178         PR ipa/60243
 179         * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
 180         (ipa_modify_call_arguments): Emit an argument load explicitely and
 181         preserve virtual SSA form there and for the replacement call.
 182         Do not update SSA form nor free dominance info.
 183
 184 2014-02-18  Jan Hubicka  <hubicka@ucw.cz>
 185
 186         * ipa.c (function_and_variable_visibility): Also clear WEAK
 187         flag when disolving COMDAT_GROUP.
 188
 189 2014-02-18  Jan Hubicka  <hubicka@ucw.cz>
 190
 191         * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
 192         * ipa-prop.c (ipa_set_jf_known_type): Return early when
 193         not devirtualizing.
 194         (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
 195         do more sanity checks.
 196         (detect_type_change): Return true when giving up early.
 197         (compute_complex_assign_jump_func): Fix type parameter of
 198         ipa_set_ancestor_jf.
 199         (compute_complex_ancestor_jump_func): Likewise.
 200         (update_jump_functions_after_inlining): Fix updating of
 201         ancestor function.
 202         * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
 203
 204 2014-02-18  Jan Hubicka  <hubicka@ucw.cz>
 205
 206         * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
 207         inline clones when edge disappears.
 208
 209 2014-02-18  Michael Meissner  <meissner@linux.vnet.ibm.com>
 210
 211         PR target/60203
 212         * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
 213         Split 64-bit moves into 2 patterns.  Do not allow the use of
 214         direct move for TDmode in little endian, since the decimal value
 215         has little endian bytes within a word, but the 64-bit pieces are
 216         ordered in a big endian fashion, and normal subreg's of TDmode are
 217         not allowed.
 218         (mov<mode>_64bit_dm): Likewise.
 219         (movtd_64bit_nodm): Likewise.
 220
 221 2014-02-18  Eric Botcazou  <ebotcazou@adacore.com>
 222
 223         PR tree-optimization/60174
 224         * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
 225         statement of an SSA_NAME that occurs in an abnormal PHI node.
 226
 227 2014-02-18  Jakub Jelinek  <jakub@redhat.com>
 228
 229         PR sanitizer/60142
 230         * final.c (SEEN_BB): Remove.
 231         (SEEN_NOTE, SEEN_EMITTED): Renumber.
 232         (final_scan_insn): Don't force_source_line on second
 233         NOTE_INSN_BASIC_BLOCK.
 234
 235 2014-02-18  Uros Bizjak  <ubizjak@gmail.com>
 236
 237         PR target/60205
 238         * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
 239         * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
 240         (type_natural_mode): Warn ABI change when %zmm register is not
 241         available for AVX512F vector value passing.
 242
 243 2014-02-18  Kai Tietz  <ktietz@redhat.com>
 244
 245         PR target/60193
 246         * config/i386/i386.c (ix86_expand_prologue): Use value in
 247         rax register as displacement when restoring %r10 or %rax.
 248         Fix wrong offset when restoring both registers.
 249
 250 2014-02-18  Eric Botcazou  <ebotcazou@adacore.com>
 251
 252         * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
 253         assertion with conditional return.
 254
 255 2014-02-18  Jakub Jelinek  <jakub@redhat.com>
 256             Uros Bizjak  <ubizjak@gmail.com>
 257
 258         PR driver/60233
 259         * config/i386/driver-i386.c (host_detect_local_cpu): If
 260         YMM state is not saved by the OS, also clear has_f16c.  Move
 261         CPUID 0x80000001 handling before YMM state saving checking.
 262
 263 2014-02-18  Andrey Belevantsev  <abel@ispras.ru>
 264
 265         PR rtl-optimization/58960
 266         * haifa-sched.c (alloc_global_sched_pressure_data): New,
 267         factored out from ...
 268         (sched_init): ... here.
 269         (free_global_sched_pressure_data): New, factored out from ...
 270         (sched_finish): ... here.
 271         * sched-int.h (free_global_sched_pressure_data): Declare.
 272         * sched-rgn.c (nr_regions_initial): New static global.
 273         (haifa_find_rgns): Initialize it.
 274         (schedule_region): Disable sched-pressure for the newly
 275         generated regions.
 276
 277 2014-02-17  Richard Biener  <rguenther@suse.de>
 278
 279         * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
 280         release SSA defs of pattern stmts.
 281
 282 2014-02-17  Richard Biener  <rguenther@suse.de>
 283
 284         * tree-inline.c (expand_call_inline): Release the virtual
 285         operand defined by the call we are about to inline.
 286
 287 2014-02-17  Richard Biener  <rguenther@suse.de>
 288
 289         * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
 290
 291 2014-02-17  Kirill Yukhin  <kirill.yukhin@intel.com>
 292             Ilya Tocar  <ilya.tocar@intel.com>
 293
 294         * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
 295         arguments order in builtin.
 296         (_mm512_permutexvar_epi64): Ditto.
 297         (_mm512_mask_permutexvar_epi64): Ditto
 298         (_mm512_maskz_permutexvar_epi32): Ditto
 299         (_mm512_permutexvar_epi32): Ditto
 300         (_mm512_mask_permutexvar_epi32): Ditto
 301
 302 2014-02-16  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
 303
 304         * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
 305         (p8_vmrgow): Likewise.
 306
 307 2014-02-16  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
 308
 309         * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
 310         endian targets.
 311
 312 2014-02-15  Michael Meissner  <meissner@linux.vnet.ibm.com>
 313
 314         PR target/60203
 315         * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
 316         (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
 317         into 64-bit and 32-bit moves.  On 64-bit moves, add support for
 318         using direct move instructions on ISA 2.07.  Also adjust
 319         instruction length for 64-bit.
 320         (mov<mode>_64bit, TFmode/TDmode): Likewise.
 321         (mov<mode>_32bit, TFmode/TDmode): Likewise.
 322
 323 2014-02-15  Alan Modra  <amodra@gmail.com>
 324
 325         PR target/58675
 326         PR target/57935
 327         * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
 328         find_replacement on parts of insn rtl that might be reloaded.
 329
 330 2014-02-15  Richard Biener  <rguenther@suse.de>
 331
 332         PR tree-optimization/60183
 333         * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
 334         (tree_ssa_phiprop): Calculate and free post-dominators.
 335
 336 2014-02-14  Jeff Law  <law@redhat.com>
 337
 338         PR rtl-optimization/60131
 339         * ree.c (get_extended_src_reg): New function.
 340         (combine_reaching_defs): Use it rather than assuming location of REG.
 341         (find_and_remove_re): Verify first operand of extension is
 342         a REG before adding the insns to the copy list.
 343
 344 2014-02-14  Roland McGrath  <mcgrathr@google.com>
 345
 346         * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
 347         * configure: Regenerated.
 348         * config.in: Regenerated.
 349         * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
 350         instead of ASM_SHORT.
 351
 352 2014-02-14  Vladimir Makarov  <vmakarov@redhat.com>
 353             Richard Earnshaw  <rearnsha@arm.com>
 354
 355         PR rtl-optimization/59535
 356         * lra-constraints.c (process_alt_operands): Encourage alternative
 357         when unassigned pseudo class is superset of the alternative class.
 358         (inherit_reload_reg): Don't inherit when optimizing for code size.
 359         * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
 360         returning CORE_REGS for anything but Thumb1 and BASE_REGS for
 361         modes not less than 4 for Thumb1.
 362
 363 2014-02-14  Kyle McMartin  <kyle@redhat.com>
 364
 365         PR pch/60010
 366         * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
 367
 368 2014-02-14  Richard Biener  <rguenther@suse.de>
 369
 370         * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
 371         (get_frame_arg): Drop the assert with langhook types_compatible_p.
 372         Do not strip INDIRECT_REFs.
 373
 374 2014-02-14  Richard Biener  <rguenther@suse.de>
 375
 376         PR lto/60179
 377         * lto-streamer-out.c (DFS_write_tree_body): Do not follow
 378         DECL_FUNCTION_SPECIFIC_TARGET.
 379         (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
 380         * tree-streamer-out.c (pack_ts_target_option): Remove.
 381         (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
 382         (write_ts_function_decl_tree_pointers): Do not stream
 383         DECL_FUNCTION_SPECIFIC_TARGET.
 384         * tree-streamer-in.c (unpack_ts_target_option): Remove.
 385         (unpack_value_fields): Do not stream TS_TARGET_OPTION.
 386         (lto_input_ts_function_decl_tree_pointers): Do not stream
 387         DECL_FUNCTION_SPECIFIC_TARGET.
 388
 389 2014-02-14  Jakub Jelinek  <jakub@redhat.com>
 390
 391         * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
 392         (get_initial_def_for_induction, vectorizable_induction): Ignore
 393         debug stmts when looking for exit_phi.
 394         (vectorizable_live_operation): Fix up condition.
 395
 396 2014-02-14  Chung-Ju Wu  <jasonwucj@gmail.com>
 397
 398         * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
 399         nreverse() because it changes the content of original tree list.
 400
 401 2014-02-14  Chung-Ju Wu  <jasonwucj@gmail.com>
 402
 403         * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
 404         * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
 405
 406 2014-02-14  Chung-Ju Wu  <jasonwucj@gmail.com>
 407
 408         * config/nds32/nds32.c (nds32_naked_function_p): Follow the
 409         GNU coding standards.
 410
 411 2014-02-13  Jakub Jelinek  <jakub@redhat.com>
 412
 413         PR debug/60152
 414         * dwarf2out.c (gen_subprogram_die): Don't call
 415         add_calling_convention_attribute if subr_die is old_die.
 416
 417 2014-02-13  Sharad Singhai  <singhai@google.com>
 418
 419         * doc/optinfo.texi: Fix order of nodes.
 420
 421 2014-02-13  Uros Bizjak  <ubizjak@gmail.com>
 422
 423         * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
 424         operands[2], not operands[3].
 425
 426 2014-02-13  Richard Biener  <rguenther@suse.de>
 427
 428         PR bootstrap/59878
 429         * doc/install.texi (ISL): Update recommended version to 0.12.2,
 430         mention the possibility of an in-tree build.
 431         (CLooG): Update recommended version to 0.18.1, mention the
 432         possibility of an in-tree build and clarify that the ISL
 433         bundled with CLooG does not work.
 434
 435 2014-02-13  Jakub Jelinek  <jakub@redhat.com>
 436
 437         PR target/43546
 438         * expr.c (compress_float_constant): If x is a hard register,
 439         extend into a pseudo and then move to x.
 440
 441 2014-02-13  Dominik Vogt  <vogt@linux.vnet.ibm.com>
 442
 443         * config/s390/s390.c (s390_asm_output_function_label): Fix crash
 444         caused by bad second argument to warning_at() with -mhotpatch and
 445         nested functions (e.g. with gfortran).
 446
 447 2014-02-13  Richard Sandiford  <rdsandiford@googlemail.com>
 448
 449         * opts.c (option_name): Remove "enabled by default" rider.
 450
 451 2014-02-12  John David Anglin  <danglin@gcc.gnu.org>
 452
 453         * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
 454
 455 2014-02-12  H.J. Lu  <hongjiu.lu@intel.com>
 456             Uros Bizjak  <ubizjak@gmail.com>
 457
 458         PR target/60151
 459         * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
 460         * configure: Regenerated.
 461
 462 2014-02-12  Richard Biener  <rguenther@suse.de>
 463
 464         * vec.c (vec_prefix::calculate_allocation): Move as
 465         inline variant to vec.h.
 466         (vec_prefix::calculate_allocation_1): New out-of-line version.
 467         * vec.h (vec_prefix::calculate_allocation_1): Declare.
 468         (vec_prefix::m_has_auto_buf): Rename to ...
 469         (vec_prefix::m_using_auto_storage): ... this.
 470         (vec_prefix::calculate_allocation): Inline the easy cases
 471         and dispatch to calculate_allocation_1 which doesn't need the
 472         prefix address.
 473         (va_heap::reserve): Use gcc_checking_assert.
 474         (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
 475         m_using_auto_storage.
 476         (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
 477         member and adjust.
 478         (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
 479         (vec<T, va_heap, vl_ptr>::release): Avoid casting.
 480         (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
 481
 482 2014-02-12  Richard Biener  <rguenther@suse.de>
 483
 484         * gcse.c (compute_transp): break from loop over canon_modify_mem_list
 485         when we found a dependence.
 486
 487 2014-02-12  Thomas Schwinge  <thomas@codesourcery.com>
 488
 489         * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
 490         common code...
 491         (maybe_fold_stmt): ... into this new function.
 492         * omp-low.c (lower_omp): Update comment.
 493
 494         * omp-low.c (lower_omp_target): Add clobber for sizes array, after
 495         last use.
 496
 497         * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
 498         dereference.
 499
 500 2014-02-12  James Greenhalgh  <james.greenhalgh@arm.com>
 501
 502         * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
 503         identifiers in comments.
 504         (cortexa53_extra_costs): Likewise.
 505         * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
 506         (cortexa7_extra_costs): Likewise.
 507         (cortexa12_extra_costs): Likewise.
 508         (cortexa15_extra_costs): Likewise.
 509         (v7m_extra_costs): Likewise.
 510
 511 2014-02-12  Richard Biener  <rguenther@suse.de>
 512
 513         PR middle-end/60092
 514         * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
 515         of posix_memalign being successful.
 516         (lower_stmt): Restrict lowering of posix_memalign to when
 517         -ftree-bit-ccp is enabled.
 518
 519 2014-02-12  Senthil Kumar Selvaraj  <senthil_kumar.selvaraj@atmel.com>
 520
 521         * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
 522         arg_loc.
 523         * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
 524
 525 2014-02-12  Eric Botcazou  <ebotcazou@adacore.com>
 526
 527         PR rtl-optimization/60116
 528         * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
 529         other_insn once the combination has been validated.
 530
 531 2014-02-11  Jan Hubicka  <hubicka@ucw.cz>
 532
 533         PR lto/59468
 534         * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
 535         and wrapper.
 536         * ipa-devirt.c: Include demangle.h
 537         (odr_violation_reported): New static variable.
 538         (add_type_duplicate): Update odr_violations.
 539         (maybe_record_node): Add completep parameter; update it.
 540         (record_target_from_binfo): Add COMPLETEP parameter;
 541         update it as needed.
 542         (possible_polymorphic_call_targets_1): Likewise.
 543         (struct polymorphic_call_target_d): Add nonconstruction_targets;
 544         rename FINAL to COMPLETE.
 545         (record_targets_from_bases): Sanity check we found the binfo;
 546         fix COMPLETEP updating.
 547         (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
 548         parameter, fix computing of COMPLETEP.
 549         (dump_possible_polymorphic_call_targets): Imrove readability of dump;
 550         at LTO time do demangling.
 551         (ipa_devirt): Use nonconstruction_targets; Improve dumps.
 552         * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
 553         parameter.
 554         (gimple_get_virt_method_for_binfo): Likewise.
 555         * gimple-fold.h (gimple_get_virt_method_for_binfo,
 556         gimple_get_virt_method_for_vtable): Update prototypes.
 557
 558 2014-02-11  Vladimir Makarov  <vmakarov@redhat.com>
 559
 560         PR target/49008
 561         * genautomata.c (add_presence_absence): Fix typo with
 562         {final_}presence_list.
 563
 564 2014-02-11  Michael Meissner  <meissner@linux.vnet.ibm.com>
 565
 566         PR target/60137
 567         * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
 568         for VSX/Altivec vectors that land in GPR registers.
 569
 570 2014-02-11  Richard Henderson  <rth@redhat.com>
 571             Jakub Jelinek  <jakub@redhat.com>
 572
 573         PR debug/59776
 574         * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
 575         around drhs if type conversion to lacc->type is not useless.
 576
 577 2014-02-11  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 578
 579         * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
 580         tuning struct.
 581         (cortex-a57.cortex-a53): Likewise.
 582         * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
 583
 584 2014-02-11  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 585
 586         * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
 587         arm_restrict_it.
 588
 589 2014-02-11  Renlin Li  <Renlin.Li@arm.com>
 590
 591         * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
 592         add_options_for_arm_vfp3.
 593
 594 2014-02-11  Jeff Law  <law@redhat.com>
 595
 596         PR middle-end/54041
 597         * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
 598         object with an undesirable mode.
 599
 600 2014-02-11  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 601
 602         PR libgomp/60107
 603         * config/i386/sol2-9.h: New file.
 604         * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
 605         *-*-solaris2.9*): Use it.
 606
 607 2014-02-10  Nagaraju Mekala <nagaraju.mekala@xilinx.com>
 608
 609         * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
 610         * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
 611
 612 2014-02-10  Nagaraju Mekala <nagaraju.mekala@xilinx.com>
 613
 614         * config/microblaze/microblaze.c: Extend mcpu version format
 615
 616 2014-02-10  David Holsgrove <david.holsgrove@xilinx.com>
 617
 618         * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
 619
 620 2014-02-10  Richard Henderson  <rth@redhat.com>
 621
 622         PR target/59927
 623         * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
 624         * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
 625         ms-abi vs -mno-accumulate-outgoing-args.
 626         (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
 627         * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
 628         respect to ms-abi.
 629
 630 2014-02-10  Bernd Edlinger  <bernd.edlinger@hotmail.de>
 631
 632         PR middle-end/60080
 633         * cfgexpand.c (expand_asm_operands): Attach source location to
 634         ASM_INPUT rtx objects.
 635         * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
 636
 637 2014-02-10  Nick Clifton  <nickc@redhat.com>
 638
 639         * config/mn10300/mn10300.c (popcount): New function.
 640         (mn10300_expand_prologue): Include saved registers in stack usage
 641         count.
 642
 643 2014-02-10  Jeff Law  <law@redhat.com>
 644
 645         PR middle-end/52306
 646         * reload1.c (emit_input_reload_insns): Do not create invalid RTL
 647         when changing the SET_DEST of a prior insn to avoid an input reload.
