1 2017-03-03 Martin Liska <mliska@suse.cz>
3 PR tree-optimization/79803
4 * tree-ssa-loop-prefetch.c (tree_ssa_prefetch_arrays): Remove
6 (pass_loop_prefetch::execute): Disabled optimization if an
7 assumption about L1 cache size is not met.
9 2017-03-03 Martin Liska <mliska@suse.cz>
11 PR rtl-optimization/79574
12 * gcse.c (struct gcse_expr): Use HOST_WIDE_INT instead of int.
13 (hash_scan_set): Likewise.
14 (dump_hash_table): Likewise.
15 (hoist_code): Likewise.
17 2017-03-03 Richard Biener <rguenther@suse.de>
19 * fixed-value.c (fixed_from_string): Restore use of elt (1)
21 (fixed_convert_from_real): Likewise.
23 2017-03-03 Uros Bizjak <ubizjak@gmail.com>
26 * config/i386/i386.md (*pushxf_rounded): Use Pmode instead of DImode.
28 2017-03-03 Richard Biener <rguenther@suse.de>
31 * match.pd ( X +- C1 CMP C2 -> X CMP C2 -+ C1): Add missing
32 TYPE_OVERFLOW_UNDEFINED check.
34 2017-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
36 * config/rs6000/vector.md (vector_ne_<mode>_p): Correct operand
38 (vector_ae_<mode>_p): Likewise.
39 (vector_nez_<mode>_p): Likewise.
40 (vector_ne_v2di_p): Likewise.
41 (vector_ae_v2di_p): Likewise.
42 (vector_ne_<mode>_p): Likewise.
43 * config/rs6000/vsx.md (vsx_tsqrt<mode>2_fg): Correct operand
45 (vsx_tsqrt<mode>2_fe): Likewise.
47 2017-03-02 Uros Bizjak <ubizjak@gmail.com>
50 * config/i386/i386.md (*pushxf_rounded): New insn_and_split pattern.
52 2017-03-02 Jakub Jelinek <jakub@redhat.com>
54 PR rtl-optimization/79780
55 * cprop.c (one_cprop_pass): When second and further conditional trap
56 in a single basic block is turned into an unconditional trap, turn it
57 into a deleted note to avoid RTL verification failures.
59 2017-03-02 Richard Biener <rguenther@suse.de>
61 * fold-const.c (const_binop): Use ulow () instead of elt (0).
63 2017-03-02 Richard Biener <rguenther@suse.de>
65 PR tree-optimization/79345
67 * tree-ssa-alias.c (walk_aliased_vdefs_1): Take a limit
68 param and abort the walk, returning -1 if it is hit.
69 (walk_aliased_vdefs): Take a limit param and pass it on.
70 * tree-ssa-alias.h (walk_aliased_vdefs): Add a limit param,
71 defaulting to 0 and return a signed int.
72 * tree-ssa-uninit.c (struct check_defs_data): New struct.
73 (check_defs): New helper.
74 (warn_uninitialized_vars): Use walk_aliased_vdefs to warn
75 about uninitialized memory.
76 * fixed-value.c (fixed_from_string): Use ulow/uhigh to avoid
77 bogus uninitialized warning.
78 (fixed_convert_from_real): Likewise.
80 2017-03-02 Bin Cheng <bin.cheng@arm.com>
82 PR tree-optimization/66768
83 * tree-ssa-loop-ivopts.c (find_interesting_uses_address): Skip addr
84 iv_use if base object can't be determined.
86 2017-03-02 Jakub Jelinek <jakub@redhat.com>
88 PR tree-optimization/79345
89 * gensupport.h (struct pattern_stats): Add min_scratch_opno field.
90 * gensupport.c (get_pattern_stats_1) <case MATCH_SCRATCH>: Update it.
91 (get_pattern_stats): Initialize it.
92 * genemit.c (gen_expand): Verify match_scratch numbers come after
93 match_operand/match_dup numbers.
94 * config/i386/i386.md (<s>mul<mode>3_highpart): Swap match_dup and
95 match_scratch numbers.
96 * config/i386/sse.md (avx2_gathersi<mode>, avx2_gatherdi<mode>):
98 * config/s390/s390.md (trunctdsd2): Likewise.
100 2017-03-02 Richard Biener <rguenther@suse.de>
102 * wide-int.h (wide_int_storage::operator=): Implement in terms
105 2017-03-02 Richard Biener <rguenther@suse.de>
107 PR tree-optimization/79777
108 * tree-ssa-pre.c (eliminate_insert): Give up if we simplify
109 the to insert expression to sth existing.
111 2017-03-01 Martin Sebor <msebor@redhat.com>
114 * gimple-ssa-sprintf.c
115 (directive::known_width_and_precision): New function.
116 (format_integer): Use it.
117 (get_mpfr_format_length): Consider the full range of precision
118 when computing %g output with the # flag. Set the likely byte
119 count to 3 rather than 1 when precision is indeterminate.
120 (format_floating): Correct the lower bound of precision.
122 2017-03-01 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
124 * doc/invoke.texi: Document default code model for 64-bit Linux.
126 2017-03-01 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
129 * config/rs6000/rs6000.md (peephole2 for udiv/umod): Should emit
130 udiv rather than div since input pattern is unsigned.
132 2017-03-01 Uros Bizjak <ubizjak@gmail.com>
134 * config/i386/i386.c (print_reg): Warn for values of
135 unsupported size in integer register.
137 2017-03-01 Michael Meissner <meissner@linux.vnet.ibm.com>
140 * config/rs6000/predicates.md (current_file_function_operand): Do
141 not allow self calls to be local if the function is replaceable.
143 2017-03-01 Kelvin Nilsen <kelvin@gcc.gnu.org>
146 * config/rs6000/altivec.h (vec_ctz and others): Change the
147 preprocessor macro that controls conditional compilation from
148 _ARCH_PWR9 to __POWER9_VECTOR__.
149 (vec_all_ne): Change parameterization of __altivec_scalar_pred
150 macro expansion under preprocessor #ifdef __POWER9_VECTOR__
151 control (instead of _ARCH_PWR9 control) so that template
152 definition uses power9-specific function.
153 (vec_any_eq): Likewise.
154 (vec_all_ne): Change macro definition to use a power9-specific
155 expansion under #ifdef __POWER9_VECTOR__ control (instead of
157 (vec_any_eq) Likewise.
158 * config/rs6000/rs6000-builtin.def (CMPNEF): Remove BU_P9V_AV_2
159 expansion for CMPNEF to remove support for xvcmpnesp instruction.
160 (CMPNED): Remove BU_P9V_AV2 expansion for CMPNED to remove
161 support for xvcmpnedp instruction.
162 (VCMPNEB_P): Replace BU_P9V_AV_P macro expansion with BU_P9V_AV_2
163 macro expansion so that Power9 implementation of vec_all_ne does
164 not use the AltiVec predicate framework.
165 (VCMPNEH_P): Likewise.
166 (VCMPNEW_P): Likewise.
167 (VCMPNED_P): Likewise.
168 (VCMPNEFP_P): Likewise.
169 (VCMPNEDP_P): Likewise.
170 (VCMPAEB_P): Add BU_P9V_AV_2 macro expansion to change
171 implementation of vec_any_eq to not use AltiVec predicate
173 (VCMPAEH_P): Likewise.
174 (VCMPAEW_P): Likewise.
175 (VCMPAED_P): Likewise.
176 (VCMPAEFP_P): Likewise.
177 (VCMPAEDP_P): Likewise.
178 (VCMPNE_P): Replace BU_P9V_OVERLOAD_P macro expansion with
179 BU_P9V_OVERLOAD_2 so that Power9 implementation of vec_all_ne does
180 not use the AltiVec predicate framework.
181 (VCMPAE_P): Add BU_P9V_OVERLOAD_2 macro to change implementation
182 of vec_any_eq to not use AltiVec predicate framework.
183 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Add
184 support for predefined __POWER9_VECTOR__ macro to indicate that
185 Power9 instruction selection is enabled.
186 (altivec_overloaded_builtins): Remove extraneous
187 ALTIVEC_BUILTIN_VEC_CMPNE entry for overloaded
188 function argument types RS6000_BTI_bool_V16QI and
189 RS6000_BTI_bool_V16QI. Remove erroneous ALTIVEC_BUILTIN_VEC_CMPNE
190 entry for overloaded function argument types RS6000_BTI_bool_V4SI
191 andRS6000_BTI_bool_V4SI, mapping to P9V_BUILTIN_CMPNEB. Remove
192 two entries mapping to P9V_BUITIN_CMPNED and one entry mapping to
193 P9V_BUILTIN_CMPNEF to force use of instructions not specific to
194 Power9 for implementations of vec_cmpne. Change the signature for
195 all definitions of the overloaded P9V_BUILTIN_VEC_CMPNE_P function
196 (representing vec_all_ne) to remove the previously described first
197 argument of type RS6000_BTI_INTSI, as this was an artifact of
198 reliance on the AltiVec predicate framework, which is no longer
199 used in the implementation of these functions. Add
200 P9V_BUILTIN_VEC_VCMPAE_P entries (representing the vec_anyeq
201 function) to match all of the P9V_BUILTIN_VEC_VCMNE_P entries
202 since, unlike the AltiVec predicate framework implementation, we
203 do not share function descriptors between vec_alle and vec_anyeq.
204 (altivec_resolve_overloaded_builtin): Add SFmode and DFmode to the
205 set of modes that receive special treatment even when
206 TARGET_P9_VECTOR is true. The special treatment emits code that
207 does not depend on Power9 instructions.
208 * config/rs6000/vector.md (vector_ne_<mode>_p): Change this
209 define_expand to not rely on AltiVec predicate framework.
210 (vector_ae_<mode>p): New define_expand to represent vec_any_eq
212 (vector_ne_v2di_p): Change this define_expand to not rely on
213 AltiVec predicate framework.
214 (vector_ae_v2di_p): New define_expand to represent vec_any_eq
216 (vector_ne_<mode>_p): Change this define_expand to not rely on
217 AltiVec predicate framework.
218 (vector_ae_<mode>p): New define_expand to represent vec_any_eq
220 * config/rs6000/vsx.md (*vsx_ne_<mode>_p): For modes VSX_EXTRACT_I
221 (V16QI, V8HI, V4SI), correct a typo in the code emitted for this
223 (*vsx_ne_<mode>_p): For modes VSX_F (V4SF and V2DF), remove this
224 define_insn pattern because the xvcmpne<VSs>. instruction is not
226 (vcmpne<VSs>): Remove this define_insn because xvcmpne<VSs>
227 instruction is not supported.
229 2017-03-01 Jakub Jelinek <jakub@redhat.com>
231 * config/nvptx/nvptx.c: Include intl.h.
233 2017-03-01 Martin Jambor <mjambor@suse.cz>
236 * ipa-prop.h (ipa_bits): Removed field known.
237 (ipa_jump_func): Removed field vr_known. Changed fields bits and m_vr
238 to pointers. Adjusted their comments to warn about their sharing.
239 (ipcp_transformation_summary): Change bits to a vector of pointers.
240 (ipa_check_create_edge_args): Moved to ipa-prop.c, declare.
241 (ipa_get_ipa_bits_for_value): Declare.
242 * tree-vrp.h (value_range): Mark as GTY((for_user)).
243 * ipa-prop.c (ipa_bit_ggc_hash_traits): New.
244 (ipa_bits_hash_table): Likewise.
245 (ipa_vr_ggc_hash_traits): Likewise.
246 (ipa_vr_hash_table): Likewise.
247 (ipa_print_node_jump_functions_for_edge): Adjust for bits and m_vr
248 being pointers and vr_known being removed.
249 (ipa_set_jf_unknown): Likewise.
250 (ipa_get_ipa_bits_for_value): New function.
251 (ipa_set_jfunc_bits): Likewise.
252 (ipa_get_value_range): New overloaded functions.
253 (ipa_set_jfunc_vr): Likewise.
254 (ipa_compute_jump_functions_for_edge): Use the above functions to
255 construct bits and vr parts of jump functions.
256 (ipa_check_create_edge_args): Move here from ipa-prop.h, also allocate
257 ipa_bits_hash_table and ipa_vr_hash_table if they do not already
259 (ipcp_grow_transformations_if_necessary): Also allocate
260 ipa_bits_hash_table and ipa_vr_hash_table if they do not already
262 (ipa_node_params_t::duplicate): Do not copy bits, just pointers to
263 them. Fix too long lines.
264 (ipa_write_jump_function): Adjust for bits and m_vr being pointers and
265 vr_known being removed.
266 (ipa_read_jump_function): Use new setter functions to construct bits
267 and vr parts of jump functions or set them to NULL.
268 (write_ipcp_transformation_info): Adjust for bits being pointers.
269 (read_ipcp_transformation_info): Likewise.
270 (ipcp_update_bits): Likewise. Fix excessively long lines a trailing
272 Include gt-ipa-prop.h.
273 * ipa-cp.c (propagate_bits_across_jump_function): Adjust for bits
275 (ipcp_store_bits_results): Likewise.
276 (propagate_vr_across_jump_function): Adjust for m_vr being a pointer.
277 Do not write to existing jump functions but use a temporary instead.
279 2017-03-01 Jakub Jelinek <jakub@redhat.com>
282 * fold-const.c (make_bit_field_ref): If orig_inner is COMPONENT_REF,
283 attempt to use its first operand as BIT_FIELD_REF base.
285 2017-03-01 Richard Biener <rguenther@suse.de>
288 * tree-chrec.c (chrec_evaluate): Perform computation of Newtons
289 interpolating formula in wrapping arithmetic.
290 (chrec_apply): Convert chrec_evaluate return value to wanted type.
292 2017-03-01 Jakub Jelinek <jakub@redhat.com>
294 PR tree-optimization/79734
295 * tree-vect-generic.c (expand_vector_condition): Optimize
296 AVX512 vector boolean VEC_COND_EXPRs into bitwise operations.
297 Handle VEC_COND_EXPR where comparison has different inner width from
300 2017-02-28 Sandra Loosemore <sandra@codesourcery.com>
302 * doc/invoke.texi (ARC Options): Copy-edit to fix punctuation,
303 markup, and similar issues. Remove @opindex entries for things
304 that aren't options. Add missing -mmpy-option entries.
306 2017-02-28 Jakub Jelinek <jakub@redhat.com>
308 PR tree-optimization/79737
309 * gimple-ssa-store-merging.c (encode_tree_to_bitpos): If bitlen is
310 a multiple of BITS_PER_UNIT and !BYTES_BIG_ENDIAN, clear
311 tmpbuf[byte_size - 1]. Call natice_encode_expr with byte_size - 1
312 instead of byte_size. Formatting fix.
313 (shift_bytes_in_array_right): Formatting fix.
315 2017-02-28 Eric Botcazou <ebotcazou@adacore.com>
318 * config/sparc/sparc.c (sparc_frame_pointer_required): Add missing
319 condition on optimize for the leaf function test.
321 2017-02-28 Martin Liska <mliska@suse.cz>
324 * read-rtl-function.c (function_reader::handle_unknown_directive):
325 Bail out when one uses -flto.
327 2017-02-28 Martin Liska <mliska@suse.cz>
329 * common.opt: Replace space with tabular for options of <number>
331 * config/i386/i386.opt: Show <number> value for
332 -mlarge-data-threshold.
333 * opts.c (print_filtered_help): Do not display number in hexadecimal
336 2017-02-28 Martin Liska <mliska@suse.cz>
338 * common.opt: Fix --help=option -Q for options which are of
341 2017-02-28 Uros Bizjak <ubizjak@gmail.com>
343 * config/i386/i386.c (print_reg): Error out for values
344 of 8-bit size in invalid integer register.
346 2017-02-28 Martin Sebor <msebor@redhat.com>
348 PR tree-optimization/79691
349 * passes.def (pass_all_optimizations_g): Enable pass_sprintf_length.
351 2017-02-28 Jakub Jelinek <jakub@redhat.com>
354 * config/i386/i386.c (ix86_print_operand) <case 'R'>: Replace
355 gcc_unreachable with output_operand_lossage.
357 2017-02-28 Richard Biener <rguenther@suse.de>
359 PR tree-optimization/79740
360 * tree-ssa-sccvn.c (vn_nary_op_insert_into): Allow redundant
362 (visit_nary_op): Insert the nary into the hashtable if we
364 * tree-ssa-pre.c (eliminate_insert): Robustify.
366 2017-02-28 Richard Biener <rguenther@suse.de>
369 * fold-const.c (decode_field_reference): Reject out-of-bound
372 2017-02-28 Jakub Jelinek <jakub@redhat.com>
374 * config/i386/i386.c: Include intl.h.
375 (ix86_option_override_internal): Use cond ? G_("...") : G_("...")
376 instead of just cond ? "..." : "...".
377 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims): Likewise.
378 * coverage.c (read_counts_file): Likewise.
379 * omp-offload.c: Include intl.h.
380 (oacc_loop_fixed_partitions): Use cond ? G_("...") : G_("...") instead
381 of just cond ? "..." : "...".
382 * gcov.c (read_count_file): Use cond ? N_("...") : N_("...") instead
383 of just cond ? "..." : "...".
385 2017-02-28 Richard Earnshaw <rearnsha@arm.com>
388 * config/arm/parsecpu.awk (gen_data): Set tuning target to 'tune for'
390 * config/arm/arm-cpus.in (cortex-m0plus.small-multiply): Correct
392 * config/arm/arm-cpu-data.h: Regenerated.
394 2017-02-28 Richard Biener <rguenther@suse.de>
396 PR tree-optimization/79732
397 * tree-inline.c (expand_call_inline): Do not shadow var.
399 2017-02-28 Richard Biener <rguenther@suse.de>
401 PR tree-optimization/79723
402 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Preserve
403 address-space properly.
405 2017-02-28 Thomas Schwinge <thomas@codesourcery.com>
407 * doc/optinfo.texi (Optimization groups): Fix option used for
409 * doc/invoke.texi (-fopt-info): Document "omp".
410 * dumpfile.h: Sort OPTGROUP_OMP before OPTGROUP_VEC.
411 (OPTGROUP_ALL): Add OPTGROUP_OMP.
412 * hsa-gen.c (pass_data_gen_hsail): Use OPTGROUP_OMP.
413 * ipa-hsa.c (pass_data_ipa_hsa): Likewise.
414 * omp-simd-clone.c (pass_data_omp_simd_clone): Likewise.
416 * dumpfile.h (OPTGROUP_OPENMP): Rename to OPTGROUP_OMP. Adjust
418 * dumpfile.c (optgroup_options): Instead of "openmp", associate
419 OPTGROUP_OMP with "omp".
421 2017-02-27 Pat Haugen <pthaugen@us.ibm.com>
424 * config/rs6000/rs6000-c.c (struct altivec_builtin_types): Use VSRAD
425 for arithmetic shift of unsigned V2DI.
427 2017-02-27 Claudiu Zissulescu <claziss@synopsys.com>
429 * config.gcc (arc*-): Clean up, use arc/big.h, arc/elf.h, and
431 * config/arc/arc.h (TARGET_OS_CPP_BUILTINS): Remove.
432 (LINK_SPEC): Likewise.
433 (ARC_TLS_EXTRA_START_SPEC): Likewise.
434 (EXTRA_SPECS): Likewise.
435 (STARTFILE_SPEC): Likewise.
436 (ENDFILE_SPEC): Likewise.
437 (LIB_SPEC): Likewise.
438 (TARGET_SDATA_DEFAULT): Likewise.
439 (TARGET_MMEDIUM_CALLS_DEFAULT): Likewise.