 648
 649 2014-02-10  Ulrich Weigand  <Ulrich.Weigand@de.ibm.com>
 650
 651         * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
 652         big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
 653         -mcall-openbsd, or -mcall-linux.
 654         (CC1_ENDIAN_BIG_SPEC): Remove.
 655         (CC1_ENDIAN_LITTLE_SPEC): Remove.
 656         (CC1_ENDIAN_DEFAULT_SPEC): Remove.
 657         (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
 658         (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
 659         and %cc1_endian_default.
 660         * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
 661
 662 2014-02-10  Richard Biener  <rguenther@suse.de>
 663
 664         PR tree-optimization/60115
 665         * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
 666         MEM_REF handling.  Properly verify that the accesses are not
 667         out of the objects bound.
 668
 669 2014-02-10  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 670
 671         * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
 672         coretex to cortex.
 673
 674 2014-02-10  Eric Botcazou  <ebotcazou@adacore.com>
 675
 676         * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
 677         proper constants and fix formatting.
 678         (possible_polymorphic_call_targets): Fix formatting.
 679
 680 2014-02-10  Kirill Yukhin  <kirill.yukhin@intel.com>
 681             Ilya Tocar  <ilya.tocar@intel.com>
 682
 683         * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
 684         (_mm512_loadu_epi32): Renamed into...
 685         (_mm512_loadu_si512): This.
 686         (_mm512_storeu_epi32): Renamed into...
 687         (_mm512_storeu_si512): This.
 688         (_mm512_maskz_ceil_ps): Removed.
 689         (_mm512_maskz_ceil_pd): Ditto.
 690         (_mm512_maskz_floor_ps): Ditto.
 691         (_mm512_maskz_floor_pd): Ditto.
 692         (_mm512_floor_round_ps): Ditto.
 693         (_mm512_floor_round_pd): Ditto.
 694         (_mm512_ceil_round_ps): Ditto.
 695         (_mm512_ceil_round_pd): Ditto.
 696         (_mm512_mask_floor_round_ps): Ditto.
 697         (_mm512_mask_floor_round_pd): Ditto.
 698         (_mm512_mask_ceil_round_ps): Ditto.
 699         (_mm512_mask_ceil_round_pd): Ditto.
 700         (_mm512_maskz_floor_round_ps): Ditto.
 701         (_mm512_maskz_floor_round_pd): Ditto.
 702         (_mm512_maskz_ceil_round_ps): Ditto.
 703         (_mm512_maskz_ceil_round_pd): Ditto.
 704         (_mm512_expand_pd): Ditto.
 705         (_mm512_expand_ps): Ditto.
 706         * config/i386/i386.c (ix86_builtins): Remove
 707         IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
 708         (bdesc_args): Ditto.
 709         * config/i386/predicates.md (const1256_operand): New.
 710         (const_1_to_2_operand): Ditto.
 711         * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
 712         (*avx512pf_gatherpf<mode>sf_mask): Ditto.
 713         (*avx512pf_gatherpf<mode>sf): Ditto.
 714         (avx512pf_gatherpf<mode>df): Ditto.
 715         (*avx512pf_gatherpf<mode>df_mask): Ditto.
 716         (*avx512pf_gatherpf<mode>df): Ditto.
 717         (avx512pf_scatterpf<mode>sf): Ditto.
 718         (*avx512pf_scatterpf<mode>sf_mask): Ditto.
 719         (*avx512pf_scatterpf<mode>sf): Ditto.
 720         (avx512pf_scatterpf<mode>df): Ditto.
 721         (*avx512pf_scatterpf<mode>df_mask): Ditto.
 722         (*avx512pf_scatterpf<mode>df): Ditto.
 723         (avx512f_expand<mode>): Removed.
 724         (<shift_insn><mode>3<mask_name>): Change predicate type.
 725
 726 2014-02-08  Jakub Jelinek  <jakub@redhat.com>
 727
 728         * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
 729         not at the end of datarefs vector use ordered_remove to avoid
 730         reordering datarefs vector.
 731
 732         PR c/59984
 733         * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
 734         mark local addressable non-static vars as GOVD_PRIVATE
 735         instead of GOVD_LOCAL.
 736         * omp-low.c (lower_omp_for): Move gimple_bind_vars
 737         and BLOCK_VARS of gimple_bind_block to new_stmt rather
 738         than copying them.
 739
 740         PR middle-end/60092
 741         * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
 742         if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
 743         assume_aligned or alloc_align attributes.
 744         (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
 745         arguments.  Handle also assume_aligned and alloc_align attributes.
 746         (evaluate_stmt): Adjust bit_value_assume_aligned caller.  Handle
 747         calls to functions with assume_aligned or alloc_align attributes.
 748         * doc/extend.texi: Document assume_aligned and alloc_align attributes.
 749
 750 2014-02-08  Terry Guo  <terry.guo@arm.com>
 751
 752         * doc/invoke.texi: Document ARM -march=armv7e-m.
 753
 754 2014-02-08  Jakub Jelinek  <jakub@redhat.com>
 755
 756         * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
 757         flag on __cilkrts_rethrow builtin.
 758
 759         PR ipa/60026
 760         * ipa-cp.c (determine_versionability): Fail at -O0
 761         or __attribute__((optimize (0))) or -fno-ipa-cp functions.
 762         * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
 763
 764         Revert:
 765         2014-02-04  Jakub Jelinek  <jakub@redhat.com>
 766
 767         PR ipa/60026
 768         * tree-inline.c (copy_forbidden): Fail for
 769         __attribute__((optimize (0))) functions.
 770
 771 2014-02-07  Jan Hubicka  <hubicka@ucw.cz>
 772
 773         * varpool.c: Include pointer-set.h.
 774         (varpool_remove_unreferenced_decls): Variables in other partitions
 775         will not be output; be however careful to not lose information
 776         about partitioning.
 777
 778 2014-02-07  Jan Hubicka  <hubicka@ucw.cz>
 779
 780         * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
 781         lookup in the vtable constructor.
 782
 783 2014-02-07  Jeff Law  <law@redhat.com>
 784
 785         PR target/40977
 786         * config/m68k/m68k.md (ashldi_extsi): Turn into a
 787         define_insn_and_split.
 788
 789         * ipa-inline.c (inline_small_functions): Fix typos.
 790
 791 2014-02-07  Richard Sandiford  <rsandifo@linux.vnet.ibm.com>
 792
 793         * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
 794         (s390_can_use_return_insn): Declare.
 795         * config/s390/s390.h (EPILOGUE_USES): Define.
 796         * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
 797         instructions.
 798         (s390_chunkify_start): Handle return JUMP_LABELs.
 799         (s390_early_mach): Emit a main_pool instruction on the entry edge.
 800         (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
 801         (s390_can_use_return_insn): New functions.
 802         (s390_fix_long_loop_prediction): Handle conditional returns.
 803         (TARGET_SET_UP_BY_PROLOGUE): Define.
 804         * config/s390/s390.md (ANY_RETURN): New code iterator.
 805         (*creturn, *csimple_return, return, simple_return): New patterns.
 806
 807 2014-02-07  Richard Sandiford  <rsandifo@linux.vnet.ibm.com>
 808
 809         * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
 810         notes to each restore.  Also add REG_CFA_DEF_CFA when restoring %r15.
 811         (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P.  Update the
 812         REG_CFA_RESTORE list when deciding not to restore a register.
 813
 814 2014-02-07  Richard Sandiford  <rsandifo@linux.vnet.ibm.com>
 815
 816         * config/s390/s390.c: Include tree-pass.h and context.h.
 817         (s390_early_mach): New function, split out from...
 818         (s390_emit_prologue): ...here.
 819         (pass_data_s390_early_mach): New pass structure.
 820         (pass_s390_early_mach): New class.
 821         (s390_option_override): Create and register early_mach pass.
 822         Move to end of file.
 823
 824 2014-02-07  Richard Sandiford  <rsandifo@linux.vnet.ibm.com>
 825
 826         * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
 827         to match for the exit block.
 828
 829 2014-02-07  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>
 830
 831         * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
 832         ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
 833         Reject misaligned operands.
 834
 835 2014-02-07  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>
 836
 837         * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
 838
 839 2014-02-07  Richard Biener  <rguenther@suse.de>
 840
 841         PR middle-end/60092
 842         * gimple-low.c (lower_builtin_posix_memalign): New function.
 843         (lower_stmt): Call it to lower posix_memalign in a way
 844         to make alignment info accessible.
 845
 846 2014-02-07  Jakub Jelinek  <jakub@redhat.com>
 847
 848         PR c++/60082
 849         * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
 850         __builtin_setjmp_receiver.
 851
 852 2014-02-07  Richard Biener  <rguenther@suse.de>
 853
 854         PR middle-end/60092
 855         * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
 856         * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
 857         * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
 858         Handle BUILT_IN_POSIX_MEMALIGN.
 859         (find_func_clobbers): Likewise.
 860         * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
 861         (call_may_clobber_ref_p_1): Likewise.
 862
 863 2014-02-06  Jan Hubicka  <hubicka@ucw.cz>
 864
 865         PR ipa/59918
 866         * ipa-devirt.c (record_target_from_binfo): Remove overactive
 867         sanity check.
 868
 869 2014-02-06  Jan Hubicka  <hubicka@ucw.cz>
 870
 871         PR ipa/59469
 872         * lto-cgraph.c (lto_output_node): Use
 873         symtab_get_symbol_partitioning_class.
 874         (lto_output_varpool_node): likewise.
 875         (symtab_get_symbol_partitioning_class): Move here from
 876         lto/lto-partition.c
 877         * cgraph.h (symbol_partitioning_class): Likewise.
 878         (symtab_get_symbol_partitioning_class): Declare.
 879
 880 2014-02-06  Jan Hubicka  <hubicka@ucw.cz>
 881
 882         * ggc.h (ggc_internal_cleared_alloc): New macro.
 883         * vec.h (vec_safe_copy): Handle memory stats.
 884         * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
 885         * target-globals.c (save_target_globals): Likewise.
 886
 887 2014-02-06  Jan Hubicka  <hubicka@ucw.cz>
 888
 889         PR target/60077
 890         * expr.c (emit_move_resolve_push): Export; be bit more selective
 891         on when to clear alias set.
 892         * expr.h (emit_move_resolve_push): Declare.
 893         * function.h (struct function): Add tail_call_marked.
 894         * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
 895         * config/i386/i386-protos.h (ix86_expand_push): Remove.
 896         * config/i386/i386.md (TImode move expander): De not call
 897         ix86_expand_push.
 898         (FP push expanders): Preserve memory attributes.
 899         * config/i386/sse.md (push<mode>1): Remove.
 900         * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
 901         (ix86_expand_push): Remove.
 902         * config/i386/mmx.md (push<mode>1): Remove.
 903
 904 2014-02-06  Jakub Jelinek  <jakub@redhat.com>
 905
 906         PR rtl-optimization/60030
 907         * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
 908         lopart with paradoxical subreg before shifting it up by hprec.
 909
 910 2014-02-06  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 911
 912         * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
 913         Remove extra newline at end of file.
 914         * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
 915         (arm_issue_rate): Handle cortexa57.
 916         * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
 917         (cortex-a57.cortex-a53): Likewise.
 918
 919 2014-02-06  Jakub Jelinek  <jakub@redhat.com>
 920
 921         PR target/59575
 922         * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
 923         don't record in REG_FRAME_RELATED_EXPR registers not set in that
 924         bitmask.
 925         (arm_expand_prologue): Adjust all callers.
 926         (arm_unwind_emit_sequence): Allow saved, but not important for unwind
 927         info, registers also at the lowest numbered registers side.  Use
 928         gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
 929         XEXP.
 930
 931         PR debug/59992
 932         * var-tracking.c (adjust_mems): Before adding a SET to
 933         amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
 934
 935 2014-02-06  Alan Modra  <amodra@gmail.com>
 936
 937         PR target/60032
 938         * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
 939         change SDmode to DDmode when lra_in_progress.
 940
 941 2014-02-06  Jakub Jelinek  <jakub@redhat.com>
 942
 943         PR middle-end/59150
 944         * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
 945         free_data_ref on the dr first, and before goto again also set dr
 946         to the next dr.  For simd_lane_access, free old datarefs[i] before
 947         overwriting it.  For get_vectype_for_scalar_type failure, don't
 948         free_data_ref if simd_lane_access.
 949
 950         * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
 951
 952         PR target/60062
 953         * tree.h (opts_for_fn): New inline function.
 954         (opt_for_fn): Define.
 955         * config/i386/i386.c (ix86_function_regparm): Use
 956         opt_for_fn (decl, optimize) instead of optimize.
 957
 958 2014-02-06  Marcus Shawcroft  <marcus.shawcroft@arm.com>
 959
 960         * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
 961         for SYMBOL_REF in large memory model.
 962
 963 2014-02-06  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 964
 965         * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
 966         and crypto support.
 967         (cortex-a57): Likewise.
 968         (cortex-a57.cortex-a53): Likewise.
 969
 970 2014-02-06  Yury Gribov  <y.gribov@samsung.com>
 971             Kugan Vivekanandarajah  <kuganv@linaro.org>
 972
 973         * config/arm/arm.c (arm_vector_alignment_reachable): Check
 974         unaligned_access.
 975         * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
 976
 977 2014-02-06  Richard Biener  <rguenther@suse.de>
 978
 979         * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
 980         set_loop_copy and initialize_original_copy_tables.
 981
 982 2014-02-06  Alex Velenko  <Alex.Velenko@arm.com>
 983
 984         * config/aarch64/aarch64-simd.md
 985         (aarch64_ashr_simddi): Change QI to SI.
 986
 987 2014-02-05  Jan Hubicka  <hubicka@ucw.cz>
 988             Jakub Jelinek  <jakub@redhat.com>
 989
 990         PR middle-end/60013
 991         * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
 992         of the dataflow.
 993
 994 2014-02-05  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
 995
 996         * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
 997         CODE_FOR_altivec_vpku[hw]um to
 998         CODE_FOR_altivec_vpku[hw]um_direct.
 999         * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
1000         UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
1001         (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
1002         UNSPEC_VUNPACK_LO_SIGN_DIRECT.
1003
1004 2014-02-05  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1005
1006         * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
1007         generation for -maltivec=be.
1008         (altivec_vsumsws): Simplify redundant test.
1009
1010 2014-02-05  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1011
1012         * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
1013         (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
1014         (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
1015         (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
1016         gen_altivec_vpkuwum.
1017         (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
1018         BYTES_BIG_ENDIAN.
1019         (altivec_vpks<VI_char>ss): Likewise.
1020         (altivec_vpks<VI_char>us): Likewise.
1021         (altivec_vpku<VI_char>us): Likewise.
1022         (altivec_vpku<VI_char>um): Likewise.
1023         (altivec_vpku<VI_char>um_direct): New (copy of
1024         altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
1025         internal use).
1026         (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
1027         target is little endian and -maltivec=be is not specified.
1028         (*altivec_vupkhs<VU_char>_direct): New (copy of
1029         altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
1030         (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
1031         target is little endian and -maltivec=be is not specified.
1032         (*altivec_vupkls<VU_char>_direct): New (copy of
1033         altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
1034         (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
1035         little endian and -maltivec=be is not specified.
1036         (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
1037         little endian and -maltivec=be is not specified.
1038
1039 2014-02-05  Richard Henderson  <rth@redhat.com>
1040
1041         PR debug/52727
1042         * combine-stack-adj.c: Revert r206943.
1043         * sched-int.h (struct deps_desc): Add last_args_size.
1044         * sched-deps.c (init_deps): Initialize it.
1045         (sched_analyze_insn): Add OUTPUT dependencies between insns that
1046         contain REG_ARGS_SIZE notes.
1047
1048 2014-02-05  Jan Hubicka  <hubicka@ucw.cz>
1049
1050         * lto-cgraph.c (asm_nodes_output): Make global.
1051         * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
1052         * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
1053         (driver_handle_option): Handle OPT_fwpa.
1054
1055 2014-02-05  Jakub Jelinek  <jakub@redhat.com>
1056
1057         PR ipa/59947
1058         * ipa-devirt.c (possible_polymorphic_call_targets): Fix
1059         a comment typo and formatting issue.  If odr_hash hasn't been
1060         created, return vNULL and set *completep to false.
1061
1062         PR middle-end/57499
1063         * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
1064         bb with no successors.
1065
1066 2014-02-05  James Greenhalgh  <james.greenhalgh@arm.com>
1067
1068         PR target/59718
1069         * doc/invoke.texi (-march): Clarify documentation for ARM.
1070         (-mtune): Likewise.
1071         (-mcpu): Likewise.
1072
1073 2014-02-05  Richard Biener  <rguenther@suse.de>
1074
1075         * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
1076         when not vectorizing because of too many alias checks.
1077         * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
1078         Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
1079
1080 2014-02-05  Nick Clifton  <nickc@redhat.com>
1081
1082         * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
1083         accept extended registers in any mode when compiling for the MN10300.
1084
1085 2014-02-05  Yury Gribov  <y.gribov@samsung.com>
1086
1087         * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
1088         * ipa-inline.c (report_inline_failed_reason): Handle mismatched
1089         sanitization attributes.
1090         (can_inline_edge_p): Likewise.
1091         (sanitize_attrs_match_for_inline_p): New function.
1092
1093 2014-02-04  Jan Hubicka  <hubicka@ucw.cz>
1094
1095         * ipa-prop.c (detect_type_change): Shor circuit testing of
1096         type changes on THIS pointer.
1097
1098 2014-02-04  John David Anglin  <danglin@gcc.gnu.org>
1099
1100         PR target/59777
1101         * config/pa/pa.c (legitimize_tls_address): Return original address
1102         if not passed a SYMBOL_REF rtx.
1103         (hppa_legitimize_address): Call legitimize_tls_address for all TLS
1104         addresses.
1105         (pa_emit_move_sequence): Simplify TLS source operands.
1106         (pa_legitimate_constant_p): Reject all TLS constants.
1107         * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
1108         (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
1109
1110 2014-02-04  Jan Hubicka  <hubicka@ucw.cz>
1111
1112         * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
1113         groups when we know they are controlled by LTO.
1114         * varasm.c (default_binds_local_p_1): If object is in other partition,
1115         it will be resolved locally.
1116
1117 2014-02-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>
1118
1119         * config/host-linux.c (linux_gt_pch_use_address): Don't
1120         use SSIZE_MAX because it is not always defined.
1121
1122 2014-02-04  Vladimir Makarov  <vmakarov@redhat.com>
1123
1124         PR bootstrap/59913
1125         * lra-constraints.c (need_for_split_p): Use more 3 reloads as
1126         threshold for pseudo splitting.
1127         (update_ebb_live_info): Process call argument hard registers and
1128         hard registers from insn definition too.
1129         (max_small_class_regs_num): New constant.
1130         (inherit_in_ebb): Update live hard regs through EBBs.  Update
1131         reloads_num only for small register classes.  Don't split for
1132         outputs of jumps.
1133
1134 2014-02-04  Markus Trippelsdorf  <markus@trippelsdorf.de>
1135
1136         PR ipa/60058
1137         * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
1138         is non-null.
1139
1140 2014-02-04  Jan Hubicka  <hubicka@ucw.cz>
1141
1142         * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
1143         visibility is safe.
1144
1145 2014-02-04  Marek Polacek  <polacek@redhat.com>
1146
1147         * gdbinit.in (pel): Define.
1148
1149 2014-02-04  Bernd Edlinger  <bernd.edlinger@hotmail.de>
1150
1151         * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
1152         behavior.
1153
1154 2014-02-04  Richard Biener  <rguenther@suse.de>
1155
1156         PR lto/59723
1157         * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
1158         in function context local.
1159         (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
1160         * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
1161         similar to LTO_imported_decl_ref.
1162
1163 2014-02-04  Jakub Jelinek  <jakub@redhat.com>
1164
1165         PR tree-optimization/60002
1166         * cgraphclones.c (build_function_decl_skip_args): Clear
1167         DECL_LANG_SPECIFIC.
1168
1169         PR tree-optimization/60023
1170         * tree-if-conv.c (predicate_mem_writes): Pass true instead of
1171         false to gsi_replace.
1172         * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
1173         has been in some EH region and vec_stmt could throw, add
1174         vec_stmt into the same EH region.
1175         * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
1176         has no lhs, ignore it.
1177         * internal-fn.c (expand_MASK_LOAD): Likewise.
1178
1179         PR ipa/60026
1180         * tree-inline.c (copy_forbidden): Fail for
1181         __attribute__((optimize (0))) functions.
1182
1183         PR other/58712
1184         * omp-low.c (simd_clone_struct_copy): If from->inbranch
1185         is set, copy one less argument.
1186         (expand_simd_clones): Don't subtract clone_info->inbranch
1187         from simd_clone_struct_alloc argument.
1188
1189         PR rtl-optimization/57915
1190         * recog.c (simplify_while_replacing): If all unary/binary/relational
1191         operation arguments are constant, attempt to simplify those.
1192
1193         PR middle-end/59261
1194         * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
1195         if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
1196
1197 2014-02-04  Richard Biener  <rguenther@suse.de>
1198
1199         PR tree-optimization/60012
1200         * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
1201         TBAA disambiguation to all DDRs.
1202
1203 2014-02-04  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
1204
1205         PR target/59788
1206         * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
1207         (LINK_SPEC): Use it for -shared, -shared-libgcc.
1208
1209 2014-02-03  Jan Hubicka  <hubicka@ucw.cz>
1210
1211         PR ipa/59882
1212         * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
1213
1214 2014-02-03  Jan Hubicka  <hubicka@ucw.cz>
1215
1216         * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
1217         * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
1218
1219 2014-02-03  Jan Hubicka  <hubicka@ucw.cz>
1220
1221         PR ipa/59831
1222         * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
1223         to figure out targets of polymorphic calls with known decl.
1224         * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
1225         * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
1226         * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
1227         (get_polymorphic_call_info): ... here.
1228         (get_polymorphic_call_info_from_invariant): New function.
1229
1230 2014-02-03  Jan Hubicka  <hubicka@ucw.cz>
1231
1232         * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
1233         lookup via vtable pointer; check for type consistency
1234         and turn inconsitent facts into UNREACHABLE.
1235         * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
1236         * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
1237         type inconsistent querries; return UNREACHABLE instead.
1238
1239 2014-02-03  Richard Henderson  <rth@twiddle.net>
1240
1241         PR tree-opt/59924
1242         * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
1243         already processed this node.
1244         (normalize_one_pred_1): Pass along mark_set.
1245         (normalize_one_pred): Create and destroy a pointer_set_t.
1246         (normalize_one_pred_chain): Likewise.
1247
1248 2014-02-03  Laurent Aflonsi  <laurent.alfonsi@st.com>
1249
1250         PR gcov-profile/58602
1251         * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
1252
1253 2014-02-03  Jan Hubicka  <hubicka@ucw.cz>
1254
1255         PR ipa/59831
1256         * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
1257         -fno-devirtualize; try to devirtualize by the knowledge of
1258         virtual table pointer given by aggregate propagation.
1259         * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
1260         (ipa_print_node_jump_functions): Dump also offset that
1261         is relevant for polymorphic calls.
1262         (determine_known_aggregate_parts): Add arg_type parameter; use it
1263         instead of determining the type from pointer type.
1264         (ipa_compute_jump_functions_for_edge): Update call of
1265         determine_known_aggregate_parts.
1266         * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
1267         (gimple_get_virt_method_for_binfo): ... here; simplify using
1268         vtable_pointer_value_to_vtable.
1269         * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
1270         * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
1271         to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
1272         (vtable_pointer_value_to_vtable): Break out from ...; handle also
1273         POINTER_PLUS_EXPR.
1274         (vtable_pointer_value_to_binfo): ... here.
1275         * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
1276
1277 2014-02-03  Teresa Johnson  <tejohnson@google.com>
1278
1279         * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
1280         redef of outer loop index variable.
1281
1282 2014-02-03  Marc Glisse  <marc.glisse@inria.fr>
1283
1284         PR c++/53017
1285         PR c++/59211
1286         * doc/extend.texi (Function Attributes): Typo.
1287
1288 2014-02-03  Cong Hou  <congh@google.com>
1289
1290         PR tree-optimization/60000
1291         * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
1292         if the vectorized statement is a store.  A store statement can only
1293         appear at the end of pattern statements.
1294
1295 2014-02-03  H.J. Lu  <hongjiu.lu@intel.com>
1296
1297         * config/i386/i386.c (flag_opts): Add -mlong-double-128.
1298         (ix86_option_override_internal): Default long double to 64-bit for
1299         32-bit Bionic and to 128-bit for 64-bit Bionic.
1300
1301         * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
1302         TARGET_LONG_DOUBLE_128 is true.
1303         (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
1304
1305         * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
1306         (mlong-double-64): Negate -mlong-double-128.
1307         (mlong-double-128): New option.
1308
1309         * config/i386/i386-c.c (ix86_target_macros): Define
1310         __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
1311
1312         * doc/invoke.texi: Document -mlong-double-128.
1313
1314 2014-02-03  H.J. Lu  <hongjiu.lu@intel.com>
1315
1316         PR rtl-optimization/60024
1317         * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
1318
1319 2014-02-03  Markus Trippelsdorf  <markus@trippelsdorf.de>
1320
1321         * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
1322
1323 2014-02-03  Andrey Belevantsev  <abel@ispras.ru>
1324
1325         PR rtl-optimization/57662
1326         * sel-sched.c (code_motion_path_driver): Do not mark already not
1327         existing blocks in the visiting bitmap.
1328
1329 2014-02-03  Andrey Belevantsev  <abel@ispras.ru>
1330
1331         * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
1332         on the insn being emitted.
1333
1334 2014-02-03  James Greenhalgh  <james.greenhalgh@arm.com>
1335             Will Deacon  <will.deacon@arm.com>
1336
1337         * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
1338
1339 2014-02-03  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
1340
1341         * config/arm/arm-tables.opt: Regenerate.
1342
1343 2014-02-02  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1344
1345         * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
1346         for vector types other than V16QImode.
1347         * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
1348         define_expand, and call altivec_expand_vec_perm_le when producing
1349         code with little endian element order.
1350         (*altivec_vperm_<mode>_internal): New insn having previous
1351         behavior of altivec_vperm_<mode>.
1352         (altivec_vperm_<mode>_uns): Change to a define_expand, and call
1353         altivec_expand_vec_perm_le when producing code with little endian
1354         element order.
1355         (*altivec_vperm_<mode>_uns_internal): New insn having previous
1356         behavior of altivec_vperm_<mode>_uns.
1357
1358 2014-02-02  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1359
1360         * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
1361         (altivec_vsumsws): Add handling for -maltivec=be with a little
1362         endian target.
1363         (altivec_vsumsws_direct): New.
1364         (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
1365         gen_altivec_vsumsws.
1366
1367 2014-02-02  Jan Hubicka  <hubicka@ucw.cz>
1368
1369         * ipa-devirt.c (subbinfo_with_vtable_at_offset,
1370         vtable_pointer_value_to_binfo): New functions.
1371         * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
1372         * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
1373
1374 2014-02-02  Sandra Loosemore  <sandra@codesourcery.com>
1375
1376         * config/nios2/nios2.md (load_got_register): Initialize GOT
1377         pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
1378         * config/nios2/nios2.c (nios2_function_profiler): Likewise.
1379
1380 2014-02-02  Jan Hubicka  <hubicka@ucw.cz>
1381
1382         * ipa-prop.c (update_jump_functions_after_inlining): When type is not
1383         preserverd by passthrough, do not propagate the type.
1384
1385 2014-02-02  Richard Sandiford  <rdsandiford@googlemail.com>
1386
1387         * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
1388         (mips_atomic_assign_expand_fenv): New function.
1389         (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
1390
1391 2014-02-02  Richard Sandiford  <rdsandiford@googlemail.com>
1392
1393         * doc/extend.texi (__builtin_mips_get_fcsr): Document.
1394         (__builtin_mips_set_fcsr): Likewise.
1395         * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
1396         MIPS_USI_FTYPE_VOID.
1397         * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
1398         (mips16_expand_set_fcsr): Likewise.
1399         * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
1400         (mips16_set_fcsr_stub): Likewise.
1401         (mips16_get_fcsr_one_only_stub): New class.
1402         (mips16_set_fcsr_one_only_stub): Likewise.
1403         (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
1404         (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
1405         (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
1406         (hard_float): New availability predicate.
1407         (mips_builtins): Add get_fcsr and set_fcsr.
1408         (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
1409         * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
1410         (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
1411         (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
1412         (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
1413         patterns.
1414
1415 2014-02-02  Richard Sandiford  <rdsandiford@googlemail.com>
1416
1417         * config/mips/mips.c (mips_one_only_stub): New class.
1418         (mips_need_mips16_rdhwr_p): Replace with...
1419         (mips16_rdhwr_stub): ...this new variable.
1420         (mips16_stub_call_address): New function.
1421         (mips16_rdhwr_one_only_stub): New class.
1422         (mips_expand_thread_pointer): Use mips16_stub_call_address.
1423         (mips_output_mips16_rdhwr): Delete.
1424         (mips_finish_stub): New function.
1425         (mips_code_end): Use it to handle rdhwr stubs.
1426
1427 2014-02-02  Uros Bizjak  <ubizjak@gmail.com>
1428
1429         PR target/60017
1430         * config/i386/i386.c (classify_argument): Fix handling of bit_offset
1431         when calculating size of integer atomic types.
1432
1433 2014-02-02  H.J. Lu  <hongjiu.lu@intel.com>
1434
1435         * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
1436
1437 2014-02-01  Jakub Jelinek  <jakub@redhat.com>
1438
1439         PR tree-optimization/60003
1440         * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
1441         * profile.c (branch_prob): Use gimple_call_builtin_p
1442         to check for BUILT_IN_SETJMP_RECEIVER.
1443         * tree-inline.c (copy_bb): Call notice_special_calls.
1444
1445 2014-01-31  Vladimir Makarov  <vmakarov@redhat.com>
1446
1447         PR bootstrap/59985
1448         * lra-constraints.c (process_alt_operands): Update reload_sum only
1449         on the first pass.
1450
1451 2014-01-31  Richard Henderson  <rth@redhat.com>
1452
1453         PR middle-end/60004
1454         * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
1455         until after else_eh is processed.
1456
1457 2014-01-31  Ilya Tocar  <ilya.tocar@intel.com>
1458
1459         * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
1460         (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
1461         (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
1462         in smmintrin.h, remove them.
1463         (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
1464         * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
1465         * config/i386/i386.md (ROUND_SAE): Fix value.
1466         * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
1467         (const48_operand): New.
1468         * config/i386/subst.md (round), (round_expand): Use
1469         const_4_or_8_to_11_operand.
1470         (round_saeonly), (round_saeonly_expand): Use const48_operand.
1471
1472 2014-01-31  Ilya Tocar  <ilya.tocar@intel.com>
1473
1474         * config/i386/constraints.md (Yk): Swap meaning with k.
1475         * config/i386/i386.md (movhi_internal): Change Yk to k.
1476         (movqi_internal): Ditto.
1477         (*k<logic><mode>): Ditto.
1478         (*andhi_1): Ditto.
1479         (*andqi_1): Ditto.
1480         (kandn<mode>): Ditto.
1481         (*<code>hi_1): Ditto.
1482         (*<code>qi_1): Ditto.
1483         (kxnor<mode>): Ditto.
1484         (kortestzhi): Ditto.
1485         (kortestchi): Ditto.
1486         (kunpckhi): Ditto.
1487         (*one_cmplhi2_1): Ditto.
1488         (*one_cmplqi2_1): Ditto.
1489         * config/i386/sse.md (): Change k to Yk.
1490         (avx512f_load<mode>_mask): Ditto.
1491         (avx512f_blendm<mode>): Ditto.
1492         (avx512f_store<mode>_mask): Ditto.
1493         (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1494         (avx512f_storedqu<mode>_mask): Ditto.
1495         (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
1496         Ditto.
1497         (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
1498         (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
1499         (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
1500         (avx512f_maskcmp<mode>3): Ditto.
1501         (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
1502         (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
1503         (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
1504         (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
1505         (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
1506         (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
1507         (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
1508         (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
1509         (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
1510         (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
1511         (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
1512         (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
1513         (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1514         (vec_extract_lo_<mode>_maskm): Ditto.
1515         (vec_extract_hi_<mode>_maskm): Ditto.
1516         (avx512f_vternlog<mode>_mask): Ditto.
1517         (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
1518         (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
1519         (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1520         (avx512f_<code>v8div16qi2_mask): Ditto.
1521         (avx512f_<code>v8div16qi2_mask_store): Ditto.
1522         (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
1523         (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
1524         (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
1525         (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
1526         (*avx512pf_gatherpf<mode>sf_mask): Ditto.
1527         (*avx512pf_gatherpf<mode>df_mask): Ditto.
1528         (*avx512pf_scatterpf<mode>sf_mask): Ditto.
1529         (*avx512pf_scatterpf<mode>df_mask): Ditto.
1530         (avx512cd_maskb_vec_dupv8di): Ditto.
1531         (avx512cd_maskw_vec_dupv16si): Ditto.
1532         (avx512f_vpermi2var<mode>3_maskz): Ditto.
1533         (avx512f_vpermi2var<mode>3_mask): Ditto.
1534         (avx512f_vpermi2var<mode>3_mask): Ditto.
1535         (avx512f_vpermt2var<mode>3_maskz): Ditto.
1536         (*avx512f_gathersi<mode>): Ditto.
1537         (*avx512f_gathersi<mode>_2): Ditto.
1538         (*avx512f_gatherdi<mode>): Ditto.
1539         (*avx512f_gatherdi<mode>_2): Ditto.
1540         (*avx512f_scattersi<mode>): Ditto.
1541         (*avx512f_scatterdi<mode>): Ditto.
1542         (avx512f_compress<mode>_mask): Ditto.
1543         (avx512f_compressstore<mode>_mask): Ditto.
1544         (avx512f_expand<mode>_mask): Ditto.
1545         * config/i386/subst.md (mask): Change k to Yk.
1546         (mask_scalar_merge): Ditto.
1547         (sd): Ditto.
1548
1549 2014-01-31  Marc Glisse  <marc.glisse@inria.fr>
1550
1551         * doc/extend.texi (Vector Extensions): Document ?: in C++.
1552
1553 2014-01-31  Richard Biener  <rguenther@suse.de>
1554
1555         PR middle-end/59990
1556         * builtins.c (fold_builtin_memory_op): Make sure to not
1557         use a floating-point mode or a boolean or enumeral type for
1558         the copy operation.
1559
1560 2014-01-30  DJ Delorie  <dj@redhat.com>
1561
1562         * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
1563         * config/msp430/msp430.md (msp430_refsym_need_exit): New.
1564         * config/msp430/msp430.c (msp430_expand_epilogue): Call it
1565         whenever main() has an epilogue.
1566
1567 2014-01-30  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1568
1569         * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
1570         unused variable "field".
1571         * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
1572         (vsx_mergeh_<mode>): Likewise.
1573         * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
1574         (altivec_vmrghh): Likewise.