440 (MULTILIB_DEFAULTS): Likewise.
441 (DWARF2_UNWIND_INFO): Likewise.
442 * config/arc/big.h: New file.
443 * config/arc/elf.h: Likewise.
444 * config/arc/linux.h: Likewise.
445 * config/arc/t-uClibc: Remove.
447 2017-02-27 Bin Cheng <bin.cheng@arm.com>
449 PR tree-optimization/77536
450 * tree-ssa-loop-manip.c (niter_for_unrolled_loop): New function.
451 (tree_transform_and_unroll_loop): Use above function to compute the
452 estimated niter of unrolled loop and use it when scaling profile.
453 Also use count info rather than frequency if it's non-zero.
454 * tree-ssa-loop-manip.h niter_for_unrolled_loop(): New declaration.
455 * tree-vect-loop.c (scale_profile_for_vect_loop): New function.
456 (vect_transform_loop): Call above function.
458 2017-02-27 Richard Biener <rguenther@suse.de>
460 PR tree-optimization/45397
461 * tree-ssa-pre.c (eliminate_insert): Handle BIT_AND_EXPR.
462 * tree-ssa-sccvn.c (valueized_wider_op): New helper.
463 (visit_nary_op): Add pattern matching for CSEing sign-changed
464 or truncated operations with wider ones.
466 2017-02-27 Richard Biener <rguenther@suse.de>
468 PR tree-optimization/79690
469 * tree-vect-stmts.c (vectorizable_store): Use vector type
470 built from the DR with address-space.
472 2017-02-26 Gerald Pfeifer <gerald@pfeifer.com>
474 * doc/invoke.texi (Optimize Options): Refine the description
475 of asan-use-after-return.
477 2017-02-25 Alan Modra <amodra@gmail.com>
479 PR rtl-optimization/79584
480 * lra-constraints.c (base_to_reg): Reload ad->base, the entire
481 base, not ad->base_term, the reg within base. Remove assertion
482 that ad->base == ad->base_term. Replace gen_int_mode using
483 bogus mode with const0_rtx.
485 2017-02-25 Jakub Jelinek <jakub@redhat.com>
488 * tree-eh.c (operation_could_trap_p, stmt_could_throw_1_p): Handle
489 FMA_EXPR like tcc_binary or tcc_unary.
491 * tree-ssa-loop-niter.c (number_of_iterations_exit): Simplify warning.
494 * dwarf2out.c (struct dw_loc_list_struct): Add noted_variable_value
496 (size_of_loc_descr): Handle DW_OP_GNU_variable_value.
497 (output_loc_operands): Handle DW_OP_call_ref and
498 DW_OP_GNU_variable_value.
499 (struct variable_value_struct): New type.
500 (struct variable_value_hasher): Likewise.
501 (variable_value_hash): New variable.
502 (string_types): Remove.
503 (copy_loc_descr): New function.
504 (add_loc_descr_to_each): Clarify comment. Use copy_loc_descr.
505 (prepend_loc_descr_to_each): New function.
506 (add_loc_list): Fix comment typo. Use prepend_loc_descr_to_each
507 instead of add_loc_descr_to_each if the first argument is single
508 location list and the second has multiple.
509 (resolve_args_picking_1): Handle DW_OP_GNU_variable_value.
510 (loc_list_from_tree_1): For early_dwarf, emit DW_OP_GNU_variable_value
511 when looking for variable value which doesn't have other location info.
512 (loc_list_from_tree): Formatting fix.
513 (gen_array_type_die): Simplify DW_AT_string_length handling.
514 (adjust_string_types): Remove.
515 (gen_subprogram_die): Don't call adjust_string_types nor test/set
516 string_types. Call resolve_variable_values.
517 (prune_unused_types_walk_loc_descr): Handle DW_OP_GNU_variable_value.
518 (resolve_addr_in_expr): Likewise. Add A argument.
519 (copy_deref_exprloc): Remove deref argument. Adjust for the
520 original expression being DW_OP_GNU_variable_value with optionally
521 DW_OP_stack_value after it instead of DW_OP_call4 with DW_OP_deref
523 (optimize_string_length): Rework for DW_OP_GNU_variable_value.
524 (resolve_addr): Adjust optimize_string_length and resolve_addr_in_expr
525 callers. Set remove_AT_byte_size if removing DW_AT_string_length.
526 (variable_value_hasher::hash, variable_value_hasher::equal): New
528 (resolve_variable_value_in_expr, resolve_variable_value,
529 resolve_variable_values, note_variable_value_in_expr,
530 note_variable_value): New functions.
531 (dwarf2out_early_finish): Call note_variable_value on all toplevel
534 2017-02-24 Jakub Jelinek <jakub@redhat.com>
537 * opts.h (handle_generated_option): Add GENERATED_P argument.
538 * opts-common.c (handle_option): Adjust function comment.
539 (handle_generated_option): Add GENERATED_P argument, pass it to
541 (control_warning_option): Pass false to handle_generated_option
543 * opts.c (maybe_default_option): Pass true to handle_generated_option
545 * optc-gen.awk: Likewise.
547 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
549 * config/sh/sh.md (tstsi_t): If operands[0] is a SUBREG instead of
550 a REG, look at the REG it is a SUBREG of.
551 (splitter for cmpeqsi_t): Ditto.
553 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
555 * config/pa/pa.c (pa_combine_instructions): Do not share RTL. Make
556 the special USEs with the pattern of the insn, not the insn itself.
558 2017-02-24 Matthew Fortune <matthew.fortune@imgtec.com>
561 * doc/invoke.texi: Document -mload-store-pairs.
563 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
564 Sandra Loosemore <sandra@codesourcery.com>
566 * config/nios2/nios2.c (nios2_simple_const_p): Returns false if the
567 argument isn't a CONST_INT.
568 (nios2_alternate_compare_const): Assert op is a CONST_INT.
569 (nios2_valid_compare_const_p): Assert op is a CONST_INT.
570 (nios2_validate_compare): Bypass alternate compare logic if *op2
572 (ldstwm_operation_p): Return false if first_base is not a REG or
573 if first_offset is not a CONST_INT.
575 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
577 * config/cris/cris.md: Use correct operand in a define_peephole2.
579 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
581 * config/c6x/c6x.c (predicate_insn): Do not incorrectly share RTL.
583 2017-02-24 Segher Boessenkool <segher@kernel.crashing.org>
585 * config/arc/arc.c (arc_ccfsm_advance): Only take the PATTERN of
586 this_insn if it is an INSN or JUMP_INSN.
587 (force_offsettable): Look at base, not at addr.
588 * config/arc/predicates.md (brcc_nolimm_operator): Don't call INTVAL
589 on things that aren't necessarily CONST_INTs.
591 2017-02-24 Uros Bizjak <ubizjak@gmail.com>
593 * doc/invoke.texi (x86 Options, -mfpmath=sse): Mention that
594 -mfpmath=sse is the default also for x86-32 targets with SSE2
595 instruction set when @option{-ffast-math} is enabled
597 2017-02-24 Jeff Law <law@redhat.com>
599 PR rtl-optimizatoin/79286
600 * ira.c (update_equiv_regs): Drop may_trap_p exception to
603 2017-02-24 Richard Biener <rguenther@suse.de>
605 PR tree-optimization/79389
606 * gimple-ssa-split-paths.c (is_feasible_trace): Properly skip
609 2017-02-24 Aldy Hernandez <aldyh@redhat.com>
611 * tree-ssa-loop-niter.c (number_of_iterations_exit): Update
612 function comment to reflect reality.
613 (loop_exits_before_overflow): Fix typo in function description.
615 2017-02-24 Richard Biener <rguenther@suse.de>
617 PR tree-optimization/79389
618 * gimple-ssa-split-paths.c (is_feasible_trace): Verify more
619 properly that a threading opportunity exists. Detect conditional
620 copy/constant propagation opportunities.
622 2017-02-23 Eric Botcazou <ebotcazou@adacore.com>
624 * config/visium/visium.md (type): Add trap.
625 (b): New mode attribute.
626 (*btst): Rename into...
627 (*btst<mode>): ...this and adjust.
628 (*cbranchsi4_btst_insn): Rename into...
629 (*cbranch<mode>4_btst_insn): ...this and adjust.
630 (trap): New define_insn.
632 2017-02-23 Jakub Jelinek <jakub@redhat.com>
634 PR tree-optimization/79389
635 * ifcvt.c (struct noce_if_info): Add rev_cond field.
636 (noce_reversed_cond_code): New function.
637 (noce_emit_store_flag): Use rev_cond if non-NULL instead of
638 reversed_comparison_code. Formatting fix.
639 (noce_try_store_flag): Test rev_cond != NULL in addition to
640 reversed_comparison_code.
641 (noce_try_store_flag_constants): Likewise.
642 (noce_try_store_flag_mask): Likewise.
643 (noce_try_addcc): Use rev_cond if non-NULL instead of
644 reversed_comparison_code.
645 (noce_try_cmove_arith): Likewise. Formatting fixes.
646 (noce_try_minmax, noce_try_abs): Clear rev_cond.
647 (noce_find_if_block): Initialize rev_cond.
648 (find_cond_trap): Call noce_get_condition with then_bb == trap_bb
649 instead of false as last argument never attempt to reverse it
652 2017-02-23 Bin Cheng <bin.cheng@arm.com>
654 PR tree-optimization/79663
655 * tree-predcom.c (combine_chains): Process refs in reverse order
656 only for ZERO length chains, and add explaining comment.
658 2017-02-23 Jeff Law <law@redhat.com>
660 PR tree-optimization/79578
661 * tree-ssa-dse.c (clear_bytes_written_by): Use OEP_ADDRESS_OF
662 in call to operand_equal_p.
664 2017-01-23 Dominique d'Humieres <dominiq@lps.ens.fr>
667 * config/i386/cpuid.h: Fix another undefined behavior.
669 2017-02-23 Richard Biener <rguenther@suse.de>
671 PR tree-optimization/79683
672 * tree-vect-stmts.c (vect_analyze_stmt): Do not overwrite
673 vector types for data-refs.
675 2017-02-23 Martin Liska <mliska@suse.cz>
677 * params.def (PARAM_MIN_NONDEBUG_INSN_UID): Change default to 0.
679 2017-02-23 Jakub Jelinek <jakub@redhat.com>
682 * internal-fn.c (get_range_pos_neg): Moved to ...
683 * tree.c (get_range_pos_neg): ... here. No longer static.
684 * tree.h (get_range_pos_neg): New prototype.
685 * expr.c (expand_expr_real_2) <case TRUNC_DIV_EXPR>: If both arguments
686 are known to be in between 0 and signed maximum inclusive, try to
687 expand both unsigned and signed divmod and use the cheaper one from
690 2017-02-22 Jeff Law <law@redhat.com>
692 PR tree-optimization/79578
693 * tree-ssa-dse.c (clear_bytes_written_by): Use operand_equal_p
694 to compare base operands.
696 2017-02-22 Segher Boessenkool <segher@kernel.crashing.org>
699 * config/rs6000/rs6000.md (*fsel<SFDF:mode><SFDF2:mode>4): Use
700 gpc_reg_operand instead of fpr_reg_operand.
702 2017-02-22 Sameera Deshpande <sameera.deshpande@imgtec.com>
704 * config/mips/mips.c (mips_return_in_memory): Force FP
705 vector types to be returned in memory for o32 ABI.
707 2017-02-22 Jakub Jelinek <jakub@redhat.com>
709 * dwarf2out.c (gen_variable_die): For -gdwarf-5, use DW_TAG_variable
710 instead of DW_TAG_member for static data member declarations and don't
711 set no_linkage_name for static inline data members.
712 (gen_member_die): For -gdwarf-5 don't change DW_TAG_variable
715 2017-02-22 Martin Liska <mliska@suse.cz>
717 * doc/invoke.texi: Replace inequality signs with square brackets
720 2017-02-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
722 PR tree-optimization/68644
723 * gcc.dg/tree-ssa/ivopts-lt-2.c: Skip for powerpc*-*-*.
725 2017-02-22 Matthew Fortune <matthew.fortune@imgtec.com>
728 * lra-constraints.c (simplify_operand_subreg): Handle
729 WORD_REGISTER_OPERATIONS targets.
731 2017-02-22 Jakub Jelinek <jakub@redhat.com>
734 * reg-stack.c (emit_swap_insn): Treat (float_extend:?F (mem:?F))
735 and (const_double:?F) like (mem:?F) for the purpose of fxch %st(1)
736 elimination by swapping fld*.
738 2017-02-22 Richard Biener <rguenther@suse.de>
740 PR tree-optimization/79673
741 * tree-ssa-pre.c (compute_avail): Use wide_int_to_tree to
742 convert the [TARGET_]MEM_REF offset INTEGER_CST, scrapping off
743 irrelevant address-space qualifiers and avoiding a
744 ADDR_SPACE_CONVERT_EXPR from fold_convert.
746 2017-02-22 Richard Biener <rguenther@suse.de>
748 PR tree-optimization/79666
749 * tree-vrp.c (extract_range_from_binary_expr_1): Make sure
750 to not symbolically negate if that may introduce undefined
753 2017-02-22 Martin Liska <mliska@suse.cz>
756 * data-streamer-in.c (streamer_read_gcov_count): Remove assert.
757 * data-streamer-out.c (streamer_write_gcov_count_stream):
759 * value-prof.c (stream_out_histogram_value): Make assert more
760 precise based on type of counter.
762 2017-02-21 Uros Bizjak <ubizjak@gmail.com>
765 * config/i386/i386.md (standard_x87sse_constant_load splitter):
766 Use nonimmediate_operand instead of memory_operand for operand 1.
767 (float-extend standard_x87sse_constant_load splitter): Ditto.
769 2017-02-21 Jeff Law <law@redhat.com>
771 PR tree-optimization/79621
772 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behavior): Ignore
773 blocks with edges to themselves.
775 2017-02-21 Jakub Jelinek <jakub@redhat.com>
778 * tree-chkp-opt.c (chkp_optimize_string_function_calls): Use
779 is_gimple_call instead of comparing gimple_code with GIMPLE_CALL.
780 Use gimple_call_builtin_p.
783 * sel-sched.c (moveup_expr_cached): Don't call sel_bb_head
784 on temporarily removed DEBUG_INSNs.
786 PR tree-optimization/79649
787 * tree-loop-distribution.c (classify_partition): Give up on
788 non-generic address space loads/stores.
790 2017-02-21 Aldy Hernandez <aldyh@redhat.com>
792 * doc/loop.texi (Loop manipulation): Remove nonexistent
793 tree_ssa_loop_version from the documentation.
794 * cfgloopmanip.c (loop_version): Document CONDITION_BB argument.
796 2017-02-21 Jakub Jelinek <jakub@redhat.com>
799 * config/i386/i386.c (ix86_expand_split_stack_prologue): Call
800 make_reg_eh_region_note_nothrow_nononlocal on call_insn.
801 * config/rs6000/rs6000.c: Include except.h.
802 (rs6000_expand_split_stack_prologue): Call
803 make_reg_eh_region_note_nothrow_nononlocal on the call insn.
805 2017-02-21 Martin Jambor <mjambor@suse.cz>
808 * ipa-prop.c (ipa_prop_write_jump_functions): Bail out if no edges
811 2017-02-21 Martin Jambor <mjambor@suse.cz>
813 * common.opt (-fipa-cp-alignment): Mark as ignored and preserved
814 for backward compatibility only.
815 * doc/invoke.texi (Option Summary): Remove all references to
818 2017-02-21 Matthew Fortune <matthew.fortune@imgtec.com>
822 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
824 * lra-constraints.c (curr_insn_transform): Handle
825 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
827 2017-02-21 Martin Liska <mliska@suse.cz>
829 * config/i386/i386.opt: Replace -masm-dialect with -masm.
831 2017-02-21 Thomas Schwinge <thomas@codesourcery.com>
834 * config/nvptx/nvptx.c (ENTRY_TEMPLATE): Single out "%ntid.y".
836 2017-02-21 Eric Botcazou <ebotcazou@adacore.com>
839 * config/arm/arm.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define.
840 (arm_function_ok_for_sibcall): Return false for an indirect call by
841 descriptor if all the argument registers are used.
842 (arm_relayout_function): Use FUNCTION_ALIGNMENT macro to adjust the
843 alignment of the function.
845 2017-02-21 Jakub Jelinek <jakub@redhat.com>
847 PR tree-optimization/61441
848 * simplify-rtx.c (simplify_const_unary_operation): For
849 -fsignaling-nans and sNaN operand, return NULL_RTX rather than
852 2017-02-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
854 * Makefile.in (BUILD_SYSTEM_HEADER_DIR): New make variabe.
855 (LIMITS_H_TEST, if_multiarch, stmp-fixinc): Use BUILD_SYSTEM_HEADER_DIR
856 instead of SYSTEM_HEADER_DIR.
858 2017-02-20 Gerald Pfeifer <gerald@pfeifer.com>
859 Martin Liška <mliska@suse.cz>
861 * doc/invoke.texi (use-after-scope-direct-emission-threshold):
862 Fix typos and grammar, use active voice, and clarify.
864 2017-02-20 Marek Polacek <polacek@redhat.com>
867 * gimplify.c (gimplify_expr): Handle unused *&&L;.
870 * ubsan.c (ubsan_type_descriptor): Check if TYPE_MAX_VALUE is null.
872 2017-02-20 Jakub Jelinek <jakub@redhat.com>
875 * config/i386/i386.c (ix86_expand_builtin): Handle
876 OPTION_MASK_ISA_AVX512VL and OPTION_MASK_ISA_64BIT in
877 ix86_builtins_isa[fcode].isa as a requirement of those
878 flags and any other flag in the bitmask.
879 (ix86_init_mmx_sse_builtins): Use 0 instead of
880 ~OPTION_MASK_ISA_64BIT as mask.
881 * config/i386/i386-builtin.def (__builtin_ia32_rdtsc,
882 __builtin_ia32_rdtscp, __builtin_ia32_pause, __builtin_ia32_bsrsi,
883 __builtin_ia32_rdpmc, __builtin_ia32_rolqi, __builtin_ia32_rolhi,
884 __builtin_ia32_rorqi, __builtin_ia32_rorhi): Likewise.
886 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
889 * lra-constraints.c (split_reg): Check requested split mode
890 is supported by the register.
892 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
894 * lra-constraints.c (simplify_operand_subreg): Remove early
897 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
900 * lra-constraints.c (curr_insn_transform): Tighten condition
901 for converting SUBREG reloads from OP_OUT to OP_INOUT.
903 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
906 * lra-constraints.c (curr_insn_transform): Handle
907 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
909 2017-02-19 Uros Bizjak <ubizjak@gmail.com>
912 2016-05-30 Uros Bizjak <ubizjak@gmail.com>
914 * config/i386/sync.md (mfence_nosse): Use "lock orl $0, -4(%esp)".
916 2017-02-19 Jonathan Wakely <jwakely@redhat.com>
919 * doc/invoke.texi (C++ Dialect Options) [-Wliteral-suffix]: Update
922 2017-02-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
924 * gimple-pretty-print.c (dump_ternary_rhs): Adjust gimple dump format
927 2017-02-18 Jakub Jelinek <jakub@redhat.com>
929 * final.c (last_columnnum, override_columnnum): New variables.
930 (final_start_function): Set last_columnnum, pass it to begin_prologue
931 hook and pass 0 to dwarf2out_begin_prologue.
932 (final_scan_insn): Update override_columnnum. Pass last_columnnum
933 to source_line debug hook.
934 (notice_source_line): Compute last_columnnum and for debug_column_info
935 return true on column changes.