1575         (altivec_vmrghw): Likewise.
1576         (altivec_vmrglb): Likewise.
1577         (altivec_vmrglh): Likewise.
1578         (altivec_vmrglw): Likewise.
1579         (altivec_vspltb): Add missing uses.
1580         (altivec_vsplth): Likewise.
1581         (altivec_vspltw): Likewise.
1582         (altivec_vspltsf): Likewise.
1583
1584 2014-01-30  Jakub Jelinek  <jakub@redhat.com>
1585
1586         PR target/59923
1587         * ifcvt.c (cond_exec_process_insns): Don't conditionalize
1588         frame related instructions.
1589
1590 2014-01-30  Vladimir Makarov  <vmakarov@redhat.com>
1591
1592         PR rtl-optimization/59959
1593         * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
1594         any reload of register whose subreg is invalid.
1595
1596 2014-01-30  Jakub Jelinek  <jakub@redhat.com>
1597
1598         * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
1599         * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
1600         Add missing return type - void.
1601
1602 2014-01-30  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1603
1604         * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
1605         gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
1606         remove element index adjustment for endian (now handled in vsx.md
1607         and altivec.md).
1608         (altivec_expand_vec_perm_const): Use
1609         gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
1610         * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
1611         (vsx_xxspltw_<mode>): Adjust element index for little endian.
1612         * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
1613         define_expand and a new define_insn *altivec_vspltb_internal;
1614         adjust for -maltivec=be on a little endian target.
1615         (altivec_vspltb_direct): New.
1616         (altivec_vsplth): Divide into a define_expand and a new
1617         define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
1618         little endian target.
1619         (altivec_vsplth_direct): New.
1620         (altivec_vspltw): Divide into a define_expand and a new
1621         define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
1622         little endian target.
1623         (altivec_vspltw_direct): New.
1624         (altivec_vspltsf): Divide into a define_expand and a new
1625         define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
1626         a little endian target.
1627
1628 2014-01-30  Richard Biener  <rguenther@suse.de>
1629
1630         PR tree-optimization/59993
1631         * tree-ssa-forwprop.c (associate_pointerplus): Check we
1632         can propagate form the earlier stmt and avoid the transform
1633         when the intermediate result is needed.
1634
1635 2014-01-30  Alangi Derick  <alangiderick@gmail.com>
1636
1637         * README.Portability: Fix typo.
1638
1639 2014-01-30  David Holsgrove <david.holsgrove@xilinx.com>
1640
1641         * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
1642         comparison_operator with ordered_comparison_operator.
1643
1644 2014-01-30  Nick Clifton  <nickc@redhat.com>
1645
1646         * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
1647         Rename to mn10300_store_multiple_regs.
1648         * config/mn10300/mn10300.c: Likewise.
1649         * config/mn10300/mn10300.md (store_movm): Fix typo: call
1650         store_multiple_regs.
1651         * config/mn10300/predicates.md (mn10300_store_multiple_operation):
1652         Call mn10300_store_multiple_regs.
1653
1654 2014-01-30  Nick Clifton  <nickc@redhat.com>
1655             DJ Delorie  <dj@redhat.com>
1656
1657         * config/rl78/rl78.c (register_sizes): Make the "upper half" of
1658         %fp 2 to keep registers after it properly word-aligned.
1659         (rl78_alloc_physical_registers_umul): Handle the case where both
1660         input operands are the same.
1661
1662 2014-01-30  Richard Biener  <rguenther@suse.de>
1663
1664         PR tree-optimization/59903
1665         * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
1666         check properly.
1667
1668 2014-01-30  Jason Merrill  <jason@redhat.com>
1669
1670         PR c++/59633
1671         * tree.c (walk_type_fields): Handle VECTOR_TYPE.
1672
1673         PR c++/59645
1674         * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
1675
1676 2014-01-30  Richard Biener  <rguenther@suse.de>
1677
1678         PR tree-optimization/59951
1679         * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
1680
1681 2014-01-30  Savin Zlobec  <savin.zlobec@gmail.com>
1682
1683         PR target/59784
1684         * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
1685         SFmode to DFmode case.
1686
1687 2014-01-29  DJ Delorie  <dj@redhat.com>
1688
1689         * config/msp430/msp430.opt (-minrt): New.
1690         * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
1691         if -minrt given.
1692         (ENDFILE_SPEC): Likewise.
1693
1694 2014-01-29  Jan Hubicka  <hubicka@ucw.cz>
1695
1696         * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
1697         (estimate_function_body_sizes): Use it.
1698
1699 2014-01-29  Paolo Carlini  <paolo.carlini@oracle.com>
1700
1701         PR c++/58561
1702         * dwarf2out.c (is_cxx_auto): New.
1703         (is_base_type): Use it.
1704         (gen_type_die_with_usage): Likewise.
1705
1706 2014-01-29  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
1707
1708         * config/rs6000/rs6000.c (altivec_expand_vec_perm_const):  Use
1709         CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
1710         * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
1711         -maltivec=be with LE targets.
1712         (vsx_mergeh_<mode>): Likewise.
1713         * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
1714         (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
1715         (altivec_vmrghb): Replace with define_expand and new
1716         *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
1717         (altivec_vmrghb_direct): New define_insn.
1718         (altivec_vmrghh): Replace with define_expand and new
1719         *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
1720         (altivec_vmrghh_direct): New define_insn.
1721         (altivec_vmrghw): Replace with define_expand and new
1722         *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
1723         (altivec_vmrghw_direct): New define_insn.
1724         (*altivec_vmrghsf): Adjust for endianness.
1725         (altivec_vmrglb): Replace with define_expand and new
1726         *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
1727         (altivec_vmrglb_direct): New define_insn.
1728         (altivec_vmrglh): Replace with define_expand and new
1729         *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
1730         (altivec_vmrglh_direct): New define_insn.
1731         (altivec_vmrglw): Replace with define_expand and new
1732         *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
1733         (altivec_vmrglw_direct): New define_insn.
1734         (*altivec_vmrglsf): Adjust for endianness.
1735         (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1736         (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1737         (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1738         (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1739         (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1740         (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1741         (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1742         (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1743
1744 2014-01-29  Marcus Shawcroft  <marcus.shawcroft@arm.com>
1745
1746         * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
1747         (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
1748         whitespace.
1749
1750 2014-01-29  Richard Biener  <rguenther@suse.de>
1751
1752         PR tree-optimization/58742
1753         * tree-ssa-forwprop.c (associate_pointerplus): Rename to
1754         associate_pointerplus_align.
1755         (associate_pointerplus_diff): New function.
1756         (associate_pointerplus): Likewise.  Call associate_pointerplus_align
1757         and associate_pointerplus_diff.
1758
1759 2014-01-29  Richard Biener  <rguenther@suse.de>
1760
1761         * lto-streamer.h (LTO_major_version): Bump to 3.
1762         (LTO_minor_version): Reset to 0.
1763
1764 2014-01-29  Renlin Li  <Renlin.Li@arm.com>
1765
1766         * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
1767         * config/arm/arm.c (FL_FOR_ARCH7VE): New.
1768         (arm_file_start): Generate correct asm header for armv7ve.
1769         * config/arm/bpabi.h: Add multilib support for armv7ve.
1770         * config/arm/driver-arm.c: Change the architectures of cortex-a7
1771         and cortex-a15 to armv7ve.
1772         * config/arm/t-aprofile: Add multilib support for armv7ve.
1773         * doc/invoke.texi: Document -march=armv7ve.
1774
1775 2014-01-29  Richard Biener  <rguenther@suse.de>
1776
1777         PR tree-optimization/58742
1778         * tree-ssa-forwprop.c (associate_plusminus): Return true
1779         if we changed sth, defer EH cleanup to ...
1780         (ssa_forward_propagate_and_combine): ... here.  Call simplify_mult.
1781         (simplify_mult): New function.
1782
1783 2014-01-29  Jakub Jelinek  <jakub@redhat.com>
1784
1785         PR middle-end/59917
1786         PR tree-optimization/59920
1787         * tree.c (build_common_builtin_nodes): Remove
1788         __builtin_setjmp_dispatcher initialization.
1789         * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
1790         * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
1791         instead of gsi_after_labels + manually skipping debug stmts.
1792         Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
1793         ignore bbs with IFN_ABNORMAL_DISPATCHER.
1794         * tree-inline.c (copy_edges_for_bb): Remove
1795         can_make_abnormal_goto argument, instead add abnormal_goto_dest
1796         argument.  Ignore computed_goto_p stmts.  Don't call
1797         make_abnormal_goto_edges.  If a call might need abnormal edges
1798         for non-local gotos, see if it already has an edge to
1799         IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
1800         with true argument, don't do anything then, otherwise add
1801         EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
1802         (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
1803         caller.
1804         * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
1805         (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
1806         (lower_stmt): Don't set data->calls_builtin_setjmp.
1807         (lower_builtin_setjmp): Adjust comment.
1808         * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
1809         * tree-cfg.c (found_computed_goto): Remove.
1810         (factor_computed_gotos): Remove.
1811         (make_goto_expr_edges): Return bool, true for computed gotos.
1812         Don't call make_abnormal_goto_edges.
1813         (build_gimple_cfg): Don't set found_computed_goto, don't call
1814         factor_computed_gotos.
1815         (computed_goto_p): No longer static.
1816         (make_blocks): Don't set found_computed_goto.
1817         (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
1818         (make_edges): If make_goto_expr_edges returns true, push bb
1819         into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
1820         instead of calling make_abnormal_goto_edges push bb into ab_edge_call
1821         vector.  Record mapping between bbs and OpenMP regions if there
1822         are any, adjust make_gimple_omp_edges caller.  Call
1823         handle_abnormal_edges.
1824         (make_abnormal_goto_edges): Remove.
1825         * tree-cfg.h (make_abnormal_goto_edges): Remove.
1826         (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
1827         * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
1828         * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
1829         * internal-fn.def (ABNORMAL_DISPATCHER): New.
1830         * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
1831         filling *region also set *region_idx to (*region)->entry->index.
1832
1833         PR other/58712
1834         * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
1835         For REGs set ORIGINAL_REGNO.
1836
1837 2014-01-29  Bingfeng Mei  <bmei@broadcom.com>
1838
1839         * doc/md.texi: Mention that a target shouldn't implement
1840         vec_widen_(s|u)mul_even/odd pair if it is less efficient
1841         than hi/lo pair.
1842
1843 2014-01-29  Jakub Jelinek  <jakub@redhat.com>
1844
1845         PR tree-optimization/59594
1846         * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
1847         a copy of the datarefs vector rather than the vector itself.
1848
1849 2014-01-28  Jason Merrill  <jason@redhat.com>
1850
1851         PR c++/53756
1852         * dwarf2out.c (auto_die): New static.
1853         (gen_type_die_with_usage): Handle C++1y 'auto'.
1854         (gen_subprogram_die): If in-class DIE had 'auto', emit type again
1855         on definition.
1856
1857 2014-01-28  H.J. Lu  <hongjiu.lu@intel.com>
1858
1859         PR target/59672
1860         * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
1861         (SPEC_X32): Likewise.
1862         (SPEC_64): Likewise.
1863         * config/i386/i386.c (ix86_option_override_internal): Turn off
1864         OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
1865         for TARGET_16BIT.
1866         (x86_file_start): Output .code16gcc for TARGET_16BIT.
1867         * config/i386/i386.h (TARGET_16BIT): New macro.
1868         (TARGET_16BIT_P): Likewise.
1869         * config/i386/i386.opt: Add m16.
1870         * doc/invoke.texi: Document -m16.
1871
1872 2014-01-28  Jakub Jelinek  <jakub@redhat.com>
1873
1874         PR preprocessor/59935
1875         * input.c (location_get_source_line): Bail out on when line number
1876         is zero, and test the return value of lookup_or_add_file_to_cache_tab.
1877
1878 2014-01-28  Richard Biener  <rguenther@suse.de>
1879
1880         PR tree-optimization/58742
1881         * tree-ssa-forwprop.c (associate_plusminus): Handle
1882         pointer subtraction of the form (T)(P + A) - (T)P.
1883
1884 2014-01-28  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
1885
1886         * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
1887         at const_int_cost.
1888
1889 2014-01-28  Richard Biener  <rguenther@suse.de>
1890
1891         Revert
1892         2014-01-28  Richard Biener  <rguenther@suse.de>
1893
1894         PR rtl-optimization/45364
1895         PR rtl-optimization/59890
1896         * var-tracking.c (local_get_addr_clear_given_value): Handle
1897         already cleared slot.
1898         (val_reset): Handle not allocated local_get_addr_cache.
1899         (vt_find_locations): Use post-order on the inverted CFG.
1900
1901 2014-01-28  Richard Biener  <rguenther@suse.de>
1902
1903         * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
1904
1905 2014-01-28  Richard Biener  <rguenther@suse.de>
1906
1907         PR rtl-optimization/45364
1908         PR rtl-optimization/59890
1909         * var-tracking.c (local_get_addr_clear_given_value): Handle
1910         already cleared slot.
1911         (val_reset): Handle not allocated local_get_addr_cache.
1912         (vt_find_locations): Use post-order on the inverted CFG.
1913
1914 2014-01-28  Alan Modra  <amodra@gmail.com>
1915
1916         * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
1917         * configure.ac <recursive call for build != host>: Define
1918         GENERATOR_FILE.  Comment.  Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
1919         and LD_FOR_BUILD too.
1920         * configure: Regenerate.
1921
1922 2014-01-27  Allan Sandfeld Jensen  <sandfeld@kde.org>
1923
1924         * config/i386/i386.c (get_builtin_code_for_version): Separate
1925         Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
1926         Broadwell from Haswell.
1927
1928 2014-01-27  Steve Ellcey  <sellcey@mips.com>
1929
1930         * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
1931         Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
1932         * config/mips/mips.c (mips_option_override): Change setting
1933         of TARGET_DSP.
1934         * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
1935         * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
1936         Change from Mask to Var.
1937
1938 2014-01-27  Jeff Law  <law@redhat.com>
1939
1940         * ipa-inline.c (inline_small_functions): Fix typo.
1941
1942 2014-01-27  Ilya Tocar  <ilya.tocar@intel.com>
1943
1944         * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
1945         (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
1946         (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
1947         (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
1948         (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
1949         (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
1950         (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
1951         (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
1952         (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
1953         (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
1954         (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
1955         (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
1956         (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
1957         (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
1958         (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
1959         (_mm512_storeu_epi64): Ditto.
1960         (_mm512_cmpge_epi32_mask): Ditto.
1961         (_mm512_cmpge_epu32_mask): Ditto.
1962         (_mm512_cmpge_epi64_mask): Ditto.
1963         (_mm512_cmpge_epu64_mask): Ditto.
1964         (_mm512_cmple_epi32_mask): Ditto.
1965         (_mm512_cmple_epu32_mask): Ditto.
1966         (_mm512_cmple_epi64_mask): Ditto.
1967         (_mm512_cmple_epu64_mask): Ditto.
1968         (_mm512_cmplt_epi32_mask): Ditto.
1969         (_mm512_cmplt_epu32_mask): Ditto.
1970         (_mm512_cmplt_epi64_mask): Ditto.
1971         (_mm512_cmplt_epu64_mask): Ditto.
1972         (_mm512_cmpneq_epi32_mask): Ditto.
1973         (_mm512_cmpneq_epu32_mask): Ditto.
1974         (_mm512_cmpneq_epi64_mask): Ditto.
1975         (_mm512_cmpneq_epu64_mask): Ditto.
1976         (_mm512_expand_pd): Ditto.
1977         (_mm512_expand_ps): Ditto.
1978         * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
1979         VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
1980         VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
1981         * config/i386/i386.c (ix86_builtins): Add
1982         IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
1983         IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
1984         IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
1985         IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
1986         IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
1987         IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
1988         IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
1989         IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
1990         IX86_BUILTIN_PMOVUSQW512_MEM.
1991         (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
1992         __builtin_ia32_pmovsqd512mem_mask,
1993         __builtin_ia32_pmovqd512mem_mask,
1994         __builtin_ia32_pmovusqw512mem_mask,
1995         __builtin_ia32_pmovsqw512mem_mask,
1996         __builtin_ia32_pmovqw512mem_mask,
1997         __builtin_ia32_pmovusdw512mem_mask,
1998         __builtin_ia32_pmovsdw512mem_mask,
1999         __builtin_ia32_pmovdw512mem_mask,
2000         __builtin_ia32_pmovqb512mem_mask,
2001         __builtin_ia32_pmovusqb512mem_mask,
2002         __builtin_ia32_pmovsqb512mem_mask,
2003         __builtin_ia32_pmovusdb512mem_mask,
2004         __builtin_ia32_pmovsdb512mem_mask,
2005         __builtin_ia32_pmovdb512mem_mask.
2006         (bdesc_args): Add __builtin_ia32_expanddf512,
2007         __builtin_ia32_expandsf512.
2008         (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
2009         VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
2010         VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
2011         * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
2012         (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
2013         (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
2014         (avx512f_<code>v8div16qi2_mask_store): This.
2015         (avx512f_expand<mode>): New.
2016
2017 2014-01-27  Kirill Yukhin  <kirill.yukhin@intel.com>
2018
2019         * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
2020         New.
2021         (_mm512_mask_prefetch_i64gather_pd): Ditto.
2022         (_mm512_prefetch_i32scatter_pd): Ditto.
2023         (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2024         (_mm512_prefetch_i64scatter_pd): Ditto.
2025         (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2026         (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
2027         (_mm512_mask_prefetch_i64gather_ps): Ditto.
2028         (_mm512_prefetch_i32scatter_ps): Ditto.
2029         (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2030         (_mm512_prefetch_i64scatter_ps): Ditto.
2031         (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2032         * config/i386/i386-builtin-types.def: Define
2033         VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
2034         and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
2035         * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
2036         IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
2037         IX86_BUILTIN_SCATTERPFQPD.
2038         (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
2039         __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
2040         __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
2041         __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
2042         __builtin_ia32_scatterpfqps.
2043         (ix86_expand_builtin): Expand new built-ins.
2044         * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
2045         fix memory access data type.
2046         (*avx512pf_gatherpf<mode>_mask): Ditto.
2047         (*avx512pf_gatherpf<mode>): Ditto.
2048         (avx512pf_scatterpf<mode>): Ditto.
2049         (*avx512pf_scatterpf<mode>_mask): Ditto.
2050         (*avx512pf_scatterpf<mode>): Ditto.
2051         (GATHER_SCATTER_SF_MEM_MODE): New.
2052         (avx512pf_gatherpf<mode>df): Ditto.
2053         (*avx512pf_gatherpf<mode>df_mask): Ditto.
2054         (*avx512pf_scatterpf<mode>df): Ditto.
2055
2056 2014-01-27  Jakub Jelinek  <jakub@redhat.com>
2057
2058         PR bootstrap/59934
2059         * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
2060         know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
2061         reached.
2062
2063 2014-01-27  James Greenhalgh  <james.greenhalgh@arm.com>
2064
2065         * common/config/arm/arm-common.c
2066         (arm_rewrite_mcpu): Handle multiple names.
2067         * config/arm/arm.h
2068         (BIG_LITTLE_SPEC): Do not discard mcpu switches.
2069
2070 2014-01-27  James Greenhalgh  <james.greenhalgh@arm.com>
2071
2072         * gimple-builder.h (create_gimple_tmp): Delete.
2073
2074 2014-01-27  Christian Bruel  <christian.bruel@st.com>
2075
2076         * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
2077         words comparisons.
2078
2079 2014-01-26  John David Anglin  <danglin@gcc.gnu.org>
2080
2081         * config/pa/pa.md (call): Generate indirect long calls to non-local
2082         functions when outputing 32-bit code.
2083         (call_value): Likewise except for special call to buggy powf function.
2084
2085         * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
2086         portable runtime and PIC indirect calls.
2087         (pa_output_indirect_call): Remove unnecessary nop from portable runtime
2088         and PIC call sequences.  Use ldo instead of blr to set return register
2089         in PIC call sequence.
2090
2091 2014-01-25  Walter Lee  <walt@tilera.com>
2092
2093         * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
2094         avoid clobbering a live register.
2095
2096 2014-01-25  Walter Lee  <walt@tilera.com>
2097
2098         * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
2099         Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
2100         * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
2101         Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
2102
2103 2014-01-25  Walter Lee  <walt@tilera.com>
2104
2105         * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
2106         arguments on even registers.
2107         (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
2108         STACK_BOUNDARY.
2109         * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
2110         (BIGGEST_ALIGNMENT): Ditto.
2111         (BIGGEST_FIELD_ALIGNMENT): Ditto.
2112
2113 2014-01-25  Walter Lee  <walt@tilera.com>
2114
2115         * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
2116         insns before bundling.
2117         * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
2118
2119 2014-01-25  Walter Lee  <walt@tilera.com>
2120
2121         * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
2122         PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
2123         * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
2124
2125 2014-01-25  Richard Sandiford  <rdsandiford@googlemail.com>
2126
2127         * config/mips/constraints.md (kl): Delete.
2128         * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
2129         define expands, using...
2130         (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
2131         instructions for MIPS16.
2132         (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
2133         non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
2134
2135 2014-01-25  Walter Lee  <walt@tilera.com>
2136
2137         * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
2138         (clzdi2): Ditto.
2139         (ffsdi2): Ditto.
2140
2141 2014-01-25  Walter Lee  <walt@tilera.com>
2142
2143         * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
2144         (TARGET_EXPAND_TO_RTL_HOOK): Define.
2145
2146 2014-01-25  Richard Sandiford  <rdsandiford@googlemail.com>
2147
2148         * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
2149         Handle XOR.
2150
2151 2014-01-25  Jakub Jelinek  <jakub@redhat.com>
2152
2153         * print-rtl.c (in_call_function_usage): New var.
2154         (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
2155         EXPR_LIST mode as mode and not as reg note name.
2156
2157         PR middle-end/59561
2158         * cfgloopmanip.c (copy_loop_info): If
2159         loop->warned_aggressive_loop_optimizations, make sure
2160         the flag is set in target loop too.
2161
2162 2014-01-24  Balaji V. Iyer  <balaji.v.iyer@intel.com>
2163
2164         * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
2165         flag_cilkplus.
2166         * builtins.def: Likewise.
2167         * cilk.h (fn_contains_cilk_spawn_p): Likewise.
2168         * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
2169         * ira.c (ira_setup_eliminable_regset): Likewise.
2170         * omp-low.c (gate_expand_omp): Likewise.
2171         (execute_lower_omp): Likewise.
2172         (diagnose_sb_0): Likewise.
2173         (gate_diagnose_omp_blocks): Likewise.
2174         (simd_clone_clauses_extract): Likewise.
2175         (gate): Likewise.
2176
2177 2014-01-24  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
2178
2179         * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
2180         correction for little endian...
2181         * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
2182         here.
2183
2184 2014-01-24  Jeff Law  <law@redhat.com>
2185
2186         PR tree-optimization/59919
2187         * tree-vrp.c (find_assert_locations_1): Do not register asserts
2188         for non-returning calls.
2189
2190 2014-01-24  James Greenhalgh  <james.greenhalgh@arm.com>
2191
2192         * common/config/aarch64/aarch64-common.c
2193         (aarch64_rewrite_mcpu): Handle multiple names.
2194         * config/aarch64/aarch64.h
2195         (BIG_LITTLE_SPEC): Do not discard mcpu switches.
2196
2197 2014-01-24  Dodji Seketeli  <dodji@redhat.com>
2198
2199         * input.c (add_file_to_cache_tab): Handle the case where fopen
2200         returns NULL.
2201
2202 2014-01-23  H.J. Lu  <hongjiu.lu@intel.com>
2203
2204         PR target/59929
2205         * config/i386/i386.md (pushsf splitter): Get stack adjustment
2206         from push operand if code of push isn't PRE_DEC.
2207
2208 2014-01-23  Michael Meissner  <meissner@linux.vnet.ibm.com>
2209
2210         PR target/59909
2211         * doc/invoke.texi (RS/6000 and PowerPC Options): Document
2212         -mquad-memory-atomic.  Update -mquad-memory documentation to say
2213         it is only used for non-atomic loads/stores.
2214
2215         * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
2216         -mquad-memory or -mquad-memory-atomic switches.
2217
2218         * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
2219         -mquad-memory-atomic to ISA 2.07 support.
2220
2221         * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
2222         to separate support of normal quad word memory operations (ldq, stq)
2223         from the atomic quad word memory operations.
2224
2225         * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2226         support to separate non-atomic quad word operations from atomic
2227         quad word operations.  Disable non-atomic quad word operations in
2228         little endian mode so that we don't have to swap words after the
2229         load and before the store.
2230         (quad_load_store_p): Add comment about atomic quad word support.
2231         (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
2232         options printed with -mdebug=reg.
2233
2234         * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
2235         -mquad-memory-atomic as the test for whether we have quad word
2236         atomic instructions.
2237         (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
2238         or -mp8-vector are used, allow byte/half-word atomic operations.
2239
2240         * config/rs6000/sync.md (load_lockedti): Insure that the address
2241         is a proper indexed or indirect address for the lqarx instruction.
2242         On little endian systems, swap the hi/lo registers after the lqarx
2243         instruction.
2244         (load_lockedpti): Use indexed_or_indirect_operand predicate to
2245         insure the address is valid for the lqarx instruction.
2246         (store_conditionalti): Insure that the address is a proper indexed
2247         or indirect address for the stqcrx. instruction.  On little endian
2248         systems, swap the hi/lo registers before doing the stqcrx.
2249         instruction.
2250         (store_conditionalpti): Use indexed_or_indirect_operand predicate to
2251         insure the address is valid for the stqcrx. instruction.
2252
2253         * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
2254         Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
2255         type of quad memory support is available.
2256
2257 2014-01-23  Vladimir Makarov  <vmakarov@redhat.com>
2258
2259         PR regression/59915
2260         * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
2261         there is a danger of looping.
2262
2263 2014-01-23  Pat Haugen  <pthaugen@us.ibm.com>
2264
2265         * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
2266         force flag_ira_loop_pressure if set via command line.
2267
2268 2014-01-23  Alex Velenko  <Alex.Velenko@arm.com>
2269
2270         * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
2271         (ashr_simd): New builtin handling DI mode.
2272         * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
2273         (aarch64_sshr_simddi): New match pattern.
2274         * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
2275         (vshrd_n_s64): Likewise.
2276         * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
2277
2278 2014-01-23  Nick Clifton  <nickc@redhat.com>
2279
2280         * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
2281         (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
2282         favour of mcu specific scripts.
2283         * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
2284         430x multilibs.
2285
2286 2014-01-23  James Greenhalgh  <james.greenhalgh@arm.com>
2287             Alex Velenko  <Alex.Velenko@arm.com>
2288
2289         * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
2290         (vaddv_s16): Likewise.
2291         (vaddv_s32): Likewise.
2292         (vaddv_u8): Likewise.
2293         (vaddv_u16): Likewise.
2294         (vaddv_u32): Likewise.
2295         (vaddvq_s8): Likewise.
2296         (vaddvq_s16): Likewise.
2297         (vaddvq_s32): Likewise.
2298         (vaddvq_s64): Likewise.
2299         (vaddvq_u8): Likewise.
2300         (vaddvq_u16): Likewise.
2301         (vaddvq_u32): Likewise.
2302         (vaddvq_u64): Likewise.
2303         (vaddv_f32): Likewise.
2304         (vaddvq_f32): Likewise.
2305         (vaddvq_f64): Likewise.
2306         (vmaxv_f32): Likewise.
2307         (vmaxv_s8): Likewise.
2308         (vmaxv_s16): Likewise.
2309         (vmaxv_s32): Likewise.
2310         (vmaxv_u8): Likewise.
2311         (vmaxv_u16): Likewise.
2312         (vmaxv_u32): Likewise.
2313         (vmaxvq_f32): Likewise.
2314         (vmaxvq_f64): Likewise.
2315         (vmaxvq_s8): Likewise.
2316         (vmaxvq_s16): Likewise.
2317         (vmaxvq_s32): Likewise.
2318         (vmaxvq_u8): Likewise.
2319         (vmaxvq_u16): Likewise.
2320         (vmaxvq_u32): Likewise.
2321         (vmaxnmv_f32): Likewise.
2322         (vmaxnmvq_f32): Likewise.
2323         (vmaxnmvq_f64): Likewise.
2324         (vminv_f32): Likewise.
2325         (vminv_s8): Likewise.
2326         (vminv_s16): Likewise.
2327         (vminv_s32): Likewise.
2328         (vminv_u8): Likewise.
2329         (vminv_u16): Likewise.
2330         (vminv_u32): Likewise.
2331         (vminvq_f32): Likewise.
2332         (vminvq_f64): Likewise.
2333         (vminvq_s8): Likewise.
2334         (vminvq_s16): Likewise.
2335         (vminvq_s32): Likewise.
2336         (vminvq_u8): Likewise.
2337         (vminvq_u16): Likewise.
2338         (vminvq_u32): Likewise.
2339         (vminnmv_f32): Likewise.
2340         (vminnmvq_f32): Likewise.
2341         (vminnmvq_f64): Likewise.
2342
2343 2014-01-23  James Greenhalgh  <james.greenhalgh@arm.com>
2344
2345         * config/aarch64/aarch64-simd.md
2346         (aarch64_dup_lane<mode>): Correct lane number on big-endian.
2347         (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
2348         (*aarch64_mul3_elt<mode>): Likewise.
2349         (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
2350         (*aarch64_mul3_elt_to_64v2df): Likewise.
2351         (*aarch64_mla_elt<mode>): Likewise.
2352         (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2353         (*aarch64_mls_elt<mode>): Likewise.
2354         (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2355         (*aarch64_fma4_elt<mode>): Likewise.
2356         (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2357         (*aarch64_fma4_elt_to_64v2df): Likewise.
2358         (*aarch64_fnma4_elt<mode>): Likewise.
2359         (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2360         (*aarch64_fnma4_elt_to_64v2df): Likewise.
2361         (aarch64_sq<r>dmulh_lane<mode>): Likewise.
2362         (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
2363         (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
2364         (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
2365         (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
2366         (aarch64_sqdmull_lane<mode>_internal): Likewise.
2367         (aarch64_sqdmull2_lane<mode>_internal): Likewise.
2368
2369 2013-01-23  Alex Velenko  <Alex.Velenko@arm.com>
2370
2371         * config/aarch64/aarch64-simd.md
2372         (aarch64_be_checked_get_lane<mode>): New define_expand.
2373         * config/aarch64/aarch64-simd-builtins.def
2374         (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
2375         New builtin definition.
2376         * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
2377         Use new safe be builtin.
2378
2379 2014-01-23  Alex Velenko  <Alex.Velenko@arm.com>
2380
2381         * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
2382         New define_insn.
2383         (aarch64_be_st1<mode>): Likewise.
2384         (aarch_ld1<VALL:mode>): Define_expand modified.
2385         (aarch_st1<VALL:mode>): Likewise.
2386         * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
2387         (UNSPEC_ST1): Likewise.
2388
2389 2014-01-23  David Holsgrove <david.holsgrove@xilinx.com>
2390
2391         * config/microblaze/microblaze.md: Add trap insn and attribute
2392
2393 2014-01-23  Dodji Seketeli  <dodji@redhat.com>
2394
2395         PR preprocessor/58580
2396         * input.h (location_get_source_line): Take an additional line_size
2397         parameter.
2398         (void diagnostics_file_cache_fini): Declare new function.
2399         * input.c (struct fcache): New type.
2400         (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
2401         New static constants.
2402         (diagnostic_file_cache_init, total_lines_num)
2403         (lookup_file_in_cache_tab, evicted_cache_tab_entry)
2404         (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
2405         (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
2406         (get_next_line, read_next_line, goto_next_line, read_line_num):
2407         New static function definitions.
2408         (diagnostic_file_cache_fini): New function.
2409         (location_get_source_line): Take an additional output line_len
2410         parameter.  Re-write using lookup_or_add_file_to_cache_tab and
2411         read_line_num.
2412         * diagnostic.c (diagnostic_finish): Call
2413         diagnostic_file_cache_fini.
2414         (adjust_line): Take an additional input parameter for the length
2415         of the line, rather than calculating it with strlen.
2416         (diagnostic_show_locus): Adjust the use of
2417         location_get_source_line and adjust_line with respect to their new
2418         signature.  While displaying a line now, do not stop at the first
2419         null byte.  Rather, display the zero byte as a space and keep
2420         going until we reach the size of the line.
2421         * Makefile.in: Add vec.o to OBJS-libcommon
2422
2423 2014-01-23  Kirill Yukhin  <kirill.yukhin@intel.com>
2424             Ilya Tocar     <ilya.tocar@intel.com>
2425
2426         * config/i386/avx512fintrin.h (_mm512_kmov): New.
2427         * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
2428         (__builtin_ia32_kmov16): Ditto.
2429         * config/i386/i386.md (UNSPEC_KMOV): New.
2430         (kmovw): Ditto.
2431
2432 2014-01-23  Kirill Yukhin  <kirill.yukhin@intel.com>
2433
2434         * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
2435         (_mm512_storeu_si512): Ditto.
2436
2437 2014-01-23  Richard Sandiford  <rdsandiford@googlemail.com>
2438
2439         PR target/52125
2440         * rtl.h (get_referenced_operands): Declare.
2441         * recog.c (get_referenced_operands): New function.
2442         * config/mips/mips.c (mips_reorg_process_insns): Check which asm
2443         operands have been referenced when recording LO_SUM references.
2444
2445 2014-01-22  David Holsgrove <david.holsgrove@xilinx.com>
2446
2447         * config/microblaze/microblaze.md: Correct bswaphi2 insn.
2448
2449 2014-01-22  Jan Hubicka  <hubicka@ucw.cz>
2450
2451         * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
2452         Enable for generic and recent AMD targets.
2453
2454 2014-01-22  Jan Hubicka  <hubicka@ucw.cz>
2455
2456         * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
2457         ARG_SIZE note when adjustment was eliminated.
2458
2459 2014-01-22  Jeff Law  <law@redhat.com>
2460
2461         PR tree-optimization/59597
2462         * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
2463         in file.  Accept new argument REGISTERING and use it to modify
2464         dump output appropriately.
2465         (register_jump_thread): Corresponding changes.
2466         (mark_threaded_blocks): Reinstate code to cancel unprofitable
2467         thread paths involving joiner blocks.  Add code to dump cancelled
2468         jump threading paths.
2469
2470 2014-01-22  Vladimir Makarov  <vmakarov@redhat.com>
2471
2472         PR rtl-optimization/59477
2473         * lra-constraints.c (inherit_in_ebb): Process call for living hard
2474         regs.  Update reloads_num and potential_reload_hard_regs for all insns.
2475
2476 2014-01-22  Tom Tromey  <tromey@redhat.com>
2477
2478         * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
2479         PARAMS.
2480         * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
2481
2482 2014-01-21  Vladimir Makarov  <vmakarov@redhat.com>
2483
2484         PR rtl-optimization/59896
2485         * lra-constraints.c (process_alt_operands): Check unused note for
2486         matched operands of insn with no output reloads.
2487
2488 2014-01-21  Richard Sandiford  <rdsandiford@googlemail.com>
2489
2490         * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
2491         (mips_move_from_gpr_cost): Likewise.