936 * debug.h (struct gcc_debug_hooks): Add column argument to
937 source_line and begin_prologue hooks.
938 (debug_nothing_int_charstar_int_bool): Remove prototype.
939 (debug_nothing_int_int_charstar,
940 debug_nothing_int_int_charstar_int_bool): New prototypes.
941 (dwarf2out_begin_prologue): Add column argument.
942 * debug.c (do_nothing_debug_hooks): Adjust source_line and
943 begin_prologue hooks.
944 (debug_nothing_int_charstar_int_bool): Remove.
945 (debug_nothing_int_int_charstar,
946 debug_nothing_int_int_charstar_int_bool): New functions.
947 * dwarf2out.c (dwarf2out_begin_prologue): Add column argument, pass it
948 through to dwarf2out_source_line.
949 (dwarf2_lineno_debug_hooks): Adjust begin_prologue hook.
950 (dwarf2out_source_line): Add column argument, emit it if requested.
951 * sdbout.c (sdbout_source_line, sdbout_begin_prologue): Add column
953 * xcoffout.h (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
954 * xcoffout.c (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
955 * vmsdbgout.c (vmsdbgout_begin_prologue): Add column argument, pass it
956 through to dwarf2out_begin_prologue.
957 (vmsdbgout_source_line): Add column argument, pass it through to
958 dwarf2out_source_line.
959 * dbxout.c (dbxout_begin_prologue): Add column argument, adjust
960 dbxout_source_line caller.
961 (dbxout_source_line): Add column argument.
963 * common.opt (gno-column-info, gcolumn-info): New options.
964 * dwarf2out.c (dwarf2_lineno_debug_hooks): Formatting fix.
965 (check_die): Also test for multiple DW_AT_decl_column attributes.
966 (add_src_coords_attributes, dwarf2out_imported_module_or_decl_1): Add
967 DW_AT_decl_column if requested.
968 (gen_subprogram_die): Compare and/or add also DW_AT_decl_column
970 (gen_variable_die): Likewise.
971 (add_call_src_coords_attributes): Add DW_AT_call_column if requested.
972 * doc/invoke.texi (-gcolumn-info, -gno-column-info): Document.
975 * config/i386/i386.opt (m3dnowa): Replace Undocumented with Report.
976 * common/config/i386/i386-common.c (OPTION_MASK_ISA_3DNOW_A_SET): Define.
977 (ix86_handle_option): Handle OPT_m3dnowa.
978 * doc/invoke.texi (-m3dnowa): Document.
979 * doc/extend.texi (__builtin_ia32_pmulhuw, __builtin_ia32_pf2iw): Use
980 -m3dnowa instead of -m3dnow -march=athlon.
983 * config/i386/i386.c (ix86_print_operand): Use output_operand_lossage
984 instead of gcc_assert for K, r and R code checks. Formatting fixes.
986 2017-02-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
989 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
990 support for CODE_FOR_vsx_xxpermdi_v2d[fi]_be.
991 * config/rs6000/rs6000.md (reload_gpr_from_vsx<mode>): Call
992 generator for vsx_xxpermdi_<mode>_be.
993 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Remove logic to
994 force big-endian semantics.
995 (vsx_xxpermdi_<mode>_be): New define_expand with same
996 implementation as previous version of vsx_xxpermdi_<mode>.
998 2017-02-17 Jakub Jelinek <jakub@redhat.com>
1000 PR tree-optimization/79327
1001 * gimple-ssa-sprintf.c (format_integer): Remove likely_adjust
1002 variable, its initialization and use.
1004 2017-02-17 Julia Koval <julia.koval@intel.com>
1006 * common/config/i386/i386-common.c (OPTION_MASK_ISA_RDPID_SET): New.
1007 (OPTION_MASK_ISA_PKU_UNSET): New.
1008 (ix86_handle_option): Handle -mrdpid.
1009 * config/i386/cpuid.h (bit_RDPID): New.
1010 * config/i386/driver-i386.c (host_detect_local_cpu):
1011 Detect RDPID feature.
1012 * config/i386/i386-builtin.def (__builtin_ia32_rdpid): New.
1013 * config/i386/i386-c.c (ix86_target_macros_internal):
1015 * config/i386/i386.c (ix86_target_string): Add -mrdpid to isa2_opts.
1016 (ix86_valid_target_attribute_inner_p): Add "rdpid".
1017 (ix86_expand_builtin): Handle IX86_BUILTIN_RDPID.
1018 * config/i386/i386.h (TARGET_RDPID, TARGET_RDPID_P): New.
1019 * config/i386/i386.md (define_insn "rdpid"): New.
1020 * config/i386/i386.opt Add -mrdpid.
1021 * config/i386/immintrin.h (_rdpid_u32): New.
1023 2017-02-17 Vladimir Makarov <vmakarov@redhat.com>
1025 PR rtl-optimization/79541
1026 * lra-constraints.c (curr_insn_transform): Remove wrong asm insn
1027 instead of transforming it into USE.
1029 2017-02-17 Segher Boessenkool <segher@kernel.crashing.org>
1031 * config/rs6000/rs6000.md (extendsfdf2): Remove default arguments.
1032 If HONOR_SNANS (SFmode) force the input to a register.
1033 (*extendsfdf2_fpr): Add !HONOR_SNANS (SFmode) condition.
1034 (*extendsfdf2_snan): New pattern, used when using SNaNs; it generates
1035 an frsp or similar insn.
1037 2017-02-17 Martin Liska <mliska@suse.cz>
1039 PR rtl-optimization/79577
1040 * params.def (selsched-max-sched-times): Increase minimum to 1.
1042 2017-02-17 Martin Liska <mliska@suse.cz>
1044 PR rtl-optimization/79574
1045 * gcse.c (want_to_gcse_p): Prevent integer overflow.
1047 2017-02-17 Martin Liska <mliska@suse.cz>
1049 PR tree-optimization/79529
1050 * tree-ssa-loop-unswitch.c (is_maybe_undefined): Use
1051 ssa_defined_default_def_p to handle cases which are implicitly
1053 * tree-ssa.c (ssa_defined_default_def_p): New function.
1054 (ssa_undefined_value_p): Use ssa_defined_default_def_p to handle cases
1055 which are implicitly defined.
1056 * tree-ssa.h (ssa_defined_default_def_p): Declare.
1058 2017-02-17 Richard Biener <rguenther@suse.de>
1061 * params.def (max-ssa-name-query-depth): Limit to 10.
1063 2017-02-17 Richard Biener <rguenther@suse.de>
1065 PR tree-optimization/79552
1066 * tree-ssa-structalias.c (visit_loadstore): Properly verify
1069 2017-02-17 Richard Biener <rguenther@suse.de>
1072 * genmatch.c (output_line_directive): Handle DIR_SEPARATOR_2.
1074 2017-02-17 Marek Polacek <polacek@redhat.com>
1077 * fold-const.c (fold_negate_expr_1): Renamed from fold_negate_expr.
1078 (fold_negate_expr): New wrapper.
1080 2017-02-16 Sandra Loosemore <sandra@codesourcery.com>
1082 * doc/invoke.texi (C++ Dialect Options) [-Wno-non-template-friend]:
1083 Correct terminology and de-emphasize pre-standard behavior.
1085 2017-02-16 Alan Modra <amodra@gmail.com>
1087 PR rtl-optimization/79286
1088 * ira.c (def_dominates_uses): New function.
1089 (update_equiv_regs): Don't create an equivalence for insns that
1090 may trap where the register def does not dominate the use.
1092 2017-02-16 Vladimir Makarov <vmakarov@redhat.com>
1094 PR rtl-optimization/78127
1095 * lra.c (lra): Call lra_eliminate before finish the loop after
1098 2017-02-16 Richard Biener <rguenther@suse.de>
1100 * graphite.h: Do not include isl/isl_val_gmp.h, instead include
1102 * graphite-isl-ast-to-gimple.c (gmp_cst_to_tree): Remove.
1103 (gcc_expression_from_isl_expr_int): Use generic isl_val interface.
1104 * graphite-sese-to-poly.c: Do not include isl/isl_val_gmp.h.
1105 (isl_val_int_from_wi): New function.
1106 (extract_affine_gmp): Rename to ...
1107 (extract_affine_wi): ... this, take a widest_int.
1108 (extract_affine_int): Just wrap extract_affine_wi.
1109 (add_param_constraints): Use isl_val_int_from_wi.
1110 (add_loop_constraints): Likewise, and extract_affine_wi.
1112 2017-02-15 Jeff Law <law@redhat.com>
1115 * ira-costs.c (scan_one_insn): Check have_regs_of_mode before calling
1116 ira_init_register_move_cost_if_necessary.
1118 2017-02-15 Martin Sebor <msebor@redhat.com>
1121 * doc/invoke.texi (-fdump-final-insns): Replace option accidentally
1122 removed in a prior commit.
1124 2017-02-15 Bin Cheng <bin.cheng@arm.com>
1126 PR tree-optimization/79347
1127 * tree-vect-loop-manip.c (vect_do_peeling): Maintain profile
1128 counters during peeling.
1130 2017-02-15 Thomas Schwinge <thomas@codesourcery.com>
1132 * Makefile.in (site.exp): Remove "set ISLVER".
1134 2017-02-15 Jakub Jelinek <jakub@redhat.com>
1137 * real.c (real_from_integer): Call real_convert even for decimal.
1139 2017-02-15 Dominik Vogt <vogt@linux.vnet.ibm.com>
1142 * config/s390/s390.c: define TARGET_CUSTOM_FUNCTION_DESCRIPTORS.
1144 2017-02-14 Andrew Pinski <apinski@cavium.com>
1146 * config/aarch64/aarch64-cores.def (thunderx2t99): Move to under 'C"
1147 cores and change the partno/implementer to be correct.
1148 (thunderx2t99p1): New core which replaces thunderx2t99 and still has
1149 the 'B" as the implementer.
1150 * config/aarch64/aarch64-tune.md: Regenerate.
1152 2017-02-14 Carl Love <cel@us.ibm.com>
1154 * config/rs6000/rs6000.c: Add case statement entry to make the
1155 xvcvuxdsp built-in argument unsigned.
1156 * config/rs6000/vsx.md: Fix the source and return operand types so they
1157 match the instruction definitions from the ISA document. Fix typo
1158 in the instruction generation for the (define_insn "vsx_xvcvuxdsp"
1161 2017-02-14 Vladimir Makarov <vmakarov@redhat.com>
1164 * lra-int.h (struct lra_operand_data, struct lra_insn_reg): Add
1165 member early_clobber_alts.
1166 * lra-lives.c (reg_early_clobber_p): New.
1167 (process_bb_lives): Use it.
1168 * lra.c (new_insn_reg): New arg early_clobber_alts. Use it.
1169 (debug_operand_data): Initialize early_clobber_alts.
1170 (setup_operand_alternative): Set up early_clobber_alts.
1171 (collect_non_operand_hard_regs): Ditto. Pass early clobber
1172 alternatives to new_insn_reg.
1173 (add_regs_to_insn_regno_info): Add arg early_clobber_alts. Use
1175 (lra_update_insn_regno_info): Pass the new arg.
1177 2017-02-14 Jakub Jelinek <jakub@redhat.com>
1180 * omp-offload.c (free_oacc_loop): Release loop->ifns vector.
1181 (new_oacc_loop_raw): Don't clear already cleared fields.
1184 * config/i386/avx512pfintrin.h (_mm512_prefetch_i32gather_pd,
1185 _mm512_prefetch_i32gather_ps, _mm512_prefetch_i64gather_pd,
1186 _mm512_prefetch_i64gather_ps): New inline functions and macros.
1188 2017-02-14 Uros Bizjak <ubizjak@gmail.com>
1191 * config/i386/i386.md (*movxf_internal): Add (o,rC) alternative.
1193 2017-02-14 H.J. Lu <hongjiu.lu@intel.com>
1196 * config/i386/i386.c (timode_scalar_chain::convert_insn): Insert
1197 the extra instruction to the right place to store 128-bit constant
1200 2017-02-14 Martin Sebor <msebor@redhat.com>
1203 * gimple-ssa-sprintf.c (format_directive): Avoid issuing INT_MAX
1204 warning for strings of unknown length.
1206 2017-02-13 Segher Boessenkool <segher@kernel.crashing.org>
1208 * config.gcc (supported_defaults) [powerpc*-*-*]: Update.
1210 2017-02-14 Jeff Law <law@redhat.com>
1213 * ira-costs.c (scan_one_insn): Initialize register move costs
1214 for pseudos seen in USE/CLOBBER insns.
1216 PR tree-optimization/79095
1217 * tree-vrp.c (extract_range_from_binary_expr_1): For EXACT_DIV_EXPR,
1218 if the numerator has the range ~[0,0] make the resultant range ~[0,0].
1219 (extract_range_from_binary_expr): For MINUS_EXPR with no derived range,
1220 if the operands are known to be not equal, then the resulting range
1222 (intersect_ranges): If the new range is ~[0,0] and the old range is
1223 wide, then prefer ~[0,0].
1224 * tree-vrp.c (overflow_comparison_p_1): New function.
1225 (overflow_comparison_p): New function.
1226 * tree-vrp.c (register_edge_assert_for_2): Register additional asserts
1227 if NAME is used in an overflow test.
1228 (vrp_evaluate_conditional_warnv_with_ops): If the ops represent an
1229 overflow check that can be expressed as an equality test, then adjust
1230 ops to be that equality test.
1232 2017-02-14 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1234 * config/s390/s390-builtin-types.def: Remove flags argument.
1235 * config/s390/s390.c (s390_init_builtins): Likewise.
1237 2017-02-14 Martin Liska <mliska@suse.cz>
1239 * tree-ssa-loop-unswitch.c (hoist_guard): Release get_loop_body
1240 vector. Fix trailing white spaces.
1242 2017-02-14 James Greenhalgh <james.greenhalgh@arm.com>
1244 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Handle
1247 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1249 PR rtl-optimization/68664
1250 * config/arm/arm.c (arm_sched_can_speculate_insn):
1251 New function. Declare prototype.
1252 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
1254 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1256 PR rtl-optimization/68664
1257 * config/aarch64/aarch64.c (aarch64_sched_can_speculate_insn):
1259 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
1261 2017-02-14 Amit Pawar <amit.pawar@amd.com>
1263 * config/i386/i386.c (znver1_cost): Fix the alignment for function and
1264 max skip bytes for function, loop and jump.
1266 2017-02-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
1268 * gimple-pretty-print.c (dump_unary_rhs): Adjust dump format for
1269 ABS_EXPR for gimple dump.
1271 2017-02-14 Jakub Jelinek <jakub@redhat.com>
1274 * config/sh/sh.c (expand_cbranchdi4): Don't clear operands[4].
1276 PR tree-optimization/79408
1277 * tree-vrp.c (simplify_div_or_mod_using_ranges): Handle also the
1278 case when on TRUNC_MOD_EXPR op0 is INTEGER_CST.
1279 (simplify_stmt_using_ranges): Call simplify_div_or_mod_using_ranges
1280 also if rhs1 is INTEGER_CST.
1282 2017-02-14 Richard Biener <rguenther@suse.de>
1285 * tree-into-ssa.c (insert_phi_nodes): When the function can
1286 have abnormal edges rewrite SSA names with broken use-def
1287 dominance out of SSA and register them for PHI insertion.
1289 2017-02-13 Martin Sebor <msebor@redhat.com>
1292 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call): Avoid
1293 clearing info.nowrite flag when snprintf size argument is a range.
1295 2017-02-13 Jakub Jelinek <jakub@redhat.com>
1297 * cprop.c (cprop_jump): Add missing space in string literal.
1298 * tree-ssa-structalias.c (rewrite_constraints): Likewise.
1299 (get_constraint_for_component_ref): Likewise.
1300 * df-core.c (df_worklist_dataflow_doublequeue): Likewise.
1301 * tree-outof-ssa.c (insert_partition_copy_on_edge): Likewise.
1302 * lra-constraints.c (process_alt_operands): Likewise.
1303 * ipa-inline.c (inline_small_functions): Likewise.
1304 * tree-ssa-sccvn.c (visit_reference_op_store): Likewise.
1305 * cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
1306 * trans-mem.c (diagnose_tm_1_op): Likewise.
1307 * omp-grid.c (grid_find_single_omp_among_assignments): Likewise.
1308 (grid_parallel_clauses_gridifiable): Likewise.
1310 * config/nvptx/mkoffload.c (process): Add space in between
1313 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
1314 "MOD4_SSE_REGS" and "ALL_REGS".
1316 * spellcheck.c (test_data): Add , in between "foo" and "food".
1318 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1321 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
1322 boundary crossing check and subsequent code generation agree.
1324 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1326 * config/aarch64/aarch64.c (has_memory_op): Delete.
1327 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
1330 2017-02-13 Jakub Jelinek <jakub@redhat.com>
1332 PR rtl-optimization/79388
1333 PR rtl-optimization/79450
1334 * combine.c (distribute_notes): When removing TEM_INSN for which
1335 corresponding dest has last value recorded, invalidate that last
1338 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1340 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
1341 of explicit '@'. Add missing assembly comment marker on branch costs
1344 2017-02-13 Nathan Sidwell <nathan@acm.org>
1346 * gengtype-lex.l (<in_struct>): Add '/'.
1348 2017-02-13 Martin Liska <mliska@suse.cz>
1351 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
1353 2017-02-13 Richard Biener <rguenther@suse.de>
1355 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
1357 * configure: Re-generate.
1358 * config.in: Likewise.
1359 * graphite-dependences.c: Simplify as if
1360 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
1361 * graphite-isl-ast-to-gimple.c: Likewise.
1362 * graphite-optimize-isl.c: Likewise.
1363 * graphite-poly.c: Likewise.
1364 * graphite-sese-to-poly.c: Likewise.
1365 * graphite.h: Likewise.
1366 * toplev.c: Include isl/version.h and use isl_version () for
1367 printing the ISL version.
1368 * doc/install.texi: Update ISL requirement.
1370 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
1372 * doc/standards.texi (Standards): Update reference to
1375 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
1377 * doc/extend.texi (Named Address Spaces): sourceware.org now
1379 * doc/install.texi (Binaries): Ditto.
1382 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
1384 * doc/cpp.texi: Replace "stringify"/"stringification" with C
1385 standard terminology "stringize"/"stringizing" throughout.
1386 * doc/cppinternals.texi: Likewise.
1388 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
1390 * doc/extend.texi: Fix some spelling mistakes and typos.
1391 * doc/invoke.texi: Likewise.
1393 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
1396 * params.def (inline-min-speedup) Change from 10 to 8.
1398 2017-02-11 Jakub Jelinek <jakub@redhat.com>
1400 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
1403 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
1406 * ipa-inline-analysis.c (get_minimal_bb): New function.
1407 (record_modified): Use it.
1408 (remap_edge_change_prob): Handle also ancestor functions.
1410 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
1412 * doc/contrib.texi (Contributors): Remove broken link into
1413 the Mauve CVS repository.
1415 2017-02-11 Jakub Jelinek <jakub@redhat.com>
1418 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
1419 result computation whenever lhs doesn't have vector mode, not
1420 just when it has BLKmode.
1422 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
1424 * doc/makefile.texi (profiledbootstrap): Refer to the
1425 installation instructions only in textual form.
1427 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1430 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
1432 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
1434 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
1435 (Specific): Update mingw-w64 reference.
1437 (Specific): Remove broken link to Renesas RX processor.
1439 2017-02-10 Richard Biener <rguenther@suse.de>
1441 * toplev.c (process_options): Do not mention obsolete graphite
1442 options when printing sorry message about missing graphite support.