2492
2493 2014-01-21  Vladimir Makarov  <vmakarov@redhat.com>
2494
2495         PR rtl-optimization/59858
2496         * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
2497         ira_class_hard_regs_num.
2498         (process_alt_operands): Increase reject for dying matched operand.
2499
2500 2014-01-21  Jakub Jelinek  <jakub@redhat.com>
2501
2502         PR target/59003
2503         * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
2504         smaller than size, perform several stores or loads and stores
2505         at dst + count - size to store or copy all of size bytes, rather
2506         than just last modesize bytes.
2507
2508 2014-01-20  DJ Delorie  <dj@redhat.com>
2509
2510         * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
2511         that CLOBBERs are REGs before propogating their values.
2512
2513 2014-01-20  H.J. Lu  <hongjiu.lu@intel.com>
2514
2515         PR middle-end/59789
2516         * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
2517         (cgraph_inline_failed_type): New function.
2518         * cgraph.h (DEFCIFCODE): Add type.
2519         (cgraph_inline_failed_type_t): New enum.
2520         (cgraph_inline_failed_type): New prototype.
2521         * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
2522         FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
2523         FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
2524         LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
2525         MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
2526         RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
2527         OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
2528         INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
2529         Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
2530         FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
2531         EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
2532         OPTIMIZATION_MISMATCH.
2533         * tree-inline.c (expand_call_inline): Emit errors during
2534         early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
2535
2536 2014-01-20  Uros Bizjak  <ubizjak@gmail.com>
2537
2538         PR target/59685
2539         * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
2540         mode attribute in insn output.
2541
2542 2014-01-20  Eric Botcazou  <ebotcazou@adacore.com>
2543
2544         * output.h (output_constant): Delete.
2545         * varasm.c (output_constant): Make private.
2546
2547 2014-01-20  Alex Velenko  <Alex.Velenko@arm.com>
2548
2549         * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
2550
2551 2014-01-20  Jakub Jelinek  <jakub@redhat.com>
2552
2553         PR middle-end/59860
2554         * tree.h (fold_builtin_strcat): New prototype.
2555         * builtins.c (fold_builtin_strcat): No longer static.  Add len
2556         argument, if non-NULL, don't call c_strlen.  Optimize
2557         directly into __builtin_memcpy instead of __builtin_strcpy.
2558         (fold_builtin_2): Adjust fold_builtin_strcat caller.
2559         * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
2560
2561 2014-01-20  Uros Bizjak  <ubizjak@gmail.com>
2562
2563         * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2564         for SImode_address_operand operands, having only a REG argument.
2565
2566 2014-01-20  Marcus Shawcroft  <marcus.shawcroft@arm.com>
2567
2568         * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
2569         loader name using mbig-endian.
2570         (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
2571
2572 2014-01-20  James Greenhalgh  <james.greenhalgh@arm.com>
2573
2574         * doc/invoke.texi (-march): Clarify documentation for AArch64.
2575         (-mtune): Likewise.
2576         (-mcpu): Likewise.
2577
2578 2014-01-20  Tejas Belagod  <tejas.belagod@arm.com>
2579
2580         * config/aarch64/aarch64-protos.h
2581         (aarch64_cannot_change_mode_class_ptr): Declare.
2582         * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
2583         aarch64_cannot_change_mode_class_ptr): New.
2584         * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
2585         backend hook aarch64_cannot_change_mode_class.
2586
2587 2014-01-20  James Greenhalgh  <james.greenhalgh@arm.com>
2588
2589         * common/config/aarch64/aarch64-common.c
2590         (aarch64_handle_option): Don't handle any option order logic here.
2591         * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
2592         selected_cpu, warn on architecture version mismatch.
2593         (aarch64_override_options): Fix parsing order for option strings.
2594
2595 2014-01-20  Jan-Benedict Glaw  <jbglaw@lug-owl.de>
2596             Iain Sandoe  <iain@codesourcery.com>
2597
2598         PR bootstrap/59496
2599         * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
2600         warning.  Amend comment to reflect current functionality.
2601
2602 2014-01-20  Richard Biener  <rguenther@suse.de>
2603
2604         PR middle-end/59860
2605         * builtins.c (fold_builtin_strcat): Remove case better handled
2606         by tree-ssa-strlen.c.
2607
2608 2014-01-20  Alan Lawrence  <alan.lawrence@arm.com>
2609
2610         * config/aarch64/aarch64.opt
2611         (mcpu, march, mtune): Make case-insensitive.
2612
2613 2014-01-20  Jakub Jelinek  <jakub@redhat.com>
2614
2615         PR target/59880
2616         * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2617         if operands[1] is a REG or ZERO_EXTEND of a REG.
2618
2619 2014-01-19  Jan Hubicka  <hubicka@ucw.cz>
2620
2621         * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
2622
2623 2014-01-19  John David Anglin  <danglin@gcc.gnu.org>
2624
2625         * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
2626         long non-pic millicode calls.
2627
2628 2014-01-19  Jan-Benedict Glaw  <jbglaw@lug-owl.de>
2629
2630         * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
2631
2632 2014-01-19  Kito Cheng  <kito@0xlab.org>
2633
2634         * builtins.c (expand_movstr): Check movstr expand done or fail.
2635
2636 2014-01-18  Uros Bizjak  <ubizjak@gmail.com>
2637             H.J. Lu  <hongjiu.lu@intel.com>
2638
2639         PR target/59379
2640         * config/i386/i386.md (*lea<mode>): Zero-extend return register
2641         to DImode for zero-extended addresses.
2642
2643 2014-01-19  Jakub Jelinek  <jakub@redhat.com>
2644
2645         PR rtl-optimization/57763
2646         * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
2647         on the new indirect jump_insn and increment LABEL_NUSES (label).
2648
2649 2014-01-18  H.J. Lu  <hongjiu.lu@intel.com>
2650
2651         PR bootstrap/59580
2652         PR bootstrap/59583
2653         * config.gcc (x86_archs): New variable.
2654         (x86_64_archs): Likewise.
2655         (x86_cpus): Likewise.
2656         Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
2657         --with-arch/--with-cpu= options.
2658         Support --with-arch=/--with-cpu={nehalem,westmere,
2659         sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
2660
2661 2014-01-18  Uros Bizjak  <ubizjak@gmail.com>
2662
2663         * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
2664         and PROCESSOR_ATHLON to simplify code.  Move "memory" calculation.
2665
2666 2014-01-18  Uros Bizjak  <ubizjak@gmail.com>
2667
2668         * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
2669
2670 2014-01-18  Jakub Jelinek  <jakub@redhat.com>
2671
2672         PR target/58944
2673         * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
2674         clear cpp_get_options (parse_in)->warn_unused_macros for
2675         ix86_target_macros_internal with cpp_define.
2676
2677 2014-01-18  Richard Sandiford  <rdsandiford@googlemail.com>
2678
2679         * jump.c (delete_related_insns): Keep (use (insn))s.
2680         * reorg.c (redundant_insn): Check for barriers too.
2681
2682 2014-01-17  H.J. Lu  <hongjiu.lu@intel.com>
2683
2684         * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
2685
2686 2014-01-17  John David Anglin  <danglin@gcc.gnu.org>
2687
2688         * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
2689         call to $$dyncall when TARGET_LONG_CALLS is true.
2690
2691 2014-01-17  Jeff Law  <law@redhat.com>
2692
2693         * ree.c (combine_set_extension): Temporarily disable test for
2694         changing number of hard registers.
2695
2696 2014-01-17  Jan Hubicka  <hubicka@ucw.cz>
2697
2698         PR middle-end/58125
2699         * ipa-inline-analysis.c (inline_free_summary):
2700         Do not free summary of aliases.
2701
2702 2014-01-17  Jakub Jelinek  <jakub@redhat.com>
2703
2704         PR middle-end/59706
2705         * gimplify.c (gimplify_expr): Use create_tmp_var
2706         instead of create_tmp_var_raw.  If cond doesn't have
2707         integral type, don't add the IFN_ANNOTATE builtin at all.
2708
2709 2014-01-17  Martin Jambor  <mjambor@suse.cz>
2710
2711         PR ipa/59736
2712         * ipa-cp.c (prev_edge_clone): New variable.
2713         (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
2714         Also resize prev_edge_clone vector.
2715         (ipcp_edge_duplication_hook): Also update prev_edge_clone.
2716         (ipcp_edge_removal_hook): New function.
2717         (ipcp_driver): Register ipcp_edge_removal_hook.
2718
2719 2014-01-17  Andrew Pinski <apinski@cavium.com>
2720             Steve Ellcey  <sellcey@mips.com>
2721
2722         PR target/59462
2723         * config/mips/mips.c (mips_print_operand): Check operand mode instead
2724         of operator mode.
2725
2726 2014-01-17  Jeff Law  <law@redhat.com>
2727
2728         PR middle-end/57904
2729         * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
2730         so that pass_ccp runs first.
2731
2732 2014-01-17  H.J. Lu  <hongjiu.lu@intel.com>
2733
2734         * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
2735         (ix86_adjust_cost): Use !TARGET_XXX.
2736         (do_reorder_for_imul): Likewise.
2737         (swap_top_of_ready_list): Likewise.
2738         (ix86_sched_reorder): Likewise.
2739
2740 2014-01-17  H.J. Lu  <hongjiu.lu@intel.com>
2741
2742         * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2743         PROCESSOR_INTEL.  Treat like PROCESSOR_GENERIC.
2744         * config/i386/i386.c (intel_memcpy): New.  Duplicate slm_memcpy.
2745         (intel_memset): New.  Duplicate slm_memset.
2746         (intel_cost): New.  Duplicate slm_cost.
2747         (m_INTEL): New macro.
2748         (processor_target_table): Add "intel".
2749         (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
2750         with PROCESSOR_INTEL for "intel".
2751         (ix86_lea_outperforms): Support PROCESSOR_INTEL.  Duplicate
2752         PROCESSOR_SILVERMONT.
2753         (ix86_issue_rate): Likewise.
2754         (ix86_adjust_cost): Likewise.
2755         (ia32_multipass_dfa_lookahead): Likewise.
2756         (swap_top_of_ready_list): Likewise.
2757         (ix86_sched_reorder): Likewise.
2758         (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
2759         instead of TARGET_OPT_AGU.
2760         * config/i386/i386.h (TARGET_INTEL): New.
2761         (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
2762         (processor_type): Add PROCESSOR_INTEL.
2763         * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
2764         Add X86_TUNE_AVOID_LEA_FOR_ADDR.
2765
2766 2014-01-17  Marek Polacek  <polacek@redhat.com>
2767
2768         PR c/58346
2769         * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
2770         size is zero.
2771
2772 2014-01-17  Richard Biener  <rguenther@suse.de>
2773
2774         PR tree-optimization/46590
2775         * opts.c (default_options_table): Add entries for
2776         OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
2777         all enabled at -O1 but not for -Og.
2778         * common.opt (fbranch-count-reg): Remove Init(1).
2779         (fmove-loop-invariants): Likewise.
2780         (ftree-pta): Likewise.
2781
2782 2014-01-17  Jakub Jelinek  <jakub@redhat.com>
2783
2784         * config/i386/i386.c (ix86_data_alignment): For compatibility with
2785         (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
2786         decls to at least the GCC 4.8 used alignments.
2787
2788         PR fortran/59440
2789         * tree-nested.c (convert_nonlocal_reference_stmt,
2790         convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
2791         of GIMPLE_BIND stmts, adjust associated decls.
2792
2793 2014-01-17  Richard Biener  <rguenther@suse.de>
2794
2795         PR tree-optimization/46590
2796         * vec.h (vec<>::bseach): New member function implementing
2797         binary search according to C89 bsearch.
2798         (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
2799         * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
2800         bitmap pointer again.  Make accesses_in_loop a flat array.
2801         (mem_ref_obstack): New global.
2802         (outermost_indep_loop): Adjust for mem_ref->stored changes.
2803         (mark_ref_stored): Likewise.
2804         (ref_indep_loop_p_2): Likewise.
2805         (set_ref_stored_in_loop): New helper function.
2806         (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
2807         (memref_free): Adjust.
2808         (record_mem_ref_loc): Simplify.
2809         (gather_mem_refs_stmt): Adjust.
2810         (sort_locs_in_loop_postorder_cmp): New function.
2811         (analyze_memory_references): Sort accesses_in_loop after
2812         loop postorder number.
2813         (find_ref_loc_in_loop_cmp): New function.
2814         (for_all_locs_in_loop): Find relevant cluster of locs in
2815         accesses_in_loop and iterate without recursion.
2816         (execute_sm): Avoid uninit warning.
2817         (struct ref_always_accessed): Simplify.
2818         (ref_always_accessed::operator ()): Likewise.
2819         (ref_always_accessed_p): Likewise.
2820         (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
2821         loop postorder numbers here.
2822         (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
2823         numbers.
2824
2825 2014-01-17  Jan Hubicka  <hubicka@ucw.cz>
2826
2827         PR c++/57945
2828         * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
2829         on decls for which assemble_alias has been called.
2830
2831 2014-01-17  Nick Clifton  <nickc@redhat.com>
2832
2833         * config/msp430/msp430.opt: (mcpu): New option.
2834         * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
2835         (msp430_option_override): Parse target_cpu.  If the MCU name
2836         matches a generic string, clear target_mcu.
2837         (msp430_attr): Allow numeric interrupt values up to 63.
2838         (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
2839         * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
2840         option.
2841         * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
2842         Add mcpu matches.
2843         * config/msp430/msp430.md (popm): Use %J rather than %I.
2844         (addsi3): Use msp430_nonimmediate_operand for operand 2.
2845         (addhi_cy_i): Use immediate_operand for operand 2.
2846         * doc/invoke.texi: Document -mcpu option.
2847
2848 2014-01-17  Richard Biener  <rguenther@suse.de>
2849
2850         PR rtl-optimization/38518
2851         * df.h (df_analyze_loop): Declare.
2852         * df-core.c: Include cfgloop.h.
2853         (df_analyze_1): Split out main part of df_analyze.
2854         (df_analyze): Adjust.
2855         (loop_inverted_post_order_compute): New function.
2856         (loop_post_order_compute): Likewise.
2857         (df_analyze_loop): New function avoiding whole-function
2858         postorder computes.
2859         * loop-invariant.c (find_defs): Use df_analyze_loop.
2860         (find_invariants): Adjust.
2861         * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
2862
2863 2014-01-17  Zhenqiang Chen  <zhenqiang.chen@arm.com>
2864
2865         * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
2866         (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
2867
2868 2014-01-16  Ilya Enkovich  <ilya.enkovich@intel.com>
2869
2870         * ipa-ref.c (ipa_remove_stmt_references): Fix references
2871         traversal when removing references.
2872
2873 2014-01-16  Jan Hubicka  <hubicka@ucw.cz>
2874
2875         PR ipa/59775
2876         * tree.c (get_binfo_at_offset): Look harder for virtual bases.
2877
2878 2014-01-16  Bernd Schmidt  <bernds@codesourcery.com>
2879
2880         PR middle-end/56791
2881         * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
2882         pushing a reload for an autoinc when we had previously reloaded an
2883         inner part of the address.
2884
2885 2014-01-16  Jakub Jelinek  <jakub@redhat.com>
2886
2887         * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
2888         field.
2889         (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
2890         * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
2891         when not giving up or versioning for alias only because of
2892         loop->safelen.
2893         (vect_analyze_data_ref_dependences): Set to true.
2894         * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
2895         is a GIMPLE_PHI.
2896         (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
2897         LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
2898         to the condition.
2899
2900         PR middle-end/58344
2901         * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
2902
2903         PR target/59839
2904         * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
2905         operand 0 predicate for gathers, use a new pseudo as subtarget.
2906
2907 2014-01-16  Vladimir Makarov  <vmakarov@redhat.com>
2908
2909         PR middle-end/59609
2910         * lra-constraints.c (process_alt_operands): Add printing debug info.
2911         Check absence of input/output reloads for matched operands too.
2912
2913 2014-01-16  Vladimir Makarov  <vmakarov@redhat.com>
2914
2915         PR rtl-optimization/59835
2916         * ira.c (ira_init_register_move_cost): Increase cost for
2917         impossible modes.
2918
2919 2014-01-16  Alan Lawrence  <alan.lawrence@arm.com>
2920
2921         * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
2922
2923 2014-01-16  Richard Earnshaw  <rearnsha@arm.com>
2924
2925         PR target/59780
2926         * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
2927         non-register objects.  Use gen_(high/low)part more consistently.
2928         Fix assertions.
2929
2930 2014-01-16  Michael Meissner  <meissner@linux.vnet.ibm.com>
2931
2932         PR target/59844
2933         * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
2934         endian support, remove tests for WORDS_BIG_ENDIAN.
2935         (p8_mfvsrd_3_<mode>): Likewise.
2936         (reload_gpr_from_vsx<mode>): Likewise.
2937         (reload_gpr_from_vsxsf): Likewise.
2938         (p8_mfvsrd_4_disf): Likewise.
2939
2940 2014-01-16  Richard Biener  <rguenther@suse.de>
2941
2942         PR rtl-optimization/46590
2943         * lcm.c (compute_antinout_edge): Use postorder iteration.
2944         (compute_laterin): Use inverted postorder iteration.
2945
2946 2014-01-16  Nick Clifton  <nickc@redhat.com>
2947
2948         PR middle-end/28865
2949         * varasm.c (output_constant): Return the number of bytes actually
2950         emitted.
2951         (output_constructor_array_range): Update the field size with the
2952         number of bytes emitted by output_constant.
2953         (output_constructor_regular_field): Likewise.  Also do not
2954         complain if the total number of bytes emitted is now greater
2955         than the expected fieldpos.