1443 Mention -floop-nest-optimize.
1445 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
1447 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
1448 (vtst_p16): Likewise.
1449 (vtstq_p8): Likewise.
1450 (vtstq_p16): Likewise.
1452 (vtstq_p64): Likewise.
1453 * config/arm/arm_neon.h (vgetq_lane_p64): New.
1454 (vset_lane_p64): New.
1455 (vsetq_lane_p64): New.
1457 2017-02-10 Jakub Jelinek <jakub@redhat.com>
1459 PR tree-optimization/79411
1460 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
1461 stmt operands are SSA_NAMEs used in abnormal phis.
1462 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
1465 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
1468 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
1471 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
1473 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
1475 2017-02-09 Jakub Jelinek <jakub@redhat.com>
1477 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
1481 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
1482 not arbitrary TREE_CONSTANT.
1485 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
1486 "omp declare target link" attribute unless is_global_var.
1487 * omp-offload.c (find_link_var_op): Likewise.
1489 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
1490 Chung-Lin Tang <cltang@codesourcery.com>
1492 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
1494 (gimplify_adjust_omp_clauses): Don't delete TILE.
1495 (gimplify_omp_for): Deal with TILE.
1496 * internal-fn.c (expand_GOACC_TILE): New function.
1497 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
1499 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
1500 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
1502 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
1503 avoid DIV for outermost collapse var.
1504 (expand_oacc_for): Insert tile element loop as needed. Adjust.
1505 Remove out of date comments, fix whitespace.
1506 * omp-general.c (omp_extract_for_data): Deal with tiling.
1507 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
1508 adjust OLF_DIM_BASE value.
1509 (struct omp_for_data): Add tiling field.
1510 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
1511 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
1512 for auto loops. Remove default auto determining, moved to
1513 oacc_loop_fixed_partitions.
1514 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
1515 stmts, add e_mask field.
1516 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
1517 (oacc_thread_numbers): Use oacc_dim_call.
1518 (oacc_xform_tile): New.
1519 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
1520 (finish_oacc_loop): Adjust for ifns vector.
1521 (oacc_loop_discover_walk): Append loop abstraction sites to list,
1522 add case for GOACC_TILE fns.
1523 (oacc_loop_xform_loop): Delete.
1524 (oacc_loop_process): Iterate over call list directly, and add
1525 handling for GOACC_TILE fns.
1526 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
1528 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
1529 vector partitioning to outer loops. Assign 2 partitions to loops
1530 when available. Add TILE handling.
1531 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
1532 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
1533 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
1534 * tree.c (omp_clause_num_ops): Adjust TILE ops.
1535 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
1537 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
1539 * configure.ac (ACX_BUGURL): Update.
1540 * configure: Regenerate.
1542 2017-02-09 Richard Biener <rguenther@suse.de>
1544 PR tree-optimization/69823
1545 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
1546 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
1548 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
1550 * config/arc/arc-c.def: Add __NPS400__ definition.
1551 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
1552 (TARGET_NPS400): Define.
1554 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
1556 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
1558 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
1560 (arc_cpu_types): Fill the arch_info field with a pointer into the
1561 arc_arch_types table.
1562 (arc_selected_cpu): Declare.
1563 * config/arc/arc.c (arc_selected_cpu): Make global.
1564 (arc_selected_arch): Delete.
1565 (arc_base_cpu): Delete.
1566 (arc_override_options): Remove references to deleted variables,
1567 update access to arch information.
1568 (ARC_OPT): Update access to arch information.
1569 (ARC_OPTX): Likewise.
1570 * config/arc/arc.h (arc_base_cpu): Remove declaration.
1571 (TARGET_ARC600): Update access to arch information.
1572 (TARGET_ARC601): Likewise.
1573 (TARGET_ARC700): Likewise.
1574 (TARGET_EM): Likewise.
1575 (TARGET_HS): Likewise.
1576 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
1579 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
1582 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
1583 condition/operands for integer GE/LE/GEU/LEU operations.
1585 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
1587 PR translation/79397
1588 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
1591 2017-02-08 Martin Jambor <mjambor@suse.cz>
1594 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
1595 whether allocation happened.
1596 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
1597 nothing was allocated.
1599 2017-02-08 Jakub Jelinek <jakub@redhat.com>
1601 PR tree-optimization/79408
1602 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
1603 constant, but SSA_NAME with a known integer range, use the minimum
1604 of that range instead of op1 to determine if modulo can be replaced
1605 with its first operand.
1607 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1609 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
1611 2017-02-08 Richard Biener <rguenther@suse.de>
1613 PR tree-optimization/71824
1614 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
1615 Check all loops contained in the merged region.
1617 2017-02-07 Andrew Pinski <apinski@cavium.com>
1619 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
1621 2017-02-07 Andrew Pinski <apinski@cavium.com>
1623 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
1624 (thunderxt88): Likewise.
1625 (thunderxt81): Disable LSE and change v8.1 to v8.
1626 (thunderxt83): Likewise.
1628 2017-02-07 Jakub Jelinek <jakub@redhat.com>
1629 Richard Biener <rguenther@suse.de>
1632 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
1633 type from int to size_t.
1634 * ira-costs.c (struct_costs_size): Change type from int to size_t.
1636 2017-02-07 Jakub Jelinek <jakub@redhat.com>
1638 PR rtl-optimization/79386
1639 * cprop.c (bypass_conditional_jumps): Initialize
1640 bypass_last_basic_block already before splitting bbs after
1641 unconditional traps...
1642 (bypass_conditional_jumps): ... rather than here.
1645 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
1646 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
1647 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
1648 fix -masm=intel patterns.
1650 2017-02-07 Richard Biener <rguenther@suse.de>
1652 PR tree-optimization/79256
1654 * builtins.c (get_object_alignment_2): Use min_align_of_type
1655 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
1656 and ADJUST_FIELD_ALIGN.
1658 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
1660 * doc/tm.texi: Regenerate.
1661 * stor-layout.c (layout_decl): Adjust.
1662 (update_alignment_for_field): Likewise.
1663 (place_field): Likewise.
1664 (min_align_of_type): Likewise.
1665 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
1666 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
1667 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
1668 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
1669 * config/frv/frv.c (frv_adjust_field_align): Likewise.
1670 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
1671 * config/i386/i386.c (x86_field_alignment): Likewise.
1672 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
1673 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
1674 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
1675 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
1676 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
1677 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
1681 2017-01-30 Richard Biener <rguenther@suse.de>
1683 PR tree-optimization/79256
1684 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
1685 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
1688 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
1690 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
1691 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
1692 builtins to SImode and emit a zero-extend, if necessary.
1694 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
1696 * docs/invoke.texi (RISC-V Options): Alphabetize.
1698 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
1700 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
1703 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
1705 * config/riscv/riscv.c: New file.
1706 * gcc/common/config/riscv/riscv-common.c: Likewise.
1707 * config.gcc: Likewise.
1708 * config/riscv/constraints.md: Likewise.
1709 * config/riscv/elf.h: Likewise.
1710 * config/riscv/generic.md: Likewise.
1711 * config/riscv/linux.h: Likewise.
1712 * config/riscv/multilib-generator: Likewise.
1713 * config/riscv/peephole.md: Likewise.
1714 * config/riscv/pic.md: Likewise.
1715 * config/riscv/predicates.md: Likewise.
1716 * config/riscv/riscv-builtins.c: Likewise.
1717 * config/riscv/riscv-c.c: Likewise.
1718 * config/riscv/riscv-ftypes.def: Likewise.
1719 * config/riscv/riscv-modes.def: Likewise.
1720 * config/riscv/riscv-opts.h: Likewise.
1721 * config/riscv/riscv-protos.h: Likewise.
1722 * config/riscv/riscv.h: Likewise.
1723 * config/riscv/riscv.md: Likewise.
1724 * config/riscv/riscv.opt: Likewise.
1725 * config/riscv/sync.md: Likewise.
1726 * config/riscv/t-elf-multilib: Likewise.
1727 * config/riscv/t-linux: Likewise.
1728 * config/riscv/t-linux-multilib: Likewise.
1729 * config/riscv/t-riscv: Likewise.
1730 * configure.ac: Likewise.
1731 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
1732 Waterman as RISC-V maintainers.
1733 * doc/install.texi: Add RISC-V entries.
1734 * doc/invoke.texi: Add RISC-V options section.
1735 * doc/md.texi: Add RISC-V constraints section.
1736 * configure: Regenerated.
1738 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
1741 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
1742 false values to be constant vectors with all 0 or all 1 bits set.
1743 (vcondu<mode><mode>): Likewise.
1744 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
1746 (fpmask_comparison_operator): Update comment.
1747 (vecint_comparison_operator): New predicate.
1748 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
1749 vector conditionals when the true and false values are constant
1750 vectors with all 0 bits or all 1 bits set.
1752 2017-02-06 Martin Sebor <msebor@redhat.com>
1754 PR tree-optimization/79376
1755 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
1757 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
1759 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
1760 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
1761 to simplify split condition.
1763 2017-02-06 Jakub Jelinek <jakub@redhat.com>
1765 * omp-expand.c (oxpand_omp_atomic_fetch_op,
1766 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
1769 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
1771 PR rtl-optimization/68664
1772 * target.def (can_speculate_insn): New hook.
1773 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
1774 * doc/tm.texi: Regenerate.
1775 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
1776 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
1777 (rs6000_sched_can_speculate_insn): New function.
1779 2017-02-06 Jakub Jelinek <jakub@redhat.com>
1781 PR tree-optimization/79284
1782 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
1783 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
1784 vectorizable_mask_load_store, vectorizable_operation,
1785 vect_is_simple_cond, get_same_sized_vectype): Use it instead
1786 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
1787 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
1788 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
1789 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
1790 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
1791 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
1792 is_gimple_assign (stmt). Replace another such test with
1793 is_gimple_assign (stmt).
1795 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
1798 * config/avr/avr.c (rtl-iter.h): Include it.
1799 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
1800 (avr_legitimate_combined_insn): ...and implementation.
1802 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
1804 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
1805 * config/s390/s390.c (s390_const_operand_ok)
1806 (s390_canonicalize_comparison, s390_extract_part)
1807 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
1808 (s390_contiguous_bitmask_p, s390_rtx_costs)
1809 (legitimize_pic_address): Likewise.
1810 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
1811 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
1812 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
1813 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
1814 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
1816 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
1818 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
1819 REGNO($0) == REGNO($1).
1821 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1823 * config/s390/linux.h(SIZE_TYPE): Add comment.
1825 2017-02-06 Julian Brown <julian@codesourcery.com>
1826 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
1827 Virendra Pathak <virendra.pathak@broadcom.com>
1829 * config/aarch64/aarch64-cores.def: Change the scheduler
1831 * config/aarch64/aarch64.md: Include thunderx2t99.md.
1832 * config/aarch64/thunderx2t99.md: New file.
1834 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
1836 * doc/standards.texi (Go Language): Update link to language
1839 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
1841 * tree-eh.c (lower_resx): Sanitize profile.
1842 (cleanup_empty_eh_move_lp): Likewise.
1844 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
1847 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
1849 * cfgloopmanip.h (loop_version): Update prototype.
1850 * modulo-sched.c (sms_schedule): Update call of loop_version.
1851 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
1852 * tree-parloops.c (gen_parallel_loop): Likewise.
1853 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
1854 * tree-ssa-loop-split.c (split_loop): Likewise.
1855 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
1856 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
1858 2017-02-05 Martin Liska <mliska@suse.cz>
1861 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
1863 (print_operand_address): Initialize a struct to zero.
1865 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
1867 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
1868 garbage collector only in textual form.
1870 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
1872 * doc/extend.texi (x86 specific memory model extensions for
1873 transactional memory): Simplify a phrase.
1875 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
1878 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
1879 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
1880 (atomic_storedi_1): Likewise.
1882 2017-02-04 Jakub Jelinek <jakub@redhat.com>
1884 PR tree-optimization/79338
1885 * tree-parloops.c (gather_scalar_reductions): Don't call
1886 vect_analyze_loop_form for loop->inner before destroying loop's
1889 2017-02-03 Martin Sebor <msebor@redhat.com>
1891 PR tree-optimization/79327
1892 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
1893 when precision has resulted in leading zeros.
1894 (format_integer): Adjust the likely counter to assume an unknown
1895 argument that may be zero is non-zero.
1897 2017-02-03 Jason Merrill <jason@redhat.com>
1900 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
1901 avoid copying non-taken branch.
1903 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1905 PR tree-optimization/79340
1906 * tree-vect-loop.c (vectorizable_reduction): Release
1907 vec_defs elements after safe_splicing them into other vectors.
1910 PR tree-optimization/79327
1911 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
1912 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
1914 (format_integer): Use wide_int_to_tree instead of build_int_cst
1915 + to_?hwi. If argmin is NULL, just set argmin and argmax to
1916 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
1917 of shortest and longest sequence.
1919 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
1921 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
1922 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
1924 2017-02-03 Walter Lee <walt@tilera.com>
1927 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
1928 after initial stackframe link reg save.
1929 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
1931 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1934 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
1935 wu for stxssp alternative.
1937 2017-02-03 Martin Sebor <msebor@redhat.com>
1939 PR tree-optimization/79352
1940 * gimple-fold.c (get_range_strlen): Add argument.
1941 (get_range_strlen): Change return type to bool.
1942 (get_maxval_strlen): Pass in a dummy argument.
1943 * gimple-fold.h (get_range_strlen): Change return type to bool.
1944 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
1945 * tree.h (array_at_struct_end_p): Add argument.
1946 * tree.c (array_at_struct_end_p): Handle it.
1948 2017-02-03 Martin Liska <mliska@suse.cz>
1951 * multiple_target.c (create_dispatcher_calls): Redirect edge
1952 from a caller of a dispatcher.
1953 (expand_target_clones): Make the clones local.
1954 (ipa_target_clone): Do both target clones and resolvers.
1955 (ipa_dispatcher_calls): Remove the pass.
1956 (pass_dispatcher_calls::gate): Likewise.
1957 (make_pass_dispatcher_calls): Likewise.
1958 * passes.def (pass_target_clone): Put as very first IPA early
1961 2017-02-03 Martin Liska <mliska@suse.cz>
1963 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
1964 in case of a function with ifunc attribute.
1966 2017-02-03 Martin Liska <mliska@suse.cz>
1968 * cgraph.c (cgraph_node::dump): Dump function version info.
1969 * symtab.c (symtab_node::dump_base): Add missing new line.
1971 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
1973 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
1974 (ifcombine_ifandif): Use it.
1976 2017-02-03 Martin Liska <mliska@suse.cz>
1978 * doc/invoke.texi: Document default value for
1979 use-after-scope-direct-emission-threshold.
1981 2017-02-03 Martin Liska <mliska@suse.cz>
1983 PR tree-optimization/79339
1984 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
1985 (format_floating): Likewise.
1987 2017-02-03 Martin Liska <mliska@suse.cz>
1990 * ipa-prop.c (ipa_node_params_t::insert): Remove current
1992 (ipa_node_params_t::remove): Likewise.
1993 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
1994 initialization from removed ipa_node_params_t::insert.
1995 (ipa_node_params::~ipa_node_params): Move from removed
1996 ipa_node_params_t::release.
1997 * symbol-summary.h (symbol_summary::m_released): New member.
1998 Do not release a summary twice. Do not allow to call finalizer
1999 for types of a summary that live in GGC memory.
2001 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
2003 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
2006 2017-02-02 Martin Sebor <msebor@redhat.com>
2009 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
2010 (format_string): Tighten up the range of output for non-constant
2011 strings and correct the expected range for wide non-constant strings.
2013 2017-02-02 Martin Sebor <msebor@redhat.com>
2015 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
2018 * doc/invoke.texi (-fdump-tree-): Remove pass-specific options from
2020 (-fdump-tree-@var): Add to index and document how to come up
2021 with pass-specific option and dump file names.
2022 (-fdump-passes): Clarify where to look for output.
2024 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
2027 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
2028 statistics of the analyzed path; allow threading for speed when
2029 any of BBs along the path are optimized for speed.
2031 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
2034 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
2035 settings of the virtual registers.
2038 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
2040 * explow.c (get_dynamic_stack_size): Take known alignment of stack
2041 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
2044 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2046 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
2047 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
2049 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2051 * config/s390/s390.md: Add missing comments with the expanded
2053 * config/s390/vector.md: Likewise.
2054 * config/s390/vx-builtins.md: Likewise.
2056 2017-02-02 Jakub Jelinek <jakub@redhat.com>
2059 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
2060 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
2061 conditions on a single line.
2063 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2065 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
2066 __S390_VX__ to __VX__.
2068 2017-02-01 Andrew Pinski <apinski@cavium.com>
2070 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
2071 stmt_info to record_stmt_cost.
2072 (vect_get_known_peeling_cost): Pass stmt_info if known to
2074 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
2075 cpu_vector_cost field into
2076 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
2077 field into vec_int_stmt_cost and vec_fp_stmt_cost.
2078 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
2079 splitting of scalar_stmt_cost and vec_stmt_cost.
2080 (thunderx_vector_cost): Likewise.
2081 (cortexa57_vector_cost): LIkewise.
2082 (exynosm1_vector_cost): Likewise.
2083 (xgene1_vector_cost): Likewise.
2084 (thunderx2t99_vector_cost): Improve after the splitting of the two
2086 (aarch64_builtin_vectorization_cost): Update for the splitting of
2087 scalar_stmt_cost and vec_stmt_cost.
2089 2017-02-01 Torvald Riegel <triegel@redhat.com>
2090 Richard Henderson <rth@redhat.com>
2092 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
2093 conditional on existance of a fast atomic load.
2094 * optabs-query.c (can_atomic_load_p): New function.
2095 * optabs-query.h (can_atomic_load_p): Declare it.
2096 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
2097 no fast atomic load is available for the particular size of access.
2098 (expand_atomic_compare_and_swap): Likewise.
2099 (expand_atomic_load): Likewise.
2100 (expand_atomic_store): Likewise.
2101 (expand_atomic_fetch_op): Likewise.
2102 * testsuite/lib/target-supports.exp
2103 (check_effective_target_sync_int_128): Remove x86 because it provides
2104 no fast atomic load.
2105 (check_effective_target_sync_int_128_runtime): Likewise.
2107 2017-02-01 Richard Biener <rguenther@suse.de>
2109 * graphite.c: Include tree-vectorizer.h for find_loop_location.
2110 (graphite_transform_loops): Provide opt-info for optimized nests.
2111 * tree-parloop.c (parallelize_loops): Provide opt-info for
2114 2017-02-01 Richard Biener <rguenther@suse.de>
2117 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
2120 2017-02-01 Richard Biener <rguenther@suse.de>
2122 PR tree-optimization/71824
2123 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
2124 Verify the loops are valid in the merged SESE region.
2125 (scop_detection::can_represent_loop_1): Check analyzing the
2126 evolution of the number of iterations in the region succeeds.
2128 2017-01-31 Ian Lance Taylor <iant@golang.org>
2130 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
2131 REG_ARGS_SIZE note to 32-bit push insns and call insn.
2133 2017-01-31 David Malcolm <dmalcolm@redhat.com>
2135 PR preprocessor/79210
2136 * input.c (get_substring_ranges_for_loc): Replace line_width
2137 assertion with error-handling.
2139 2017-01-31 Richard Biener <rguenther@suse.de>
2141 PR tree-optimization/77318
2142 * graphite-sese-to-poly.c (extract_affine): Fix assert.
2143 (create_pw_aff_from_tree): Take loop parameter.
2144 (add_condition_to_pbb): Pass loop of the condition to
2145 create_pw_aff_from_tree.
2147 2017-01-31 Jakub Jelinek <jakub@redhat.com>
2149 * config/s390/s390.c (s390_asan_shadow_offset): New function.
2150 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
2152 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2156 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
2158 (convert_int_to_float128): Likewise.
2159 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
2160 (convert_int_to_float128): Likewise.
2161 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
2162 (UNSPEC_IEEE128_CONVERT): Likewise.
2163 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
2164 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
2165 Use local variables for IBM extended format.
2166 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
2167 (fix_trunc<mode>si2_fprs): Likewise.
2168 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
2169 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
2170 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
2171 to know that we can now have integers of all sizes in vector
2173 (fix<uns>_<mode>di2_hw): Likewise.
2174 (float<uns>_<mode>si2_hw): Likewise.
2175 (fix_<mode>si2_hw): Likewise.
2176 (fixuns_<mode>si2_hw): Likewise.
2177 (float<uns>_<mode>di2_hw): Likewise.
2178 (float_<mode>di2_hw): Likewise.
2179 (float_<mode>si2_hw): Likewise.
2180 (floatuns_<mode>di2_hw): Likewise.
2181 (floatuns_<mode>si2_hw): Likewise.
2182 (xscvqp<su>wz_<mode>): Delete, no longer used.
2183 (xscvqp<su>dz_<mode>): Likewise.
2184 (xscv<su>dqp_<mode>): Likewise.
2185 (ieee128_mfvsrd_64bit): Likewise.
2186 (ieee128_mfvsrd_32bit): Likewise.
2187 (ieee128_mfvsrwz): Likewise.
2188 (ieee128_mtvsrw): Likewise.
2189 (ieee128_mtvsrd_64bit): Likewise.
2190 (ieee128_mtvsrd_32bit): Likewise.
2192 2017-01-31 Martin Liska <mliska@suse.cz>
2195 * ipa-prop.c (ipa_free_all_node_params): Call release method
2196 instead of ~sumbol_summary to not to trigger double times
2199 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
2201 PR tree-optimization/71691
2202 * bitmap.h (class auto_bitmap): New.
2203 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
2204 is_maybe_undefined instead of ssa_undefined_value_p.
2206 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2208 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
2209 __S390_ARCH_LEVEL__ to __ARCH__.
2211 2017-01-31 Jakub Jelinek <jakub@redhat.com>
2213 PR tree-optimization/79267
2214 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
2215 if should_remove_lhs_p is true.
2217 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
2220 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
2221 (add_alignment_attribute): New.
2222 (base_type_die): Add alignment attribute.
2223 (subrange_type_die): Likewise.
2224 (modified_type_die): Likewise.
2225 (gen_array_type_die): Likewise.
2226 (gen_descr_array_type_die: Likewise.
2227 (gen_enumeration_type_die): Likewise.
2228 (gen_subprogram_die): Likewise.
2229 (gen_variable_die): Likewise.
2230 (gen_field_die): Likewise.
2231 (gen_ptr_to_mbr_type_die): Likewise.
2232 (gen_struct_or_union_type_die): Likewise.
2233 (gen_subroutine_type_die): Likewise.
2234 (gen_typedef_die): Likewise.
2235 (base_type_cmp): Compare alignment attribute.
2237 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
2240 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
2241 (setb_unsigned) New pattern for setb with CCUNS.
2242 * config/rs6000/rs6000.c (expand_block_compare): Use a different
2243 subfc./subfe sequence to avoid overflow problems. Generate a
2244 shorter sequence with cmpld/setb for power9.
2245 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
2246 for generating subfc. instruction.
2247 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
2248 now uses this instruction.
2250 2017-01-30 Ian Lance Taylor <iant@google.com>
2253 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
2254 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
2256 2017-01-30 Martin Sebor <msebor@redhat.com>
2258 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
2259 Move constant to the right of a relational operator.
2260 (get_mpfr_format_length, format_character, format_string): Ditto.
2261 (should_warn_p, maybe_warn): Same.
2263 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
2265 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
2268 * asan.c (get_translation_unit_decl): Remove function.
2269 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
2271 2017-01-30 Martin Liska <mliska@suse.cz>
2273 PR gcov-profile/79259
2274 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
2277 2017-01-30 Martin Liska <mliska@suse.cz>
2280 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
2281 Initialize variables with NULL value.
2283 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
2286 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
2288 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
2290 2017-01-30 Richard Biener <rguenther@suse.de>
2292 PR tree-optimization/79276
2293 * tree-vrp.c (process_assert_insertions): Properly adjust common
2294 when removing a duplicate.
2296 * gcc.dg/torture/pr79276.c: New testcase.
2298 2017-01-30 Richard Biener <rguenther@suse.de>
2300 PR tree-optimization/79256
2301 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
2302 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
2304 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
2306 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
2309 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
2310 ("*r<noxa>sbg_<mode>_sll_bitmask")
2311 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
2312 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
2313 Use contiguous_bitmask_nowrap_operand.
2315 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2318 * config/rs6000/altivec.h (vec_xl): Revise #define.
2319 (vec_xst): Likewise.
2321 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
2323 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
2325 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
2327 PR rtl-optimization/79194
2328 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
2329 traps before call to bypass_conditional_jumps.
2331 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
2333 PR tree-optimization/71374
2334 * lra-constraints.c (check_conflict_input_operands): New.
2335 (match_reload): Use it.
2337 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
2340 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
2341 account to calculate conflict_set.
2343 2017-01-27 Bin Cheng <bin.cheng@arm.com>
2345 PR rtl-optimization/78559
2346 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
2347 other_insn in combine.
2349 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
2351 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
2352 uint16_type_node for BT_UINT16.
2354 2017-01-27 David Malcolm <dmalcolm@redhat.com>
2356 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
2357 "RTL Tests" to menu.
2358 (GIMPLE Tests): New node.
2359 (RTL Tests): New node.
2361 2017-01-27 Richard Biener <rguenther@suse.de>
2363 PR tree-optimization/79245
2364 * tree-loop-distribution.c (distribute_loop): Apply cost
2365 modeling also to detected patterns.
2367 2017-01-27 Richard Biener <rguenther@suse.de>
2369 PR tree-optimization/71433
2370 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
2371 (compare_assert_loc): New function.
2372 (process_assert_insertions): Sort and optimize assert locations
2373 to remove duplicates and push down identical assertions on
2374 edges to their destination block.
2376 2017-01-27 Richard Biener <rguenther@suse.de>
2378 PR tree-optimization/79244
2379 * tree-vrp.c (remove_range_assertions): Forcefully propagate
2380 out SSA names even if abnormal.
2382 2017-01-27 Jakub Jelinek <jakub@redhat.com>
2384 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
2385 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
2386 instead of MPFR_RNDN.
2388 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
2391 * arm.c (arm_option_override): Don't call build_target_option_node
2392 until after doing all option overrides.
2393 (arm_valid_target_attribute_tree): Likewise.
2395 2017-01-27 Martin Liska <mliska@suse.cz>
2397 * doc/invoke.texi (-fprofile-arcs): Document profiling support
2398 for {cd}tors and C++ {cd}tors.
2400 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
2402 * config/s390/s390.md ("*setmem_long_and")
2403 ("*setmem_long_and_31z"): Use zero_extend instead of and.
2405 2017-01-26 Martin Sebor <msebor@redhat.com>
2407 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
2410 2017-01-26 Martin Sebor <msebor@redhat.com>
2412 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
2413 HAVE_DFmode before using XFmode or DFmode.
2414 (parse_directive): Avoid using the z length modifier to avoid
2415 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
2418 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
2419 to accept adjustment as an array.
2420 (get_int_range): New function.
2421 (struct directive): Make width and prec arrays.
2422 (directive::set_width, directive::set_precision): Call get_int_range.
2423 (format_integer, format_floating): Handle width and precision ranges.
2424 (format_string, parse_directive): Same.
2426 2017-01-26 Jakub Jelinek <jakub@redhat.com>
2429 * dwarf2out.c (generate_skeleton_bottom_up): For children with
2430 comdat_type_p set, just clone them, but keep the children in the
2434 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
2435 which have direct callers with -fvar-tracking-assignments enabled
2437 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
2438 inside of type units.
2440 2017-01-26 Martin Sebor <msebor@redhat.com>
2443 * gimple-ssa-sprintf.c (struct result_range): Add likely and
2445 (struct format_result): Replace number_chars, number_chars_min,
2446 and number_chars_max with a single member of struct result_range.
2448 (format_result::operator+=): Adjust.
2449 (struct fmtresult): Remove bounded. Handle likely and unlikely
2451 (fmtresult::adjust_for_width_or_precision): New function.
2452 (fmtresult:type_max_digits): New function.
2453 (bytes_remaining): Handle likely and unlikely counters.
2454 (min_bytes_remaining): Remove.
2455 (format_percent): Simplify.
2456 (format_integer, format_floating): Set likely and unlikely counters.
2457 (get_string_length, format_character, format_string): Same.
2458 (format_plain, should_warn_p): New function.
2459 (maybe_warn): Call should_warn_p. Update diagnostic messages
2460 and handle those for all directives, including plain strings.
2461 (format_directive): Handle likely and unlikely counters.
2462 Remove unnecessary quoting from diagnostics. Add an informational
2464 (add_bytes): Remove.
2465 (pass_sprintf_length::compute_format_length): Simplify.
2466 (try_substitute_return_value): Handle likely and unlikely counters.
2468 2017-01-26 Carl Love <cel@us.ibm.com>
2470 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
2471 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
2473 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
2476 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
2477 endianess for subregs into account.
2478 * lra-constraints.c (lra_constraints): Do risky transformations
2479 always on the first iteration.
2480 * lra-lives.c (check_pseudos_live_through_calls): Add arg
2481 last_call_used_reg_set.
2482 (process_bb_lives): Define and use last_call_used_reg_set.
2483 * lra.c (lra): Always continue after lra_constraints on the first
2486 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
2488 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
2490 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
2492 2017-01-26 Jakub Jelinek <jakub@redhat.com>
2494 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
2495 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
2496 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
2497 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
2498 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
2499 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
2500 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
2501 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
2502 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
2504 2017-01-26 Marek Polacek <polacek@redhat.com>
2507 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
2508 for the third operand.
2510 2017-01-26 Jakub Jelinek <jakub@redhat.com>
2513 * omp-low.c (struct omp_context): Add simt_stmt field.
2514 (scan_omp_for): Return omp_context *.
2515 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
2516 context to the _simt_ SIMD stmt.
2517 (lower_omp_for): For combined SIMD with sibling _simt_
2518 SIMD, make sure to use the same decls in _looptemp_
2519 clauses as in the sibling.
2521 2017-01-26 David Sherwood <david.sherwood@arm.com>
2524 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
2527 2017-01-26 Jakub Jelinek <jakub@redhat.com>
2530 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
2531 emit fld b; fld a; if possible.
2533 * brig-builtins.def: Update copyright years.
2534 * config/arm/arm_acle_builtins.def: Update copyright years.
2536 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
2539 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
2540 constraint instead of o for the stxsd instruction.
2542 2017-01-25 Carl Love <cel@us.ibm.com>
2544 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
2545 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
2547 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
2549 * doc/invoke.texi (C++ Dialect Options): Fix typo.
2551 2017-01-25 Richard Biener <rguenther@suse.de>
2553 PR tree-optimization/69264
2554 * target.def (vector_alignment_reachable): Improve documentation.
2555 * doc/tm.texi: Regenerate.
2556 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
2558 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
2559 earlier changes with respect to TYPE_USER_ALIGN.
2560 (vector_alignment_reachable_p): Likewise. Improve dumping.
2562 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2565 * config/arm/arm.md (xordi3): Force constant operand into a register
2568 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2570 * doc/invoke.texi (-fstore-merging): Correct default optimization
2571 levels at which it is enabled.
2572 (-O): Move -fstore-merging from list to...
2575 2017-01-25 Richard Biener <rguenther@suse.de>
2578 * omp-expand.c: Include debug.h.
2579 (expand_omp_taskreg): Make sure to generate early debug before
2580 outlining anything from a function.
2581 (expand_omp_target): Likewise.
2582 (grid_expand_target_grid_body): Likewise.
2584 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
2587 * asan.c (get_translation_unit_decl): New function.
2588 (asan_add_global): Extract modules file name from globals
2589 TRANSLATION_UNIT_DECL name.
2591 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
2594 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
2595 for long calls with APCS frame and VFP.
2597 2017-01-24 David Malcolm <dmalcolm@redhat.com>
2599 * cfg.c (original_copy_tables_initialized_p): New function.
2600 * cfg.h (original_copy_tables_initialized_p): New decl.
2601 * cfgrtl.c (relink_block_chain): Guard the call to
2602 free_original_copy_tables with a call to
2603 original_copy_tables_initialized_p.
2604 * cgraph.h (symtab_node::native_rtl_p): New decl.
2605 * cgraphunit.c (symtab_node::native_rtl_p): New function.
2606 (symtab_node::needed_p): Don't assert for early assembly output
2607 for __RTL functions.
2608 (cgraph_node::finalize_function): Set "force_output" for __RTL
2610 (cgraph_node::analyze): Bail out early for __RTL functions.
2611 (analyze_functions): Update assertion to support __RTL functions.
2612 (cgraph_node::expand): Bail out early for __RTL functions.
2613 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
2615 * function.h (struct function): Update comment for field
2617 * gimple-expr.c: Include "tree-pass.h".
2618 (gimple_has_body_p): Return false for __RTL functions.
2619 * Makefile.in (OBJS): Add run-rtl-passes.o.
2620 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
2622 (gcc::pass_manager::get_clean_slate): New accessor.
2623 * passes.c: Include "insn-addr.h".
2624 (should_skip_pass_p): Add logging. Update logic for running
2625 "expand" to be compatible with both __GIMPLE and __RTL. Guard
2626 property-provider override so it is only done for gimple passes.
2628 (skip_pass): New function.
2629 (execute_one_pass): Call skip_pass when skipping passes.
2630 * read-md.c (md_reader::read_char): Support filtering
2631 the input to a subset of line numbers.
2632 (md_reader::md_reader): Initialize fields
2633 m_first_line and m_last_line.
2634 (md_reader::read_file_fragment): New function.
2635 * read-md.h (md_reader::read_file_fragment): New decl.
2636 (md_reader::m_first_line): New field.
2637 (md_reader::m_last_line): New field.
2638 * read-rtl-function.c (function_reader::create_function): Only
2639 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
2640 curr_properties. Set DECL_INITIAL to a dummy block.
2641 (read_rtl_function_body_from_file_range): New function.
2642 * read-rtl-function.h (read_rtl_function_body_from_file_range):
2644 * run-rtl-passes.c: New file.
2645 * run-rtl-passes.h: New file.
2647 2017-01-24 Jeff Law <law@redhat.com>
2649 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
2652 2017-01-24 Bin Cheng <bin.cheng@arm.com>
2654 PR tree-optimization/79159
2655 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
2656 (record_nonwrapping_iv): Improve boundary using above function if no
2657 value range information.
2659 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
2660 Martin Jambor <mjambor@suse.cz>
2662 * brig-builtins.def: New file.
2663 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
2664 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
2665 (DEF_HSAIL_SAT_BUILTIN): Likewise.
2666 (DEF_HSAIL_INTR_BUILTIN): Likewise.
2667 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
2668 * builtin-types.def (BT_INT8): New.
2669 (BT_INT16): Likewise.
2670 (BT_UINT8): Likewise.
2671 (BT_UINT16): Likewise.
2672 (BT_FN_ULONG): Likewise.
2673 (BT_FN_UINT_INT): Likewise.
2674 (BT_FN_UINT_ULONG): Likewise.
2675 (BT_FN_UINT_LONG): Likewise.
2676 (BT_FN_UINT_PTR): Likewise.
2677 (BT_FN_ULONG_PTR): Likewise.
2678 (BT_FN_INT8_FLOAT): Likewise.
2679 (BT_FN_INT16_FLOAT): Likewise.
2680 (BT_FN_UINT32_FLOAT): Likewise.
2681 (BT_FN_UINT16_FLOAT): Likewise.
2682 (BT_FN_UINT8_FLOAT): Likewise.
2683 (BT_FN_UINT64_FLOAT): Likewise.
2684 (BT_FN_UINT16_UINT32): Likewise.
2685 (BT_FN_UINT32_UINT16): Likewise.
2686 (BT_FN_UINT16_UINT16_UINT16): Likewise.
2687 (BT_FN_INT_PTR_INT): Likewise.
2688 (BT_FN_UINT_PTR_UINT): Likewise.
2689 (BT_FN_LONG_PTR_LONG): Likewise.
2690 (BT_FN_ULONG_PTR_ULONG): Likewise.
2691 (BT_FN_VOID_UINT64_UINT64): Likewise.
2692 (BT_FN_UINT8_UINT8_UINT8): Likewise.
2693 (BT_FN_INT8_INT8_INT8): Likewise.
2694 (BT_FN_INT16_INT16_INT16): Likewise.
2695 (BT_FN_INT_INT_INT): Likewise.
2696 (BT_FN_UINT_FLOAT_UINT): Likewise.
2697 (BT_FN_FLOAT_UINT_UINT): Likewise.
2698 (BT_FN_ULONG_UINT_UINT): Likewise.
2699 (BT_FN_ULONG_UINT_PTR): Likewise.
2700 (BT_FN_ULONG_ULONG_ULONG): Likewise.
2701 (BT_FN_UINT_UINT_UINT): Likewise.
2702 (BT_FN_VOID_UINT_PTR): Likewise.
2703 (BT_FN_UINT_UINT_PTR: Likewise.
2704 (BT_FN_UINT32_UINT64_PTR): Likewise.
2705 (BT_FN_INT_INT_UINT_UINT): Likewise.
2706 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
2707 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
2708 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
2709 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
2710 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
2711 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
2712 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
2713 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
2714 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
2715 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
2716 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
2717 * doc/frontends.texi: List BRIG FE.
2718 * doc/install.texi (Testing): Add BRIG tesring requirements.
2719 * doc/invoke.texi (Overall Options): Mention BRIG.
2720 * doc/standards.texi (Standards): Doucment BRIG HSA version.
2722 2017-01-24 Richard Biener <rguenther@suse.de>
2724 PR translation/79208
2725 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
2727 2017-01-24 Martin Jambor <mjambor@suse.cz>
2730 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
2731 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
2734 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
2737 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
2738 casts from signed to unsigned really don't have a range.
2740 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2742 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
2743 GMP_RNDx for compatiblity.
2745 2017-01-24 Martin Liska <mliska@suse.cz>
2748 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
2749 that would prevent us to call alloca with -1 as argument.
2751 2017-01-24 Jakub Jelinek <jakub@redhat.com>
2753 * dwarf2out.c (output_compilation_unit_header, output_file_names):
2754 Avoid -Wformat-security warning.
2756 2017-01-23 Andrew Pinski <apinski@cavium.com>
2758 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
2761 2017-01-23 Martin Sebor <msebor@redhat.com>
2764 * gimple-ssa-sprintf.c (warn_level): New global.
2765 (format_integer): Use it here and throughout the rest of the file.
2766 Use the same switch to compute sign as base.
2767 (maybe_warn): New function.
2768 (format_directive): Factor out warnings into maybe_warn.
2769 Add debugging output. Use warn_level.
2770 (add_bytes): Use warn_level.
2771 (pass_sprintf_length::compute_format_length): Add debugging output.
2772 (try_substitute_return_value): Same.