2956         * output.h (output_constant): Update prototype and descriptive comment.
2957
2958 2014-01-16  Marek Polacek  <polacek@redhat.com>
2959
2960         PR middle-end/59827
2961         * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
2962         it is error_mark_node.
2963
2964 2014-01-15  Uros Bizjak  <ubizjak@gmail.com>
2965
2966         * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
2967         VALID_AVX256_REG_OR_OI_MODE.
2968
2969 2014-01-15  Pat Haugen  <pthaugen@us.ibm.com>
2970
2971         * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
2972         current procedure should be profiled.
2973
2974 2014-01-15  Andrew Pinski  <apinski@cavium.com>
2975
2976         * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
2977         of moving from/to the STACK_REG register class.
2978
2979 2014-01-15  Richard Henderson  <rth@redhat.com>
2980
2981         PR debug/54694
2982         * reginfo.c (global_regs_decl): Globalize.
2983         * rtl.h (global_regs_decl): Declare.
2984         * ira.c (do_reload): Diagnose frame_pointer_needed and it
2985         reserved via global_regs.
2986
2987 2014-01-15  Teresa Johnson  <tejohnson@google.com>
2988
2989         * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
2990
2991 2014-01-15  Bill Schmidt  <wschmidt@vnet.linux.ibm.com>
2992
2993         * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
2994         and vmulosh rather than call gen_vec_widen_smult_*.
2995         (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
2996         than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
2997         (vec_widen_smult_even_v16qi): Likewise.
2998         (vec_widen_umult_even_v8hi): Likewise.
2999         (vec_widen_smult_even_v8hi): Likewise.
3000         (vec_widen_umult_odd_v16qi): Likewise.
3001         (vec_widen_smult_odd_v16qi): Likewise.
3002         (vec_widen_umult_odd_v8hi): Likewise.
3003         (vec_widen_smult_odd_v8hi): Likewise.
3004         (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
3005         vmuloub rather than call gen_vec_widen_umult_*.
3006         (vec_widen_umult_lo_v16qi): Likewise.
3007         (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
3008         vmulosb rather than call gen_vec_widen_smult_*.
3009         (vec_widen_smult_lo_v16qi): Likewise.
3010         (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
3011         rather than call gen_vec_widen_umult_*.
3012         (vec_widen_umult_lo_v8hi): Likewise.
3013         (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
3014         rather than call gen_vec_widen_smult_*.
3015         (vec_widen_smult_lo_v8hi): Likewise.
3016
3017 2014-01-15  Jeff Law  <law@redhat.com>
3018
3019         PR tree-optimization/59747
3020         * ree.c (find_and_remove_re): Properly handle case where a second
3021         eliminated extension requires widening a copy created for elimination
3022         of a prior extension.
3023         (combine_set_extension): Ensure that the number of hard regs needed
3024         for a destination register does not change when we widen it.
3025
3026 2014-01-15  Sebastian Huber  <sebastian.huber@embedded-brains.de>
3027
3028         * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
3029         (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
3030         (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
3031         (arm*-*-rtems*): Use t-rtems from existing tmake_file.
3032         (avr-*-rtems*): Likewise.
3033         (bfin*-rtems*): Likewise.
3034         (moxie-*-rtems*): Likewise.
3035         (h8300-*-rtems*): Likewise.
3036         (i[34567]86-*-rtems*): Likewise.
3037         (lm32-*-rtems*): Likewise.
3038         (m32r-*-rtems*): Likewise.
3039         (m68k-*-rtems*): Likewise.
3040         (microblaze*-*-rtems*): Likewise.
3041         (mips*-*-rtems*): Likewise.
3042         (powerpc-*-rtems*): Likewise.
3043         (sh-*-rtems*): Likewise.
3044         (sparc-*-rtems*): Likewise.
3045         (sparc64-*-rtems*): Likewise.
3046         (v850-*-rtems*): Likewise.
3047         (m32c-*-rtems*): Likewise.
3048
3049 2014-01-15  Vladimir Makarov  <vmakarov@redhat.com>
3050
3051         PR rtl-optimization/59511
3052         * ira.c (ira_init_register_move_cost): Use memory costs for some
3053         cases of register move cost calculations.
3054         * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
3055         instead of BB frequency.
3056         * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
3057         * lra-assigns.c (find_hard_regno_for): Ditto.
3058
3059 2014-01-15  Richard Biener  <rguenther@suse.de>
3060
3061         PR tree-optimization/59822
3062         * tree-vect-stmts.c (hoist_defs_of_uses): New function.
3063         (vectorizable_load): Use it to hoist defs of uses of invariant
3064         loads out of the loop.
3065
3066 2014-01-15  Matthew Gretton-Dann  <matthew.gretton-dann@linaro.org>
3067             Kugan Vivekanandarajah  <kuganv@linaro.org>
3068
3069         PR target/59695
3070         * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
3071         truncation.
3072
3073 2014-01-15  Richard Biener  <rguenther@suse.de>
3074
3075         PR rtl-optimization/59802
3076         * lcm.c (compute_available): Use inverted postorder to seed
3077         the initial worklist.
3078
3079 2014-01-15  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>
3080
3081         PR target/59803
3082         * config/s390/s390.c (s390_preferred_reload_class): Don't return
3083         ADDR_REGS for invalid symrefs in non-PIC code.
3084
3085 2014-01-15  Jakub Jelinek  <jakub@redhat.com>
3086
3087         PR other/58712
3088         * builtins.c (determine_block_size): Initialize *probable_max_size
3089         even if len_rtx is CONST_INT.
3090
3091 2014-01-14  Andrew Pinski  <apinski@cavium.com>
3092
3093         * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
3094         * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
3095         (cortexa53_tunings): Likewise.
3096         (aarch64_sched_issue_rate): New function.
3097         (TARGET_SCHED_ISSUE_RATE): Define.
3098
3099 2014-01-14  Vladimir Makarov  <vmakarov@redhat.com>
3100
3101         * ira-costs.c (find_costs_and_classes): Add missed
3102         ira_init_register_move_cost_if_necessary.
3103
3104 2014-01-14  Vladimir Makarov  <vmakarov@redhat.com>
3105
3106         PR target/59787
3107         * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
3108
3109 2014-01-14  H.J. Lu  <hongjiu.lu@intel.com>
3110
3111         PR target/59794
3112         * config/i386/i386.c (type_natural_mode): Add a bool parameter
3113         to indicate if type is used for function return value.  Warn ABI
3114         change if the vector mode isn't available for function return value.
3115         (ix86_function_arg_advance): Pass false to type_natural_mode.
3116         (ix86_function_arg): Likewise.
3117         (ix86_gimplify_va_arg): Likewise.
3118         (function_arg_32): Don't warn ABI change.
3119         (ix86_function_value): Pass true to type_natural_mode.
3120         (ix86_return_in_memory): Likewise.
3121         (ix86_struct_value_rtx): Removed.
3122         (TARGET_STRUCT_VALUE_RTX): Likewise.
3123
3124 2014-01-14  Richard Sandiford  <rsandifo@linux.vnet.ibm.com>
3125
3126         * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
3127         converting a conditional jump into a conditional return.
3128
3129 2014-01-14  Richard Biener  <rguenther@suse.de>
3130
3131         PR tree-optimization/58921
3132         PR tree-optimization/59006
3133         * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
3134         hoisting invariant stmts.
3135         * tree-vect-stmts.c (vectorizable_load): Insert the splat of
3136         invariant loads on the preheader edge if possible.
3137
3138 2014-01-14  Joey Ye  <joey.ye@arm.com>
3139
3140         * doc/plugin.texi (Building GCC plugins): Update to C++.
3141
3142 2014-01-14  Kirill Yukhin  <kirill.yukhin@intel.com>
3143
3144         * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
3145         (_mm_rcp28_round_ss): Ditto.
3146         (_mm_rsqrt28_round_sd): Ditto.
3147         (_mm_rsqrt28_round_ss): Ditto.
3148         (_mm_rcp28_sd): Ditto.
3149         (_mm_rcp28_ss): Ditto.
3150         (_mm_rsqrt28_sd): Ditto.
3151         (_mm_rsqrt28_ss): Ditto.
3152         * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
3153         * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
3154         * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
3155         (IX86_BUILTIN_RCP28SD): Ditto.
3156         (IX86_BUILTIN_RCP28SS): Ditto.
3157         (IX86_BUILTIN_RSQRT28SD): Ditto.
3158         (IX86_BUILTIN_RSQRT28SS): Ditto.
3159         (bdesc_special_args): Define __builtin_ia32_movntdqa512,
3160         __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
3161         __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
3162         (ix86_expand_special_args_builtin): Expand new FTYPE.
3163         * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
3164         (srcp14<mode>): Make insn unary.
3165         (avx512f_vmscalef<mode><round_name>): Use substed predicate.
3166         (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
3167         (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
3168         (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
3169         (avx512er_exp2<mode><mask_name><round_saeonly_name>):
3170         Fix rounding: make it SAE only.
3171         (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
3172         Ditto.
3173         (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
3174         Ditto.
3175         (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
3176         (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
3177         (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
3178         * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
3179         (round_saeonly_mask_scalar_operand4): Ditto.
3180         (round_saeonly_mask_scalar_op3): Ditto.
3181         (round_saeonly_mask_scalar_op4): Ditto.
3182
3183 2014-01-13  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
3184
3185         * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
3186         Implement -maltivec=be for vec_insert and vec_extract.
3187
3188 2014-01-10  DJ Delorie  <dj@redhat.com>
3189
3190         * config/msp430/msp430.md (call_internal): Don't allow memory
3191         references with SP as the base register.
3192         (call_value_internal): Likewise.
3193         * config/msp430/constraints.md (Yc): New.  For memory references
3194         that don't use SP as a base register.
3195
3196         * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
3197         "an integer without a # prefix"
3198         * config/msp430/msp430.md (epilogue_helper): Use it.
3199
3200 2014-01-13  Jakub Jelinek  <jakub@redhat.com>
3201
3202         PR target/59617
3203         * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
3204         AVX512F gather builtins.
3205         * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
3206         on gather decls with INTEGER_TYPE masktype.
3207         (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
3208         directly into the builtin rather than hoisting it before loop.
3209
3210         PR tree-optimization/59387
3211         * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
3212         (scev_const_prop): If folded_casts and type has undefined overflow,
3213         use force_gimple_operand instead of force_gimple_operand_gsi and
3214         for each added stmt if it is assign with
3215         arith_code_with_undefined_signed_overflow, call
3216         rewrite_to_defined_overflow.
3217         * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
3218         gimple-fold.h instead.
3219         (arith_code_with_undefined_signed_overflow,
3220         rewrite_to_defined_overflow): Moved to ...
3221         * gimple-fold.c (arith_code_with_undefined_signed_overflow,
3222         rewrite_to_defined_overflow): ... here.  No longer static.
3223         Include gimplify-me.h.
3224         * gimple-fold.h (arith_code_with_undefined_signed_overflow,
3225         rewrite_to_defined_overflow): New prototypes.
3226
3227 2014-01-13  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
3228
3229         * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
3230
3231 2014-01-13  Eric Botcazou  <ebotcazou@adacore.com>
3232
3233         * builtins.c (get_object_alignment_2): Minor tweak.
3234         * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
3235
3236 2014-01-13  Christian Bruel  <christian.bruel@st.com>
3237
3238         * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
3239         optimized non constant lengths.
3240
3241 2014-01-13  Jakub Jelinek  <jakub@redhat.com>
3242
3243         PR libgomp/59194
3244         * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
3245         load as __atomic_load_N if possible.
3246
3247 2014-01-11  David Edelsohn  <dje.gcc@gmail.com>
3248
3249         * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
3250         target parameter.
3251         (rs6000_expand_builtin): Adjust call.
3252
3253 2014-01-11  David Edelsohn  <dje.gcc@gmail.com>
3254
3255         PR target/58115
3256         * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
3257         * config/rs6000/rs6000.c: Include target-globals.h.
3258         (rs6000_set_current_function): Instead of doing target_reinit
3259         unconditionally, use save_target_globals_default_opts and
3260         restore_target_globals.
3261
3262         * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
3263         FPSCR.
3264         * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
3265         (rs6000_expand_builtin): Handle mffs and mtfsf.
3266         (rs6000_init_builtins): Define mffs and mtfsf.
3267         * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
3268         (rs6000_mffs): New pattern.
3269         (rs6000_mtfsf): New pattern.
3270
3271 2014-01-11  Bin Cheng  <bin.cheng@arm.com>
3272
3273         * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
3274         Start narrowing with START.  Apply candidate-use pair
3275         and check overall cost in narrowing.
3276         (iv_ca_prune): Pass new argument.
3277
3278 2014-01-10  Jeff Law  <law@redhat.com>
3279
3280         PR middle-end/59743
3281         * ree.c (combine_reaching_defs): Ensure the defining statement
3282         occurs before the extension when optimizing extensions with
3283         different source and destination hard registers.
3284
3285 2014-01-10  Jan Hubicka  <hubicka@ucw.cz>
3286
3287         PR ipa/58585
3288         * ipa-devirt.c (build_type_inheritance_graph): Also add types of
3289         vtables into the type inheritance graph.
3290
3291 2014-01-10  Jakub Jelinek  <jakub@redhat.com>
3292
3293         PR rtl-optimization/59754
3294         * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
3295         modes in the REGNO != REGNO case.
3296
3297 2014-01-10  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
3298
3299         * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
3300
3301 2014-01-10  Jakub Jelinek  <jakub@redhat.com>
3302
3303         PR tree-optimization/59745
3304         * tree-predcom.c (tree_predictive_commoning_loop): Call
3305         free_affine_expand_cache if giving up because components is NULL.
3306
3307         * target-globals.c (save_target_globals): Allocate < 4KB structs using
3308         GC in payload of target_globals struct instead of allocating them on
3309         the heap and the larger structs separately using GC.
3310         * target-globals.h (struct target_globals): Make regs, hard_regs,
3311         reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
3312         of GTY((skip)) and change type to void *.
3313         (reset_target_globals): Cast loads from those fields to corresponding
3314         types.
3315
3316 2014-01-10  Steve Ellcey  <sellcey@mips.com>
3317
3318         PR plugins/59335
3319         * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
3320         gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
3321         tree-ssanames.h, print-tree.h, varasm.h, and context.h.
3322
3323 2014-01-10  Richard Earnshaw  <rearnsha@arm.com>
3324
3325         PR target/59744
3326         * aarch64-modes.def (CC_Zmode): New flags mode.
3327         * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
3328         represents an equality.
3329         (aarch64_get_condition_code): Handle CC_Zmode.
3330         * aarch64.md (compare_neg<mode>): Restrict to equality operations.
3331
3332 2014-01-10  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>
3333
3334         * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
3335         extraction in good case.
3336
3337 2014-01-10  Richard Biener  <rguenther@suse.de>
3338
3339         PR tree-optimization/59374
3340         * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
3341         checking after SLP discovery.  Mark stmts not participating
3342         in any SLP instance properly.
3343
3344 2014-01-10  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
3345
3346         * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
3347         when handling a SET rtx.
3348
3349 2014-01-10  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
3350
3351         * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
3352         (cortex-a57): Likewise.
3353         (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
3354
3355 2014-01-10  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
3356
3357         * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
3358         non-iwmmxt builtins.
3359
3360 2014-01-10  Jan Hubicka  <hubicka@ucw.cz>
3361
3362         PR ipa/58252
3363         PR ipa/59226
3364         * ipa-devirt.c record_target_from_binfo): Take as argument
3365         stack of binfos and lookup matching one for virtual inheritance.
3366         (possible_polymorphic_call_targets_1): Update.
3367
3368 2014-01-10  Huacai Chen  <chenhc@lemote.com>
3369
3370         * config/mips/driver-native.c (host_detect_local_cpu): Handle new
3371         kernel strings for Loongson-2E/2F/3A.
3372
3373 2014-01-10  Jakub Jelinek  <jakub@redhat.com>
3374
3375         PR middle-end/59670
3376         * tree-vect-data-refs.c (vect_analyze_data_refs): Check
3377         is_gimple_call before calling gimple_call_internal_p.
3378
3379 2014-01-09  Steve Ellcey  <sellcey@mips.com>
3380
3381         * Makefile.in (TREE_FLOW_H): Remove.
3382         (TREE_SSA_H): Add file names from tree-flow.h.
3383         * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
3384         * tree.h: Remove tree-flow.h reference.
3385         * hash-table.h: Remove tree-flow.h reference.
3386         * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
3387         reference with tree-ssa-loop.h.
3388
3389 2014-01-09  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
3390
3391         * doc/invoke.texi: Add -maltivec={be,le} options, and document
3392         default element-order behavior for -maltivec.
3393         * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
3394         * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
3395         that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
3396         when targeting big endian, at least for now.
3397         * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
3398
3399 2014-01-09  Jakub Jelinek  <jakub@redhat.com>
3400
3401         PR middle-end/47735
3402         * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
3403         var satisfies use_register_for_decl, just take into account type
3404         alignment, rather than decl alignment.
3405
3406         PR tree-optimization/59622
3407         * gimple-fold.c (gimple_fold_call): Fix a typo in message.  For
3408         __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
3409         __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
3410         Don't devirtualize for inplace at all.  For targets.length () == 1,
3411         if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
3412
3413 2014-01-09  H.J. Lu  <hongjiu.lu@intel.com>
3414
3415         * config/i386/i386.md (cpu): Remove the unused btver1.
3416
3417 2014-01-09  H.J. Lu  <hongjiu.lu@intel.com>
3418
3419         * gdbasan.in: Put a breakpoint on __sanitizer::Report.
3420
3421 2014-01-09  Jakub Jelinek  <jakub@redhat.com>
3422
3423         PR target/58115
3424         * tree-core.h (struct target_globals): New forward declaration.