2773 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
2776 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
2777 (struct fmtresult, format_integer, format_floating): Adjust.
2778 (fmtresult::fmtresult): Set max correctly in two argument ctor.
2779 (get_string_length, format_string,format_directive): Same.
2780 (pass_sprintf_length::compute_format_length): Same.
2781 (try_substitute_return_value): Simplify slightly.
2784 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
2785 (fmtresult::operator+=): Outlined.
2786 (struct fmtresult): Add ctors.
2787 (struct conversion_spec): Rename...
2788 (struct directive): ...to this. Add and remove data members.
2789 (directive::set_width, directive::set_precision): New functions.
2790 (format_percent): Use fmtresult ctor.
2791 (get_width_and_precision): Remove.
2792 (format_integer): Make naming changes. Avoid computing width and
2794 (format_floating): Same. Adjust indentation.
2795 (format_character, format_none): New functions.
2796 (format_string): Moved character handling to format_character.
2797 (format_directive): Remove arguments, change return type.
2798 (parse_directive): New function.
2799 (pass_sprintf_length::compute_format_length): Move directive
2800 parsing to parse_directive.
2802 2017-01-23 Jakub Jelinek <jakub@redhat.com>
2804 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
2805 (assign_assembler_name_if_needed): ... this.
2806 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
2807 (assign_assembler_name_if_needed): ... this.
2808 (free_lang_data_in_cgraph): Adjust callers.
2809 * cgraphunit.c (cgraph_node::analyze): Likewise.
2810 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
2813 2017-01-23 Richard Biener <rguenther@suse.de>
2815 PR tree-optimization/79088
2816 PR tree-optimization/79188
2817 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
2818 resetting loop bounds after last path deletion. Reset loop
2819 bounds of the target loop, make code match the comments.
2820 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
2821 Make sure loops need no fixups.
2823 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
2825 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
2826 exponent support with double type for first argument.
2827 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
2828 type returned by __builtin_vec_extract_sig,
2829 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
2830 functions from "vector int" to "vector unsigned int" or from
2831 "vector long long int" to "vector unsigned long long int".
2832 Changed type returned by __builtin_vec_extract_exp,
2833 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
2834 functions from "vector int" to "vector unsigned int" or from
2835 "vector long long int" to "vector unsigned long long int".
2836 Changed return type of __builtin_vec_test_data_class,
2837 __builtin_vec_test_data_class_sp, and
2838 __builtin_vec_test_data_class_dp from "vector int" to
2839 "vector bool int" or from "vector long long int" to "vector bool
2840 long long int" and changed second argument type from "unsigned
2841 int" to "int". Added new overloaded function forms "vector float
2842 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
2843 "vector float __builtin_vec_insert_exp_sp (vector float, vector
2844 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
2845 double, vector unsigned long long int)" and "vector double
2846 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
2847 long int)". Changed return type of
2848 __builtin_scalar_test_data_class and
2849 __builtin_scalar_test_data_class_sp and
2850 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
2851 int" and changed second argument from "unsigned int" to "int".
2852 Changed type returned by __builtin_scalar_test_neg,
2853 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
2854 from "int" to "bool int". Added new overloaded function form
2855 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
2856 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
2857 exponent double-precision with floating point first argument.
2858 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
2859 documentation of scalar_test_data_class, scalar_test_neg,
2860 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
2861 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
2862 vec_test_data_class built-in functions to reflect refinements in
2863 their type signatures.
2865 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
2867 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
2869 (aarch64_elf_asm_destructor): Likewise.
2871 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
2873 PR rtl-optimization/78634
2874 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
2875 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
2876 * ifcvt.c (noce_try_cmove): Add missing cost check.
2878 PR rtl-optimization/71724
2879 * combine.c (if_then_else_cond): Look for situations where it is
2880 beneficial to undo the work of one of the recursive calls.
2882 2017-01-23 Bin Cheng <bin.cheng@arm.com>
2884 PR tree-optimization/70754
2885 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
2886 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
2887 combined stmt before it if not NULL.
2888 (combine_chains): Process refs reversely and compute dominance point
2891 2017-01-23 Martin Liska <mliska@suse.cz>
2893 PR tree-optimization/79196
2894 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
2895 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
2897 (strlen_optimize_stmt): Call the renamed function.
2899 2017-01-23 Michael Matz <matz@suse.de>
2901 PR tree-optimization/78384
2902 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
2904 2017-01-23 Richard Biener <rguenther@suse.de>
2906 PR tree-optimization/79186
2907 * tree-vrp.c (register_new_assert_for): Make sure we've seen
2908 both incoming edges before moving an assert.
2910 2017-01-23 Martin Jambor <mjambor@suse.cz>
2912 * ipa-prop.c (load_from_param_1): Removed.
2913 (load_from_unmodified_param): Bits from load_from_param_1 put back
2915 (load_from_param): Removed.
2916 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
2917 with stmt. Reverted back to use of load_from_unmodified_param.
2919 2017-01-23 Martin Jambor <mjambor@suse.cz>
2922 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
2923 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
2924 field a pointer to garbage collected vector, mark lattices and
2925 ipcp_orig_node with GTY((skip)).
2926 (ipa_get_param_count): Adjust to descriptors being a pointer.
2927 (ipa_get_param): Likewise.
2928 (ipa_get_type): Likewise.
2929 (ipa_get_param_move_cost): Likewise.
2930 (ipa_set_param_used): Likewise.
2931 (ipa_get_controlled_uses): Likewise.
2932 (ipa_set_controlled_uses): Likewise.
2933 (ipa_is_param_used): Likewise.
2934 (ipa_node_params_t): Move into garbage collector. New methods insert
2936 (ipa_node_params_sum): Annotate wth GTY(()).
2937 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
2939 (ipa_load_from_parm_agg): Adjust declaration.
2940 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
2941 * ipa-profile.c (ipa_profile): Likewise.
2942 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
2943 (ipa_populate_param_decls): Make descriptors parameter garbage
2945 (ipa_dump_param): Adjust to descriptors being a pointer.
2946 (ipa_alloc_node_params): Likewise.
2947 (ipa_initialize_node_params): Likewise.
2948 (load_from_param_1): Make descriptors parameter garbage collected.
2949 (load_from_unmodified_param): Likewise.
2950 (load_from_param): Likewise.
2951 (ipa_load_from_parm_agg): Likewise.
2952 (ipa_node_params::~ipa_node_params): Removed.
2953 (ipa_free_all_node_params): Remove call to delete operator.
2954 (ipa_node_params_t::insert): New.
2955 (ipa_node_params_t::remove): Likewise.
2956 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
2957 copy known_csts and known_contexts vectors.
2958 (ipa_read_node_info): Adjust to descriptors being a pointer.
2959 (ipcp_modif_dom_walker): Make m_descriptors field garbage
2961 (ipcp_transform_function): Make descriptors variable garbage
2964 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
2966 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
2967 * config/i386/avx512dqintrin.h: Ditto.
2968 * config/i386/avx512fintrin.h: Ditto.
2969 * gcc/config/i386/i386.c: Handle new builtins.
2970 * config/i386/i386-builtin.def: Add new builtins.
2971 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
2972 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
2974 2017-01-23 Jakub Jelinek <jakub@redhat.com>
2975 Martin Liska <mliska@suse.cz>
2977 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
2978 * asan.c (asan_expand_poison_ifn): Support stores and use
2979 appropriate ASAN report function.
2980 * internal-fn.c (expand_ASAN_POISON_USE): New function.
2981 * internal-fn.def (ASAN_POISON_USE): Declare.
2982 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
2983 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
2984 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
2985 ASAN_POISON calls w/o LHS.
2986 * tree-ssa.c (execute_update_addresses_taken): Create clobber
2987 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
2988 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
2989 * gimplify.c (asan_poison_variables): Add attribute
2990 use_after_scope_memory to variables that really needs to live
2992 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
2993 having the attribute.
2995 2017-01-23 Martin Liska <mliska@suse.cz>
2997 * asan.c (create_asan_shadow_var): New function.
2998 (asan_expand_poison_ifn): Likewise.
2999 * asan.h (asan_expand_poison_ifn): New declaration.
3000 * internal-fn.c (expand_ASAN_POISON): Likewise.
3001 * internal-fn.def (ASAN_POISON): New builtin.
3002 * sanopt.c (pass_sanopt::execute): Expand
3003 asan_expand_poison_ifn.
3004 * tree-inline.c (copy_decl_for_dup_finish): Make function
3006 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
3007 * tree-ssa.c (is_asan_mark_p): New function.
3008 (execute_update_addresses_taken): Rewrite local variables
3009 (identified just by use-after-scope as addressable) into SSA.
3011 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
3013 * doc/install.texi (Specific): opensource.apple.com uses https
3014 now. Remove trailing slash.
3016 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
3018 * README.Portability: Remove note on an Irix compatibility issue.
3020 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
3022 * gcov.c (INCLUDE_ALGORITHM): Define.
3023 (INCLUDE_VECTOR): Define.
3024 No longer include <vector> and <algorithm> directly.
3026 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
3028 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
3030 * doc/invoke.texi (Code Gen Options): Ditto.
3032 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
3035 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
3037 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
3039 rtl-optimization/79125
3040 * cprop.c (local_cprop_pass): Handle cases where we make an
3043 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
3047 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
3048 read from, for big endian.
3050 2017-01-20 Jiong Wang <jiong.wang@arm.com>
3052 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
3053 register pauth builtins for LP64 only.
3055 2017-01-20 Marek Polacek <polacek@redhat.com>
3058 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
3061 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
3063 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
3065 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
3066 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
3067 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
3069 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3072 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
3073 in big-endian mode when they are not a single duplicated value.
3075 2017-01-20 Richard Biener <rguenther@suse.de>
3077 * BASE-VER: Bump to 7.0.1.
3079 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
3081 * omp-low.c (omplow_simd_context): New struct. Use it...
3082 (lower_rec_simd_input_clauses): ...here and...
3083 (lower_rec_input_clauses): ...here to hold common data. Adjust all
3084 references to idx, lane, max_vf, is_simt.
3086 2017-01-20 Graham Markall <graham.markall@embecosm.com>
3088 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
3091 2017-01-20 Martin Jambor <mjambor@suse.cz>
3093 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
3094 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
3096 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
3097 (GTFILES): Rename hsa.c to hsa-common.c.
3098 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
3099 * hsa-dump.c: Likewise.
3100 * hsa-gen.c: Likewise.
3101 * hsa-regalloc.c: Likewise.
3102 * ipa-hsa.c: Likewise.
3103 * omp-expand.c: Likewise.
3104 * omp-low.c: Likewise.
3105 * toplev.c: Likewise.
3107 2017-01-20 Marek Polacek <polacek@redhat.com>
3110 * doc/invoke.texi: Document -Wduplicated-branches.
3111 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
3112 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
3113 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
3114 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
3115 return 0 only when not OEP_LEXICOGRAPHIC.
3116 (fold_build_cleanup_point_expr): Use the expression
3117 location when building CLEANUP_POINT_EXPR.
3118 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
3119 * tree.c (add_expr): Handle error_mark_node.
3121 2017-01-20 Martin Liska <mliska@suse.cz>
3124 * tree-profile.c (init_ic_make_global_vars): Do not call
3126 (gimple_init_gcov_profiler): Likewise.
3128 2017-01-20 Martin Liska <mliska@suse.cz>
3131 * cgraph.h (maybe_create_reference): Remove argument and
3133 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
3135 * ipa-cp.c (create_specialized_node): Likewise.
3136 * symtab.c (symtab_node::maybe_create_reference): Handle
3137 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
3139 2017-01-20 Martin Liska <mliska@suse.cz>
3141 * read-rtl-function.c (function_reader::create_function): Use
3142 build_decl instread of build_decl_stat.
3144 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
3146 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
3147 * config/i386/avx512dqintrin.h: Ditto.
3148 * config/i386/avx512fintrin.h: Ditto.
3149 * config/i386/i386-builtin-types.def: Add new types.
3150 * gcc/config/i386/i386.c: Handle new types.
3151 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
3152 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
3153 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
3154 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
3155 (__builtin_ia32_kshiftridi): New.
3156 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
3158 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
3162 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
3163 define to rs6000_init_stack_protect_guard.
3164 (rs6000_init_stack_protect_guard): New function.
3166 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
3167 Yunqiang Su <yunqiang.su@imgtec.com>
3169 * config.gcc (supported_defaults): Add madd4.
3170 (with_madd4): Add validation.
3171 (all_defaults): Add madd4.
3172 * config/mips/mips.opt (mmadd4): New option.
3173 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
3175 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
3177 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
3178 (ISA_HAS_FUSED_MADD4): Likewise.
3179 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
3180 * gcc/doc/install.texi (--with-madd4): Document the new option.
3182 2017-01-19 Jiong Wang <jiong.wang@arm.com>
3184 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
3185 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
3186 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
3187 (aarch64_init_pauth_hint_builtins): New.
3188 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
3189 (aarch64_expand_builtin): Expand new builtins.
3191 2017-01-19 Jiong Wang <jiong.wang@arm.com>
3193 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
3194 * combine-stack-adj.c (no_unhandled_cfa): Handle
3195 REG_CFA_TOGGLE_RA_MANGLE.
3196 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
3197 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
3198 info for return address signing.
3199 (aarch64_expand_epilogue): Likewise.
3201 2017-01-19 Jiong Wang <jiong.wang@arm.com>
3203 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
3204 * config/aarch64/aarch64-protos.h
3205 (aarch64_return_address_signing_enabled): New declaration.
3206 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
3208 (aarch64_expand_prologue): Sign return address before it's pushed onto
3210 (aarch64_expand_epilogue): Authenticate return address fetched from
3212 (aarch64_override_options): Sanity check for ILP32 and ISA level.
3213 (aarch64_attributes): New function attributes for "sign-return-address".
3214 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
3215 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
3216 ("*do_return"): Generate combined instructions according to key index.
3217 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
3218 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
3220 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
3221 * config/aarch64/aarch64.opt (msign-return-address=): New.
3222 * doc/extend.texi (AArch64 Function Attributes): Documents
3223 "sign-return-address=".
3224 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
3226 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
3228 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
3229 overall option summary.
3231 2017-01-19 Jiong Wang <jiong.wang@arm.com>
3233 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
3234 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
3235 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
3236 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
3238 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
3240 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
3241 -mpower9-minmax by default for -mcpu=power9.
3242 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
3243 128-bit floating point.
3245 2017-01-20 Alan Modra <amodra@gmail.com>
3247 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
3248 optimizing for size.
3250 2017-01-20 Alan Modra <amodra@gmail.com>
3253 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
3254 for strcmp and strncmp from corresponding builtin decl.
3256 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
3258 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
3259 instead of i386/rtems-64.h.
3260 * config/i386/rtems-64.h: Remove.
3262 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
3266 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
3268 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
3270 2017-01-19 Tamar Christina <tamar.christina@arm.com>
3272 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
3273 Change int to HOST_WIDE_INT.
3274 * config/aarch64/aarch64-protos.h
3275 (aarch64_simd_gen_const_vector_dup): Likewise.
3276 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
3278 2017-01-19 David Malcolm <dmalcolm@redhat.com>
3280 * langhooks-def.h (lhd_type_for_size): New decl.
3281 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
3282 * langhooks.c (lhd_type_for_size): New function, taken from
3285 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
3287 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
3288 define_bypass for CR latency.
3289 (power9-cracked-alu): Update bypass latency and remove power9-branch.
3290 (power9-alu2): Add define_bypass for CR latency.
3292 (power9-mul): Update insn latency.
3293 (power9-mul-compare): Update insn latency, bypass latency and remove
3296 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3298 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
3300 * config/aarch64/aarch64.md
3301 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
3302 aarch64_nopcrelative_literal_loads.
3303 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
3305 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
3307 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
3309 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
3311 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
3314 * config.gcc (supported_defaults): Add lxc1-sxc1.
3315 (with_lxc1_sxc1): Add validation.
3316 (all_defaults): Add lxc1-sxc1.
3317 * config/mips/mips.opt (mlxc1-sxc1): New option.
3318 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
3320 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
3321 __mips_no_lxc1_sxc1.
3322 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
3323 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
3324 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
3326 2017-01-19 Richard Biener <rguenther@suse.de>
3328 PR tree-optimization/72488
3329 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
3330 sure to restore SSA info.
3331 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
3333 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
3335 PR rtl-optimization/79121
3336 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
3337 of the inner type when shifting an extended value.
3339 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
3342 * symtab.c (symtab_node::equal_address_to): Fix comparing of
3343 interposable aliases.
3345 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
3348 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
3349 Use the evmergelohi instruction.
3350 (mov_si<mode>_e500_subreg4_2_le): Likewise.
3351 (mov_sitf_e500_subreg8_2_be): Likewise.
3352 (mov_sitf_e500_subreg12_2_le): Likewise.
3353 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
3354 (mov_si<mode>_e500_subreg4_2_be): Likewise.
3355 (mov_sitf_e500_subreg8_2_le): Likewise.
3356 (mov_sitf_e500_subreg12_2_be): Likewise.
3358 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3360 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
3361 attribute from vecsimple to vecperm.
3362 (altivec_vbpermq2): Likewise.
3364 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3367 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
3369 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
3370 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
3371 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
3372 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
3373 case where N arg is SIZE_MAX.
3374 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
3375 (cmpstrsi): Add pattern.
3377 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
3379 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3380 __builtin_vec_revb builtins.
3381 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
3382 built-in functions to support generation of the ISA 3.0 XXBR<x>
3383 vector byte reverse instructions.
3384 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
3385 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
3386 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
3387 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
3388 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
3389 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
3390 (P9V_BUILTIN_VEC_REVB): Likewise.
3391 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
3392 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
3393 (p9_xxbrq_v16qi): Likewise.
3394 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
3395 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
3396 (p9_xxbrh_v8hi): Likewise.
3397 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
3398 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
3399 vec_revb built-in functions.
3401 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
3403 PR rtl-optimization/78952
3404 * config/i386/i386.md (any_extract): New code iterator.
3405 (*insvqi_2): Use any_extract for source operand.
3406 (*insvqi_3): Use any_shiftrt for source operand.
3408 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
3410 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
3412 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
3414 2017-01-18 Matthias Klose <doko@ubuntu.com>
3416 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
3418 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3420 * config/rs6000/altivec.h (vec_bperm): Change #define.
3421 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
3422 (altivec_vbpermq2): New define_insn.
3423 (altivec_vbpermd): Likewise.
3424 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
3426 (VBPERMD): Likewise.
3427 (VBPERM): New polymorphic function interface.
3428 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
3429 Add entries for P9V_BUILTIN_VEC_VBPERM.
3430 * doc/extend.texi: Add interfaces for vec_bperm.
3432 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3434 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
3435 first letter of error messages.
3436 (s390_resolve_overloaded_builtin): Likewise.
3437 * config/s390/s390.c (s390_expand_builtin): Likewise.
3438 (s390_invalid_arg_for_unprototyped_fn): Likewise.
3439 (s390_valid_target_attribute_inner_p): Likewise.
3440 * config/s390/s390.md ("tabort"): Likewise.
3442 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
3444 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
3445 (ISA_AVOID_DIV_HILO): New macro.
3446 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
3447 (ISA_HAS_DDIV): Likewise.
3449 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
3451 * doc/invoke.texi (fabi-version): Correct number of occurrences.
3453 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
3455 * doc/invoke.texi (fabi-version): Spelling fix.
3457 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
3460 * doc/invoke.texi (fabi-version): Mention mangling fix for
3463 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
3466 * doc/invoke.texi (fabi-version): Document discriminator mangling.
3468 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
3471 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
3472 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
3474 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
3476 (stack_protect_test): Ditto.
3477 * config/rs6000/rs6000.opt (mstack-protector-guard=,
3478 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
3480 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
3481 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
3482 -mstack-protector-guard-offset=.
3483 (RS/6000 and PowerPC Options): Ditto.
3485 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
3487 * config/i386/i386.h (MASK_CLASS_P): New define.
3488 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
3489 there are no registers from different register sets also when
3490 mask registers are used. Update function comment.
3491 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
3492 to (*k/*r) and (*k/*km) alternatives.
3494 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
3496 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
3497 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
3498 (EH_RETURN_HANDLER_RTX): New define.
3499 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
3500 Force frame pointer in EH return functions.
3501 (aarch64_expand_epilogue): Add barrier for eh_return.
3502 (aarch64_final_eh_return_addr): Remove.
3503 (aarch64_eh_return_handler_rtx): New function.
3504 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
3506 (aarch64_eh_return_handler_rtx): New prototype.
3508 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3510 * config/rs6000/altivec.h (vec_rlmi): New #define.
3511 (vec_vrlnm): Likewise.
3512 (vec_rlnm): Likewise.
3513 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
3514 (UNSPEC_VRLNM): Likewise.
3515 (VIlong): New mode iterator.
3516 (altivec_vrl<VI_char>mi): New define_insn.
3517 (altivec_vrl<VI_char>nm): Likewise.
3518 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
3521 (RLNM): New polymorphic function entry.
3522 (VRLWMI): New monomorphic function entry.
3524 (RLMI): New polymorphic function entry.
3525 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
3526 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
3527 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
3530 2017-01-17 Jakub Jelinek <jakub@redhat.com>
3533 * dwarf2out.c (field_byte_offset): Restore the
3534 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
3535 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
3536 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
3539 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
3542 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
3544 2017-01-17 Jakub Jelinek <jakub@redhat.com>
3547 * dwarf2out.c (add_data_member_location_attribute): For constant
3548 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
3549 instead of DW_AT_data_member_location, DW_AT_bit_offset and
3550 DW_AT_byte_size attributes.
3552 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
3554 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
3555 after forcing to constant memory when the code model is medium.
3557 2017-01-17 Julia Koval <julia.koval@intel.com>
3560 * config/i386/avx512fintrin.h
3561 (_mm512_i32gather_ps): Change __addr type to void const*.
3562 (_mm512_mask_i32gather_ps): Ditto.
3563 (_mm512_i32gather_pd): Ditto.
3564 (_mm512_mask_i32gather_pd): Ditto.
3565 (_mm512_i64gather_ps): Ditto.
3566 (_mm512_mask_i64gather_ps): Ditto.
3567 (_mm512_i64gather_pd): Ditto.
3568 (_mm512_mask_i64gather_pd): Ditto.
3569 (_mm512_i32gather_epi32): Ditto.
3570 (_mm512_mask_i32gather_epi32): Ditto.
3571 (_mm512_i32gather_epi64): Ditto.
3572 (_mm512_mask_i32gather_epi64): Ditto.
3573 (_mm512_i64gather_epi32): Ditto.
3574 (_mm512_mask_i64gather_epi32): Ditto.
3575 (_mm512_i64gather_epi64): Ditto.
3576 (_mm512_mask_i64gather_epi64): Ditto.
3577 (_mm512_i32scatter_ps): Change __addr type to void*.
3578 (_mm512_mask_i32scatter_ps): Ditto.
3579 (_mm512_i32scatter_pd): Ditto.
3580 (_mm512_mask_i32scatter_pd): Ditto.
3581 (_mm512_i64scatter_ps): Ditto.
3582 (_mm512_mask_i64scatter_ps): Ditto.
3583 (_mm512_i64scatter_pd): Ditto.
3584 (_mm512_mask_i64scatter_pd): Ditto.
3585 (_mm512_i32scatter_epi32): Ditto.
3586 (_mm512_mask_i32scatter_epi32): Ditto.
3587 (_mm512_i32scatter_epi64): Ditto.
3588 (_mm512_mask_i32scatter_epi64): Ditto.
3589 (_mm512_i64scatter_epi32): Ditto.
3590 (_mm512_mask_i64scatter_epi32): Ditto.
3591 (_mm512_i64scatter_epi64): Ditto.
3592 (_mm512_mask_i64scatter_epi64): Ditto.
3593 * config/i386/avx512pfintrin.h
3594 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
3595 (_mm512_mask_prefetch_i32gather_ps): Ditto.
3596 (_mm512_mask_prefetch_i64gather_pd): Ditto.
3597 (_mm512_mask_prefetch_i64gather_ps): Ditto.
3598 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
3599 (_mm512_prefetch_i32scatter_ps): Ditto.
3600 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
3601 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
3602 (_mm512_prefetch_i64scatter_pd): Ditto.
3603 (_mm512_prefetch_i64scatter_ps): Ditto.
3604 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
3605 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
3606 * config/i386/avx512vlintrin.h
3607 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
3608 (_mm_mmask_i32gather_ps): Ditto.
3609 (_mm256_mmask_i32gather_pd): Ditto.
3610 (_mm_mmask_i32gather_pd): Ditto.
3611 (_mm256_mmask_i64gather_ps): Ditto.
3612 (_mm_mmask_i64gather_ps): Ditto.
3613 (_mm256_mmask_i64gather_pd): Ditto.
3614 (_mm_mmask_i64gather_pd): Ditto.
3615 (_mm256_mmask_i32gather_epi32): Ditto.
3616 (_mm_mmask_i32gather_epi32): Ditto.
3617 (_mm256_mmask_i32gather_epi64): Ditto.
3618 (_mm_mmask_i32gather_epi64): Ditto.
3619 (_mm256_mmask_i64gather_epi32): Ditto.
3620 (_mm_mmask_i64gather_epi32): Ditto.
3621 (_mm256_mmask_i64gather_epi64): Ditto.
3622 (_mm_mmask_i64gather_epi64): Ditto.
3623 (_mm256_i32scatter_ps): Change __addr type to void*.
3624 (_mm256_mask_i32scatter_ps): Ditto.
3625 (_mm_i32scatter_ps): Ditto.
3626 (_mm_mask_i32scatter_ps): Ditto.
3627 (_mm256_i32scatter_pd): Ditto.
3628 (_mm256_mask_i32scatter_pd): Ditto.
3629 (_mm_i32scatter_pd): Ditto.
3630 (_mm_mask_i32scatter_pd): Ditto.
3631 (_mm256_i64scatter_ps): Ditto.
3632 (_mm256_mask_i64scatter_ps): Ditto.
3633 (_mm_i64scatter_ps): Ditto.
3634 (_mm_mask_i64scatter_ps): Ditto.
3635 (_mm256_i64scatter_pd): Ditto.
3636 (_mm256_mask_i64scatter_pd): Ditto.
3637 (_mm_i64scatter_pd): Ditto.
3638 (_mm_mask_i64scatter_pd): Ditto.
3639 (_mm256_i32scatter_epi32): Ditto.
3640 (_mm256_mask_i32scatter_epi32): Ditto.
3641 (_mm_i32scatter_epi32): Ditto.
3642 (_mm_mask_i32scatter_epi32): Ditto.
3643 (_mm256_i32scatter_epi64): Ditto.
3644 (_mm256_mask_i32scatter_epi64): Ditto.
3645 (_mm_i32scatter_epi64): Ditto.
3646 (_mm_mask_i32scatter_epi64): Ditto.
3647 (_mm256_i64scatter_epi32): Ditto.
3648 (_mm256_mask_i64scatter_epi32): Ditto.
3649 (_mm_i64scatter_epi32): Ditto.
3650 (_mm_mask_i64scatter_epi32): Ditto.
3651 (_mm256_i64scatter_epi64): Ditto.
3652 (_mm256_mask_i64scatter_epi64): Ditto.
3653 (_mm_i64scatter_epi64): Ditto.
3654 (_mm_mask_i64scatter_epi64): Ditto.
3655 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
3656 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
3657 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
3658 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
3659 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
3660 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
3661 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
3662 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
3663 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
3664 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
3665 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
3666 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
3667 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
3668 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
3669 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
3670 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
3671 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
3672 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
3673 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
3674 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
3675 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
3676 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
3677 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
3678 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
3679 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
3680 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
3681 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
3682 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
3683 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
3684 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
3685 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
3686 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
3687 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
3688 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
3689 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
3690 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
3691 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
3692 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
3693 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
3694 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
3695 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
3696 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
3697 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
3698 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
3699 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
3700 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
3701 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
3702 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
3703 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
3704 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
3705 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
3706 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
3707 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
3708 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
3709 definitions accordingly.
3711 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
3712 Kuan-Lin Chen <kuanlinchentw@gmail.com>
3715 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
3718 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
3721 * ira-conflicts.c (ira_build_conflicts): Update total conflict
3722 hard regs for inner regno.
3724 2017-01-17 Martin Liska <mliska@suse.cz>
3727 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
3728 assumption and add comment.
3730 2017-01-17 Nathan Sidwell <nathan@acm.org>
3732 * ipa-visibility.c (localize_node): New function, broken out of ...
3733 (function_and_variable_visibility): ... here. Call it.
3735 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
3738 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
3739 correctly set frequency of oudgoing edge.
3740 (duplicate_thread_path): Fix profile updating.
3742 2017-01-17 Jakub Jelinek <jakub@redhat.com>
3745 * configure.ac: Add GCC_BASE_VER.
3746 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
3747 version from BASE-VER file.
3748 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
3749 (gcc.o): Depend on $(BASEVER).
3750 * common.opt (dumpfullversion): New option.
3751 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
3752 * doc/invoke.texi: Document -dumpfullversion.
3753 * doc/install.texi: Document --with-gcc-major-version-only.
3754 * configure: Regenerated.
3756 2017-01-17 Richard Biener <rguenther@suse.de>
3758 PR tree-optimization/71433
3759 * tree-vrp.c (register_new_assert_for): Merge same asserts
3760 on all incoming edges.
3761 (process_assert_insertions_for): Handle insertions at the
3764 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
3766 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
3767 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
3769 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
3772 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
3775 2017-01-17 Alan Modra <amodra@gmail.com>
3778 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
3779 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
3780 symbolic stack limit when pic.
3782 2017-01-16 Martin Sebor <msebor@redhat.com>
3784 PR tree-optimization/78608
3785 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
3787 2017-01-16 Jeff Law <law@redhat.com>
3790 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
3791 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
3792 for several include directories that may be relative to sysroot.
3793 * config/i386/x-mingw32 (gplus_includedir): Define.
3794 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
3795 (native_system_includedir): Likewise.
3796 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
3797 override if TARGET_SYSTEM_ROOT is defined.
3798 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
3800 PR tree-optimization/79090
3801 PR tree-optimization/33562
3802 PR tree-optimization/61912
3803 PR tree-optimization/77485
3804 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
3805 and computed trims into the dump file.
3807 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
3809 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
3811 2017-01-16 Jakub Jelinek <jakub@redhat.com>
3814 * gimplify.c (gimplify_init_constructor): If want_value and
3815 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
3819 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
3820 sequence. Formatting fixes.
3821 (doloop_optimize): Formatting fixes.
3824 * gcc.c (debug_level_greater_than_spec_func): New function.
3825 (static_spec_functions): Add debug-level-gt spec function.
3826 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
3828 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
3829 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
3830 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
3831 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
3832 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
3833 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
3835 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
3837 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
3838 QImode fixups to general and mask registers only.
3840 2017-01-16 Carl Love <cel@us.ibm.com>
3842 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
3843 for built-in functions
3844 vector signed char vec_nabs (vector signed char)
3845 vector signed short vec_nabs (vector signed short)
3846 vector signed int vec_nabs (vector signed int)
3847 vector signed long long vec_nabs (vector signed long long)
3848 vector float vec_nabs (vector float)
3849 vector double vec_nabs (vector double)
3850 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
3852 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
3853 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
3854 * doc/extend.texi: Update the documentation file for the new built-in
3857 2017-01-16 Martin Sebor <msebor@redhat.com>
3859 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
3862 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3864 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
3865 UNSPEC_VSX__XXSPLTD to require special splat handling.
3867 2017-01-16 David Malcolm <dmalcolm@redhat.com>
3870 * system.h: Poison strndup.
3872 2017-01-16 Alan Modra <amodra@gmail.com>
3875 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
3878 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
3880 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
3882 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
3884 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
3885 call recog here. Assert that INSN_CODE (insn) is non-negative.
3887 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
3890 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
3892 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
3893 in the currently scheduled RTL fragment.
3895 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
3897 PR rtl-optimization/78751
3898 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
3901 2017-01-14 Jeff Law <law@redhat.com>
3903 PR tree-optimization/79090
3904 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
3905 variable length stores.
3906 (compute_trims): Delete dead assignment to *trim_tail.
3907 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
3910 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
3912 PR rtl-optimization/78626
3913 PR rtl-optimization/78727
3914 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
3915 of a block, and split such blocks after everything else is finished.
3917 2017-01-14 Alan Modra <amodra@gmail.com>
3920 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
3921 target legitimate_combined_insn.
3922 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
3923 (rs6000_legitimate_combined_insn): New function.
3924 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
3926 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
3927 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
3928 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
3930 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
3932 * doc/frontends.texi (G++ and GCC): Remove references to Java.
3934 2017-01-13 Jeff Law <law@redhat.com>
3936 PR tree-optimization/33562
3937 PR tree-optimization/61912
3938 PR tree-optimization/77485
3939 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
3941 (delete_dead_assignment): Likewise.
3942 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
3943 statement to delete_dead_call and delete_dead_assignment.
3945 2017-01-13 David Malcolm <dmalcolm@redhat.com>
3948 * substring-locations.c (format_warning_va): Strengthen case 1 so
3949 that both endpoints of the substring must be within the format
3950 range for just the substring to be printed.
3952 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
3954 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
3955 * config/i386/i386.c (ix86_target_string): Add missing options
3956 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
3957 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
3958 flags_other and ix86_target_other to flags2_other. Display unknown
3960 (ix86_valid_target_attribute_inner_p): Add missing options and
3961 reorder options by implied ISAs, as in ix86_target_string.
3963 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3965 * hash-table.h (hash_table::too_empty_p): New function.
3966 (hash_table::expand): Use it.
3967 (hash_table::traverse): Likewise.
3968 (hash_table::empty_slot): Use sizeof (value_type) instead of
3969 sizeof (PTR) to convert bytes to elements. Shrink the table
3970 if the current size is excessive for the current number of
3973 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3975 * ira-costs.c (record_reg_classes): Break from the inner loop
3976 early once alt_fail is known to be true. Update outer loop
3977 handling accordingly.
3979 2017-01-13 Jeff Law <law@redhat.com>
3981 * tree-ssa-dse.c (decrement_count): New function.
3982 (increment_start_addr, maybe_trim_memstar_call): Likewise.
3983 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
3984 when we know the partially dead statement is a mem* function.
3986 PR tree-optimization/61912
3987 PR tree-optimization/77485
3988 * tree-ssa-dse.c: Include expr.h.
3989 (maybe_trim_constructor_store): New function.
3990 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
3992 PR tree-optimization/33562
3993 PR tree-optimization/61912
3994 PR tree-optimization/77485
3995 * doc/invoke.texi: Document new dse-max-object-size param.
3996 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
3997 * tree-ssa-dse.c: Include params.h.
3998 (dse_store_status): New enum.
3999 (initialize_ao_ref_for_dse): New, partially extracted from
4001 (valid_ao_ref_for_dse, normalize_ref): New.
4002 (setup_live_bytes_from_ref, compute_trims): Likewise.
4003 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
4004 (maybe_trim_partially_dead_store): Likewise.
4005 (maybe_trim_complex_store): Likewise.
4006 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
4007 Track what bytes live from the original store. Return tri-state
4008 for dead, partially dead or live.
4009 (dse_dom_walker): Add constructor, destructor and new private members.
4010 (delete_dead_call, delete_dead_assignment): New extracted from
4012 (dse_optimize_stmt): Make a member of dse_dom_walker.
4013 Use initialize_ao_ref_for_dse.
4015 PR tree-optimization/33562
4016 PR tree-optimization/61912
4017 PR tree-optimization/77485
4018 * sbitmap.h (bitmap_count_bits): Prototype.
4019 (bitmap_clear_range, bitmap_set_range): Likewise.
4020 * sbitmap.c (bitmap_clear_range): New function.
4021 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
4023 2017-01-13 Martin Liska <mliska@suse.cz>
4026 * function.c (set_cfun): Add new argument force.
4027 * function.h (set_cfun): Likewise.
4028 * ipa-inline-transform.c (inline_call): Use the function when
4029 strict alising from is dropped for function we inline to.
4031 2017-01-13 Richard Biener <rguenther@suse.de>
4033 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
4034 for dumping GIMPLE INTEGER_CSTs.
4036 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4038 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
4039 to 201112L since C++17.
4041 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
4044 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
4045 if -fsanitize=kernel-address is present.
4047 2017-01-13 Richard Biener <rguenther@suse.de>
4049 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
4050 as _Literal ( type ) number in case usual suffixes do not
4051 preserve all information.
4053 2017-01-13 Richard Biener <rguenther@suse.de>
4055 PR tree-optimization/77283
4056 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
4057 and ssa-iterators.h.
4058 (is_feasible_trace): Implement a cost model based on joiner
4061 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4064 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
4065 char or short to __float128/_Float128 directly.
4067 2017-01-12 Martin Sebor <msebor@redhat.com>
4069 to -Wformat-overflow.
4070 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
4071 (min_bytes_remaining): Same.
4072 (get_string_length): Same.
4073 (format_string): Same.
4074 (format_directive): Same.
4076 (pass_sprintf_length::handle_gimple_call): Same.
4078 2017-01-12 Jakub Jelinek <jakub@redhat.com>
4080 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
4081 info.nowrite calls with no lhs that can't throw. Return bool
4082 whether gsi_remove has been called or not.
4083 (pass_sprintf_length::handle_gimple_call): Return bool whether
4084 try_substitute_return_value called gsi_remove. Formatting fix.
4085 (pass_sprintf_length::execute): Don't use gsi_remove if
4086 handle_gimple_call returned true.
4089 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
4090 be removed due to side-effects, don't remove following barrier nor
4091 turn the successor edge into fallthru edge.
4093 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4096 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
4097 element-reversing loads and stores as not swappable.
4099 2017-01-12 Nathan Sidwell <nathan@acm.org>
4100 Nicolai Stange <nicstange@gmail.com>
4102 * combine.c (try_combine): Don't ignore result of overlap checking
4103 loop. Combine overlap & asm check into single loop.
4105 2017-01-12 Richard Biener <rguenther@suse.de>
4107 * tree-pretty-print.c (dump_generic_node): Provide -gimple
4108 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
4110 2017-01-12 Richard Biener <rguenther@suse.de>
4112 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
4113 and TS_TARGET_OPTION directly derive from TS_BASE.
4114 * tree-core.h (tree_optimization_option): Derive from tree_base.
4115 (tree_target_option): Likewise.
4117 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
4119 * config/i386/i386.c (memory_address_length): Increase len
4120 only when rip_relative_addr_p returns false.
4122 2017-01-11 Julia Koval <julia.koval@intel.com>
4124 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
4125 (OPTION_MASK_ISA_SGX_SET): New.
4126 (ix86_handle_option): Handle OPT_msgx.
4127 * config.gcc: Added sgxintrin.h.
4128 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
4129 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
4130 * config/i386/i386.c (ix86_target_string): Add -msgx.