3425         (struct tree_target_option): Add globals field.
3426         * tree.h (TREE_TARGET_GLOBALS): Define.
3427         (prepare_target_option_nodes_for_pch): New prototype.
3428         * target-globals.h (struct target_globals): Define even if
3429         !SWITCHABLE_TARGET.
3430         * tree.c (prepare_target_option_node_for_pch,
3431         prepare_target_option_nodes_for_pch): New functions.
3432         * config/i386/i386.h (SWITCHABLE_TARGET): Define.
3433         * config/i386/i386.c: Include target-globals.h.
3434         (ix86_set_current_function): Instead of doing target_reinit
3435         unconditionally, use save_target_globals_default_opts and
3436         restore_target_globals.
3437
3438 2014-01-09  Richard Biener  <rguenther@suse.de>
3439
3440         PR tree-optimization/59715
3441         * tree-cfg.h (split_critical_edges): Declare.
3442         * tree-cfg.c (split_critical_edges): Export.
3443         * tree-ssa-sink.c (execute_sink_code): Split critical edges.
3444
3445 2014-01-09  Max Ostapenko  <m.ostapenko@partner.samsung.com>
3446
3447         * cfgexpand.c (expand_stack_vars): Optionally disable
3448         asan stack protection.
3449         (expand_used_vars): Likewise.
3450         (partition_stack_vars): Likewise.
3451         * asan.c (asan_emit_stack_protection): Optionally disable
3452         after return stack usage.
3453         (instrument_derefs): Optionally disable memory access instrumentation.
3454         (instrument_builtin_call): Likewise.
3455         (instrument_strlen_call): Likewise.
3456         (asan_protect_global): Optionally disable global variables protection.
3457         * doc/invoke.texi: Added doc for new options.
3458         * params.def: Added new options.
3459         * params.h: Likewise.
3460
3461 2014-01-09  Jakub Jelinek  <jakub@redhat.com>
3462
3463         PR rtl-optimization/59724
3464         * ifcvt.c (cond_exec_process_if_block): Don't call
3465         flow_find_head_matching_sequence with 0 longest_match.
3466         * cfgcleanup.c (flow_find_head_matching_sequence): Count even
3467         non-active insns if !stop_after.
3468         (try_head_merge_bb): Revert 2014-01-07 changes.
3469
3470 2014-01-08  Jeff Law  <law@redhat.com>
3471
3472         * ree.c (get_sub_rtx): New function, extracted from...
3473         (merge_def_and_ext): Here.
3474         (combine_reaching_defs): Use get_sub_rtx.
3475
3476 2014-01-08  Eric Botcazou  <ebotcazou@adacore.com>
3477
3478         * cgraph.h (varpool_variable_node): Do not choke on null node.
3479
3480 2014-01-08  Catherine Moore  <clm@codesourcery.com>
3481
3482         * config/mips/mips.md (simple_return): Attempt to use JRC
3483         for microMIPS.
3484         * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
3485
3486 2014-01-08  Richard Sandiford  <rdsandiford@googlemail.com>
3487
3488         PR rtl-optimization/59137
3489         * reorg.c (steal_delay_list_from_target): Call update_block for
3490         elided insns.
3491         (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
3492
3493 2014-01-08  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
3494
3495         * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
3496         two duplicate entries.
3497
3498 2014-01-08  Richard Sandiford  <rdsandiford@googlemail.com>
3499
3500         Revert:
3501         2012-10-07  Richard Sandiford  <rdsandiford@googlemail.com>
3502
3503         * config/mips/mips.c (mips_truncated_op_cost): New function.
3504         (mips_rtx_costs): Adjust test for BADDU.
3505         * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
3506
3507         2012-10-02  Richard Sandiford  <rdsandiford@googlemail.com>
3508
3509         * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
3510         (*baddu_si): ...this new pattern.
3511
3512 2014-01-08  Jakub Jelinek  <jakub@redhat.com>
3513
3514         PR ipa/59722
3515         * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
3516
3517 2014-01-08  Bernd Edlinger  <bernd.edlinger@hotmail.de>
3518
3519         PR middle-end/57748
3520         * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
3521         inner_reference_p.
3522         (expand_expr, expand_normal): Adjust.
3523         * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
3524         inner_reference_p. Use inner_reference_p to expand inner references.
3525         (store_expr): Adjust.
3526         * cfgexpand.c (expand_call_stmt): Adjust.
3527
3528 2014-01-08  Rong Xu  <xur@google.com>
3529
3530         * gcov-io.c (gcov_var): Move from gcov-io.h.
3531         (gcov_position): Ditto.
3532         (gcov_is_error): Ditto.
3533         (gcov_rewrite): Ditto.
3534         * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
3535         only part to libgcc/libgcov.h.
3536
3537 2014-01-08  Marek Polacek  <polacek@redhat.com>
3538
3539         PR middle-end/59669
3540         * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
3541
3542 2014-01-08  Marek Polacek  <polacek@redhat.com>
3543
3544         PR sanitizer/59667
3545         * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
3546
3547 2014-01-08  Jakub Jelinek  <jakub@redhat.com>
3548
3549         PR rtl-optimization/59649
3550         * stor-layout.c (get_mode_bounds): For BImode return
3551         0 and STORE_FLAG_VALUE.
3552
3553 2014-01-08  Richard Biener  <rguenther@suse.de>
3554
3555         PR middle-end/59630
3556         * gimple.h (is_gimple_builtin_call): Remove.
3557         (gimple_builtin_call_types_compatible_p): New.
3558         (gimple_call_builtin_p): New overload.
3559         * gimple.c (is_gimple_builtin_call): Remove.
3560         (validate_call): Rename to ...
3561         (gimple_builtin_call_types_compatible_p): ... this and export.  Also
3562         check return types.
3563         (validate_type): New static function.
3564         (gimple_call_builtin_p): New overload and adjust.
3565         * gimple-fold.c (gimple_fold_builtin): Fold the return value.
3566         (gimple_fold_call): Likewise.  Use gimple_call_builtin_p.
3567         (gimple_fold_stmt_to_constant_1): Likewise.
3568         * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
3569
3570 2014-01-08  Richard Biener  <rguenther@suse.de>
3571
3572         PR middle-end/59471
3573         * gimplify.c (gimplify_expr): Gimplify register-register type
3574         VIEW_CONVERT_EXPRs to separate stmts.
3575
3576 2014-01-07  Jeff Law  <law@redhat.com>
3577
3578         PR middle-end/53623
3579         * ree.c (combine_set_extension): Handle case where source
3580         and destination registers in an extension insn are different.
3581         (combine_reaching_defs): Allow source and destination registers
3582         in extension to be different under limited circumstances.
3583         (add_removable_extension): Remove restriction that the
3584         source and destination registers in the extension are the same.
3585         (find_and_remove_re): Emit a copy from the extension's
3586         destination to its source after the defining insn if
3587         the source and destination registers are different.
3588
3589         PR middle-end/59285
3590         * ifcvt.c (merge_if_block): If we are merging a block with more than
3591         one successor with a block with no successors, remove any BARRIER
3592         after the second block.
3593
3594 2014-01-07  Dan Xio Qiang <ziyan01@163.com>
3595
3596         * hw-doloop.c (reorg_loops): Release the bitmap obstack.
3597
3598 2014-01-07  John David Anglin  <danglin@gcc.gnu.org>
3599
3600         PR target/59652
3601         * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
3602         for 14-bit register offsets when INT14_OK_STRICT is false.
3603
3604 2014-01-07  Roland Stigge  <stigge@antcom.de>
3605             Michael Meissner  <meissner@linux.vnet.ibm.com>
3606
3607         PR 57386/target
3608         * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
3609         Only check TFmode for SPE constants.  Don't check TImode or TDmode.
3610
3611 2014-01-07  James Greenhalgh  <james.greenhalgh@arm.com>
3612
3613         * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
3614         -mcpu.
3615
3616 2014-01-07  Yufeng Zhang  <yufeng.zhang@arm.com>
3617
3618         * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
3619         with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
3620         rtx is const0_rtx or not.
3621
3622 2014-01-07  Richard Sandiford  <rdsandiford@googlemail.com>
3623
3624         PR target/58115
3625         * target-globals.c (save_target_globals): Remove this_fn_optab
3626         handling.
3627         * toplev.c: Include optabs.h.
3628         (target_reinit): Temporarily restore the global options if another
3629         set of options are in force.
3630
3631 2014-01-07  Jakub Jelinek  <jakub@redhat.com>
3632
3633         PR rtl-optimization/58668
3634         * cfgcleanup.c (flow_find_cross_jump): Don't count
3635         any jumps if dir_p is NULL.  Remove p1 variable, use active_insn_p
3636         to determine what is counted.
3637         (flow_find_head_matching_sequence): Use active_insn_p to determine
3638         what is counted.
3639         (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
3640         counting change.
3641         * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
3642         determine what is counted.
3643
3644         PR tree-optimization/59643
3645         * tree-predcom.c (split_data_refs_to_components): If one dr is
3646         read and one write, determine_offset fails and the write isn't
3647         in the bad component, just put the read into the bad component.
3648
3649 2014-01-07  Mike Stump  <mikestump@comcast.net>
3650             Jakub Jelinek  <jakub@redhat.com>
3651
3652         PR pch/59436
3653         * tree-core.h (struct tree_optimization_option): Change optabs
3654         type from unsigned char * to void *.
3655         * optabs.c (init_tree_optimization_optabs): Adjust
3656         TREE_OPTIMIZATION_OPTABS initialization.
3657
3658 2014-01-06  Jakub Jelinek  <jakub@redhat.com>
3659
3660         PR target/59644
3661         * config/i386/i386.h (struct machine_function): Add
3662         no_drap_save_restore field.
3663         * config/i386/i386.c (ix86_save_reg): Use
3664         !cfun->machine->no_drap_save_restore instead of
3665         crtl->stack_realign_needed.
3666         (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
3667         this function clears frame_pointer_needed.  Set
3668         cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
3669         and DRAP reg is needed.
3670
3671 2014-01-06  Marek Polacek  <polacek@redhat.com>
3672
3673         PR c/57773
3674         * doc/implement-c.texi: Mention that other integer types are
3675         permitted as bit-field types in strictly conforming mode.
3676
3677 2014-01-06  Felix Yang  <fei.yang0953@gmail.com>
3678
3679         * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
3680         is newly allocated.
3681
3682 2014-01-06  Richard Earnshaw  <rearnsha@arm.com>
3683
3684         * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
3685
3686 2014-01-06  Martin Jambor  <mjambor@suse.cz>
3687
3688         PR ipa/59008
3689         * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
3690         to int.
3691         * ipa-prop.c (ipa_print_node_params): Fix indentation.
3692
3693 2014-01-06  Eric Botcazou  <ebotcazou@adacore.com>
3694
3695         PR debug/59350
3696         PR debug/59510
3697         * var-tracking.c (add_stores): Preserve the value of the source even if
3698         we don't record the store.
3699
3700 2014-01-06  Terry Guo  <terry.guo@arm.com>
3701
3702         * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
3703
3704 2014-01-05  Iain Sandoe  <iain@codesourcery.com>
3705
3706         PR bootstrap/59541
3707         * config/darwin.c (darwin_function_section): Adjust return values to
3708         correspond to optimisation changes made in r206070.
3709
3710 2014-01-05  Uros Bizjak  <ubizjak@gmail.com>
3711
3712         * config/i386/i386.c (ix86_data_alignment): Calculate max_align
3713         from prefetch_block tune setting.
3714         (nocona_cost): Correct size of prefetch block to 64.
3715
3716 2014-01-04  Eric Botcazou  <ebotcazou@adacore.com>
3717
3718         * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
3719         (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
3720         used to save the static chain register in the computation of the offset
3721         from which the FP registers need to be restored.
3722
3723 2014-01-04  Jakub Jelinek  <jakub@redhat.com>
3724
3725         PR tree-optimization/59519
3726         * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
3727         ICE if get_current_def (current_new_name) is already non-NULL, as long
3728         as it is a phi result of some other phi in *new_exit_bb that has
3729         the same argument.
3730
3731         * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
3732         or vmovdqu* for misaligned_operand.
3733         (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
3734         <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
3735         * config/i386/i386.c (ix86_expand_special_args_builtin): Set
3736         aligned_mem for AVX512F masked aligned load and store builtins and for
3737         non-temporal moves.
3738
3739 2014-01-03  Bingfeng Mei  <bmei@broadcom.com>
3740
3741         PR tree-optimization/59651
3742         * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
3743         Address range for negative step should be added by TYPE_SIZE_UNIT.
3744
3745 2014-01-03  Andreas Schwab  <schwab@linux-m68k.org>
3746
3747         * config/m68k/m68k.c (handle_move_double): Handle pushes with
3748         overlapping registers also for registers other than the stack pointer.
3749
3750 2014-01-03  Marek Polacek  <polacek@redhat.com>
3751
3752         PR other/59661
3753         * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
3754         __builtin_FILE.
3755
3756 2014-01-03  Jakub Jelinek  <jakub@redhat.com>
3757
3758         PR target/59625
3759         * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
3760         asm goto as jump.
3761
3762         * config/i386/i386.md (MODE_SIZE): New mode attribute.
3763         (push splitter): Use <P:MODE_SIZE> instead of
3764         GET_MODE_SIZE (<P:MODE>mode).
3765         (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
3766         (mov -1, reg peephole2): Likewise.
3767         * config/i386/sse.md (*mov<mode>_internal,
3768         <sse>_storeu<ssemodesuffix><avxsizesuffix>,
3769         <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
3770         *<code><mode>3, *andnot<mode>3<mask_name>,
3771         <mask_codefor><code><mode>3<mask_name>): Likewise.
3772         * config/i386/subst.md (mask_mode512bit_condition,
3773         sd_mask_mode512bit_condition): Likewise.
3774
3775 2014-01-02  Xinliang David Li  <davidxl@google.com>
3776
3777         PR tree-optimization/59303
3778         * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
3779         (dump_predicates): Better output format.
3780         (pred_equal_p): New function.
3781         (is_neq_relop_p): Ditto.
3782         (is_neq_zero_form_p): Ditto.
3783         (pred_expr_equal_p): Ditto.
3784         (pred_neg_p): Ditto.
3785         (simplify_pred): Ditto.
3786         (simplify_preds_2): Ditto.
3787         (simplify_preds_3): Ditto.
3788         (simplify_preds_4): Ditto.
3789         (simplify_preds): Ditto.
3790         (push_pred): Ditto.
3791         (push_to_worklist): Ditto.
3792         (get_pred_info_from_cmp): Ditto.
3793         (is_degenerated_phi): Ditto.
3794         (normalize_one_pred_1): Ditto.
3795         (normalize_one_pred): Ditto.
3796         (normalize_one_pred_chain): Ditto.
3797         (normalize_preds): Ditto.
3798         (normalize_cond_1): Remove function.
3799         (normalize_cond): Ditto.
3800         (is_gcond_subset_of): Ditto.
3801         (is_subset_of_any): Ditto.
3802         (is_or_set_subset_of): Ditto.
3803         (is_and_set_subset_of): Ditto.
3804         (is_norm_cond_subset_of): Ditto.
3805         (pred_chain_length_cmp): Ditto.
3806         (convert_control_dep_chain_into_preds): Type change.
3807         (find_predicates): Ditto.
3808         (find_def_preds): Ditto.
3809         (destroy_predicates_vecs): Ditto.
3810         (find_matching_predicates_in_rest_chains): Ditto.
3811         (use_pred_not_overlap_with_undef_path_pred): Ditto.
3812         (is_pred_expr_subset): Ditto.
3813         (is_pred_chain_subset_of): Ditto.
3814         (is_included_in): Ditto.
3815         (is_superset_of): Ditto.
3816
3817 2014-01-02  Richard Sandiford  <rdsandiford@googlemail.com>
3818
3819         Update copyright years.
3820
3821 2014-01-02  Richard Sandiford  <rdsandiford@googlemail.com>
3822
3823         * common/config/arc/arc-common.c, config/arc/arc-modes.def,
3824         config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
3825         config/arc/arc.md, config/arc/arc.opt,
3826         config/arm/arm_neon_builtins.def, config/arm/crypto.def,
3827         config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
3828         config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
3829         config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
3830         config/linux-protos.h, config/linux.c, config/winnt-c.c,
3831         diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
3832         vtable-verify.c, vtable-verify.h: Use the standard form for the
3833         copyright notice.
3834
3835 2014-01-02  Tobias Burnus  <burnus@net-b.de>
3836
3837         * gcc.c (process_command): Update copyright notice dates.
3838         * gcov-dump.c: Ditto.
3839         * gcov.c: Ditto.
3840         * doc/cpp.texi: Bump @copying's copyright year.
3841         * doc/cppinternals.texi: Ditto.
3842         * doc/gcc.texi: Ditto.
3843         * doc/gccint.texi: Ditto.
3844         * doc/gcov.texi: Ditto.
3845         * doc/install.texi: Ditto.
3846         * doc/invoke.texi: Ditto.
3847
3848 2014-01-01  Jan-Benedict Glaw  <jbglaw@lug-owl.de>
3849
3850         * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
3851
3852 2014-01-01  Jakub Jelinek  <jakub@redhat.com>
3853
3854         * config/i386/sse.md (*mov<mode>_internal): Guard
3855         EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
3856
3857         PR rtl-optimization/59647
3858         * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
3859         new_rtx into UNSIGNED_FLOAT rtxes.
3860 \f
3861 Copyright (C) 2014 Free Software Foundation, Inc.
3862
3863 Copying and distribution of this file, with or without modification,
3864 are permitted in any medium without royalty provided the copyright
3865 notice and this notice are preserved.