4132 (ix86_option_override_internal): Handle new options.
4133 (ix86_valid_target_attribute_inner_p): Add sgx.
4134 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
4135 * config/i386/i386.opt: Add msgx.
4136 * config/i386/sgxintrin.h: New file.
4137 * config/i386/x86intrin.h: Add sgxintrin.h.
4139 2017-01-11 Jakub Jelinek <jakub@redhat.com>
4142 * fold-const.c (maybe_nonzero_address): Return 1 for function
4144 (tree_single_nonzero_warnv_p): Don't handle function local objects
4148 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
4151 2017-01-11 David Malcolm <dmalcolm@redhat.com>
4154 * opts.c: Include "spellcheck.h"
4155 (struct string_fragment): New struct.
4156 (struct edit_distance_traits<const string_fragment &>): New
4158 (get_closest_sanitizer_option): New function.
4159 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
4161 2017-01-11 Jakub Jelinek <jakub@redhat.com>
4163 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
4165 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
4166 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
4167 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
4168 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
4169 for initial die_offset if dwarf_split_debug_info.
4170 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
4171 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
4173 (output_skeleton_debug_sections): Formatting fix. Use
4174 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
4175 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
4177 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
4179 * config/arm/cortex-a53.md: Add bypasses for
4181 (cortex_a53_r2f): Only use for transfers.
4182 (cortex_a53_f2r): Likewise.
4183 (cortex_a53_r2f_cvt): Add reservation for conversions.
4184 (cortex_a53_f2r_cvt): Likewise.
4186 2017-01-11 Tamar Christina <tamar.christina@arm.com>
4188 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
4189 to all inlined functions, change static to extern.
4191 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
4194 * config/arm/arm.c (legitimize_pic_address): Handle reference to
4196 (arm_assemble_integer): Likewise.
4198 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
4200 * config.gcc: Use new awk script to check CPU, FPU and architecture
4201 parameters for --with-... options.
4202 * config/arm/parsecpu.awk: New file
4203 * config/arm/arm-cpus.in: New file.
4204 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
4206 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
4208 * config/arm/t-arm: Update dependency rules.
4209 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
4210 of processing .def files.
4211 * config/arm/genopt.sh: Deleted.
4212 * config/arm/gentune.sh: Deleted.
4213 * config/arm/arm-cores.def: Deleted.
4214 * config/arm/arm-arches.def: Deleted.
4215 * config/arm/arm-fpus.def: Deleted.
4216 * config/arm/arm-tune.md: Regenerated.
4217 * config/arm/arm-tables.opt: Regenerated.
4218 * config/arm/arm-cpu.h: New generated file.
4219 * config/arm/arm-cpu-data.h: New generated file.
4220 * config/arm/arm-cpu-cdata.h: New generated file.
4222 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
4225 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
4227 (input_varpool_node): Unpack dynamically_initialized bit.
4229 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
4231 PR rtl-optimization/79032
4232 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
4233 the alignment of the adjusted memory reference against that of MODE,
4234 instead of the alignment of the original memory reference.
4236 2017-01-11 Martin Jambor <mjambor@suse.cz>
4238 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
4240 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
4241 decorated functions.
4243 2017-01-11 Richard Biener <rguenther@suse.de>
4245 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
4246 set range/nonnull info for PHI results. Do not set it on
4247 stmts marked for removal.
4249 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
4251 * expr.c (store_field): In the bitfield case, fetch the return value
4252 from the registers before applying a single big-endian adjustment.
4253 Always do a final load for a BLKmode value not larger than a word.
4255 2017-01-10 David Malcolm <dmalcolm@redhat.com>
4258 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
4259 that we correctly handle column numbers greater than
4260 LINE_MAP_MAX_COLUMN_NUMBER.
4262 2017-01-10 Martin Sebor <msebor@redhat.com>
4265 * gimple-ssa-sprintf.c (get_destination_size): Call
4266 {init,fini}object_sizes.
4267 * tree-object-size.c (addr_object_size): Adjust.
4268 (pass_through_call): Adjust.
4269 (pass_object_sizes::execute): Adjust.
4270 * tree-object-size.h (fini_object_sizes): Declare.
4272 2017-01-10 Martin Sebor <msebor@redhat.com>
4274 PR tree-optimization/78775
4275 * builtins.c (get_size_range): Move...
4276 * calls.c: ...to here.
4277 (alloc_max_size): Accept zero argument.
4278 (operand_signed_p): Remove.
4279 (maybe_warn_alloc_args_overflow): Call get_size_range.
4280 * calls.h (get_size_range): Declare.
4282 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
4284 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
4285 from TI's devices.csv file as of September 2016.
4286 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
4288 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
4290 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
4291 * doc/invoke.texi: Likewise.
4292 * doc/md.texi: Likewise.
4293 * doc/objc.texi: Likewise.
4295 2017-01-10 Joshua Conner <joshconner@google.com>
4297 * config/arm/fuchsia-elf.h: New file.
4298 * config/fuchsia.h: New file.
4299 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
4300 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
4302 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
4304 2016-01-10 Richard Biener <rguenther@suse.de>
4306 PR tree-optimization/79034
4307 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
4308 Propagate out degenerate PHIs in the joiner.
4310 2017-01-10 Martin Liska <mliska@suse.cz>
4312 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
4313 (sort_congruence_classes_by_decl_uid): Likewise.
4314 (sort_congruence_class_groups_by_decl_uid): Likewise.
4315 (sem_item_optimizer::merge_classes): Sort class, groups in these
4316 classes and members in the groups by DECL_UID of declarations.
4317 This would make merge operations stable.
4319 2017-01-10 Martin Liska <mliska@suse.cz>
4321 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
4322 usage of m_classes_vec.
4323 (sem_item_optimizer::~sem_item_optimizer): Likewise.
4324 (sem_item_optimizer::get_group_by_hash): Likewise.
4325 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
4326 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
4327 (sem_item_optimizer::verify_classes): Likewise.
4328 (sem_item_optimizer::process_cong_reduction): Likewise.
4329 (sem_item_optimizer::dump_cong_classes): Likewise.
4330 (sem_item_optimizer::merge_classes): Likewise.
4331 * ipa-icf.h (congruence_class_hash): Rename from
4332 congruence_class_group_hash. Remove declaration of m_classes_vec.
4334 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
4336 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
4337 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
4338 * config.gcc: Add avx512vpopcntdqintrin.h.
4339 * config/i386/avx512vpopcntdqintrin.h: New.
4340 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
4341 * config/i386/i386-builtin-types.def: Add new types.
4342 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
4343 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
4344 __builtin_ia32_vpopcountq_v8di_mask): New.
4345 * config/i386/i386-c.c (ix86_target_macros_internal): Define
4346 __AVX512VPOPCNTDQ__.
4347 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
4348 (PTA_AVX512VPOPCNTDQ): Define.
4349 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
4350 TARGET_AVX512VPOPCNTDQ_P): Define.
4351 * config/i386/i386.opt: Add mavx512vpopcntdq.
4352 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
4353 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
4355 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4358 * predict.def (PRED_CALL): Set to 67.
4360 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
4362 * expr.c (store_field): In the bitfield case, if the value comes from
4363 a function call and is of an aggregate type returned in registers, do
4364 not modify the field mode; extract the value in all cases if the mode
4365 is BLKmode and the size is not larger than a word.
4367 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
4370 * config/i386/cpuid.h: Fix undefined behavior.
4372 2017-01-04 Jeff Law <law@redhat.com>
4374 PR tree-optimization/79007
4375 PR tree-optimization/67955
4376 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
4377 conservative for pt.null when flag_non_call_exceptions is on.
4379 2017-01-09 Jakub Jelinek <jakub@redhat.com>
4381 PR translation/79019
4382 PR translation/79020
4383 * params.def (PARAM_INLINE_MIN_SPEEDUP,
4384 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
4385 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
4387 * config/avr/avr.opt (maccumulate-args): Likewise.
4388 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
4389 * common.opt (freport-bug): Likewise.
4390 * cif-code.def (CIF_FINAL_ERROR): Likewise.
4391 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
4392 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
4393 translatable string.
4394 * config/i386/i386.c (function_value_32): Likewise.
4395 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
4396 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
4398 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
4399 * common/config/msp430/msp430-common.c (msp430_handle_option):
4401 * symtab.c (symtab_node::verify_base): Likewise.
4402 * opts.c (set_debug_level): Likewise.
4403 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
4404 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
4405 missing whitespace to translatable strings.
4406 * config/avr/avr.md (bswapsi2): Fix typo in comment.
4407 * config/sh/superh.h: Likewise.
4408 * config/i386/xopintrin.h: Likewise.
4409 * config/i386/znver1.md: Likewise.
4410 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
4411 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
4412 * double-int.h (struct double_int): Likewise.
4413 * double-int.c (div_and_round_double): Likewise.
4414 * wide-int.cc: Likewise.
4415 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4416 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
4417 * cfgcleanup.c (crossjumps_occured): Renamed to ...
4418 (crossjumps_occurred): ... this.
4419 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
4422 PR tree-optimization/78899
4423 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
4424 returning bool return struct loop *, NULL for failure and the new
4426 (versionable_outer_loop_p): Don't version outer loop if it has
4427 dont_vectorized bit set.
4428 (tree_if_conversion): When versioning outer loop, ensure
4429 tree_if_conversion is performed also on the inner loop of the
4430 non-vectorizable outer loop copy.
4431 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
4432 LOOP_VECTORIZED in inner loop of the scalar outer loop and
4433 prevent vectorization of it.
4434 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
4435 the outer loop vectorization of the non-scalar version is attempted
4436 before vectorization of the inner loop in scalar version. If
4437 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
4438 vectorization of its inner loop.
4439 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
4440 has 2 inner loops, rename also on edges from bb whose single pred
4441 is outer_loop->header. Fix typo in function comment.
4443 2017-01-09 Martin Sebor <msebor@redhat.com>
4446 * asan.c (asan_emit_stack_protection): Increase local buffer size
4447 to avoid snprintf truncation warning.
4449 2017-01-09 Andrew Pinski <apinski@cavium.com>
4451 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
4452 to reference thunderx2t99 for the tuning structure
4453 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
4455 (thunderx2t99_extra_costs): This.
4456 * config/aarch64/aarch64-tune.md: Regenerate.
4457 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
4458 (vulcan_addrcost_table): This.
4459 (vulcan_regmove_cost): Rename to ...
4460 (thunderx2t99_regmove_cost): This.
4461 (vulcan_vector_cost): Rename to ...
4462 (thunderx2t99_vector_cost): this.
4463 (vulcan_branch_cost): Rename to ...
4464 (thunderx2t99_branch_cost): This.
4465 (vulcan_tunings): Rename to ...
4466 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
4467 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
4469 2017-01-09 Martin Jambor <mjambor@suse.cz>
4473 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
4474 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
4475 (propagate_vr_accross_jump_function): Use the above function for all
4476 value range computations for pass-through jump functions and type
4477 converasion from explicit value range values.
4478 (ipcp_propagate_stage): Do not attempt to deduce types of formal
4479 parameters from TYPE_ARG_TYPES.
4480 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
4481 (ipa_write_node_info): Stream type of the actual argument.
4482 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
4484 2017-01-09 Martin Liska <mliska@suse.cz>
4487 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
4488 (lookup_compiler): Do not show error message with have_E.
4490 2017-01-09 Jakub Jelinek <jakub@redhat.com>
4492 PR tree-optimization/78938
4493 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
4494 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
4495 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
4496 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
4499 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4501 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
4504 2017-01-09 Richard Biener <rguenther@suse.de>
4506 PR tree-optimization/78997
4507 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
4508 name condition properly.
4510 2017-01-09 Richard Biener <rguenther@suse.de>
4513 * dwarf2out.c (is_cxx): New overload with context.
4514 (is_naming_typedef_decl): Use it.
4516 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
4518 * invoke.texi (Option Summary): Correct spacing in option lists
4519 and add line breaks to fix over-long lines.
4521 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
4525 * extend.texi (Common Variable Attributes): Add xref to GCC
4526 Internals manual to explain mode attribute keywords.
4528 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
4531 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
4532 and Preprocessor Options.
4533 (Options for Linking): Document -pthread here....
4534 (RS/6000 and PowerPC Options): ...not here.
4535 (Solaris 2 Options): ...or here.
4536 * doc/cppopts.texi: Document -pthread.
4538 2017-01-08 Martin Sebor <msebor@redhat.com>
4541 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
4542 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
4543 New member functions.
4544 (format_directive): Used them.
4546 (pass_sprintf_length::handle_gimple_call): Same.
4547 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
4548 to avoid truncation for any argument.
4549 (extract_affine_mul): Same.
4550 * tree.c (get_file_function_name): Same.
4552 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4555 * predict.def (PRED_INDIR_CALL): Set to 86.
4557 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
4559 PR preprocessor/54124
4560 * doc/cppopts.texi: Reformat -d subtable to list the full name
4561 of the options. Add cross-reference to the docs for the general
4562 compiler -d options.
4563 * doc/invoke.texi (Developer Options): Add cross-reference to the
4564 preprocessor-specific -d option documentation.
4566 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
4568 PR preprocessor/13498
4569 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
4570 redudant material, and reflect new command-line options.
4571 (System Headers): Likewise.
4573 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
4575 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
4576 -isystem, and -idirafter. Copy-edit.
4577 * doc/cppopts.texi: Copy-edit. Remove contradiction about
4578 default for -ftrack-macro-expansion. Delete obsolete and
4579 badly-formatted implementation details about -fdebug-cpp output.
4580 * doc/cppwarnopts.texi: Copy-edit.
4582 2017-01-07 David Malcolm <dmalcolm@redhat.com>
4585 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
4586 that the transition from a max line width >= 1<<10 to narrower
4587 lines works correctly.
4589 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
4591 * doc/options.texi (PerFunction): New.
4592 * opt-functions.awk (switch_flags): Map both Optimization and
4593 PerFunction to CL_OPTIMIZATION.
4594 * opth-gen.awk: Test for PerFunction flag along with
4596 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
4597 it only when the latter is present. Skip those that don't in
4598 the hash function generator.
4599 * common.opt (fvar-tracking): Mark as PerFunction instead of
4601 (fvar-tracking-assignments): Likewise.
4602 (fvar-tracking-assignments-toggle): Likewise.
4603 (fvar-tracking-uninit): Likewise.
4605 2017-01-07 Jakub Jelinek <jakub@redhat.com>
4607 PR translation/79018
4608 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
4611 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
4614 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
4615 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
4616 TARGET_LONG_JUMP_TABLE_OFFSETS.
4617 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
4618 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
4619 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
4620 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
4621 * config/m68k/m68k.md (tablejump expander): Likewise.
4622 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
4623 TARGET_LONG_JUMP_TABLE_OFFSETS.
4624 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
4625 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
4627 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4628 David Holsgrove <david.holsgrove@xilinx.com>
4630 * common/config/microblaze/microblaze-common.c
4631 (TARGET_EXCEPT_UNWIND_INFO): Remove.
4632 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
4634 * config/microblaze/microblaze.c (microblaze_must_save_register)
4635 (microblaze_expand_epilogue, microblaze_return_addr): Handle
4637 (microblaze_eh_return): New function.
4638 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
4639 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
4640 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
4641 * config/microblaze/microblaze.md (eh_return): New pattern.
4643 2017-01-06 Jakub Jelinek <jakub@redhat.com>
4645 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
4646 GCC_DIAGNOSTIC_STRINGIFY): Define.
4648 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
4650 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4652 * config/arm/arm.md (<mcrr>): New.
4654 * config/arm/arm.c (arm_arch5te): New.
4655 (arm_option_override): Set arm_arch5te.
4656 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
4658 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
4659 (arm_mcrr_qualifiers): ... this. New.
4660 (MRRC_QUALIFIERS): Define to...
4661 (arm_mrrc_qualifiers): ... this. New.
4662 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
4664 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
4665 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
4666 (MRRCI, mrrc, MRRC): New.
4667 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
4668 VUNSPEC_MRRC2): New.
4670 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4672 * config/arm/arm.md (<mcr>): New.
4674 * config/arm/arm.c (arm_coproc_builtin_available): Add
4675 support for mcr, mrc, mcr2 and mrc2.
4676 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
4677 (arm_mcr_qualifiers): ... this. New.
4678 (MRC_QUALIFIERS): Define to ...
4679 (arm_mrc_qualifiers): ... this. New.
4680 (MCR_QUALIFIERS): Define to ...
4681 (arm_mcr_qualifiers): ... this. New.
4682 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
4684 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
4685 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
4686 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
4689 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4691 * config/arm/arm.md (*ldc): New.
4695 * config/arm/arm.c (arm_coproc_builtin_available): Add
4696 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
4697 (arm_coproc_ldc_stc_legitimate_address): New.
4698 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
4699 'qualifier_const_pointer'.
4700 (LDC_QUALIFIERS): Define to...
4701 (arm_ldc_qualifiers): ... this. New.
4702 (STC_QUALIFIERS): Define to...
4703 (arm_stc_qualifiers): ... this. New.
4704 * config/arm/arm-protos.h
4705 (arm_coproc_ldc_stc_legitimate_address): New.
4706 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
4707 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
4708 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
4709 stc2, stcl, stc2l): New.
4710 * config/arm/constraints.md (Uz): New.
4711 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
4712 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
4713 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
4714 VUNSPEC_STC2L): New.
4716 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4718 * config/arm/arm.md (<cdp>): New.
4719 * config/arm/arm.c (neon_const_bounds): Rename this ...
4720 (arm_const_bounds): ... this.
4721 (arm_coproc_builtin_available): New.
4722 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
4723 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
4724 (CDP_QUALIFIERS): Define to...
4725 (arm_cdp_qualifiers): ... this. New.
4727 (arm_expand_builtin_args): Add case for 6 arguments.
4728 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
4729 (arm_const_bounds): ... this.
4730 (arm_coproc_builtin_available): New.
4731 * config/arm/arm_acle.h (__arm_cdp): New.
4733 * config/arm/arm_acle_builtins.def (cdp): New.
4735 * config/arm/iterators.md (CDPI,CDP,cdp): New.
4736 * config/arm/neon.md: Rename all 'neon_const_bounds' to
4738 * config/arm/types.md (coproc): New.
4739 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
4740 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
4741 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
4742 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
4744 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4746 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
4747 (UBINOP_QUALIFIERS): New.
4749 (acle_builtin_data): New. Change comment.
4750 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
4751 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
4752 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
4753 arm_acle_builtins.def.
4754 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
4755 (arm_init_acle_builtins): New.
4756 (CRC32_BUILTIN): Remove.
4757 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
4758 crc32cb, crc32ch and crc32cw.
4759 (arm_init_crc32_builtins): Remove.
4760 (arm_init_builtins): Use arm_init_acle_builtins rather
4761 than arm_init_crc32_builtins.
4762 (arm_expand_acle_builtin): New.
4763 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
4764 * config/arm/arm_acle_builtins.def: New.
4766 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
4768 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
4769 (arm_builtin_datum): ... this.
4770 (arm_init_neon_builtin): Rename to ...
4771 (arm_init_builtin): ... this. Add a new parameters PREFIX
4772 and USE_SIG_IN_NAME.
4773 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
4774 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
4775 'arm_builtin_datum'.
4776 (arm_init_vfp_builtins): Likewise.
4777 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
4778 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
4779 (arm_expand_neon_args): Rename to ...
4780 (arm_expand_builtin_args): ... this. Rename builtin_arg
4781 enum values and differentiate between ARG_BUILTIN_MEMORY
4782 and ARG_BUILTIN_NEON_MEMORY.
4783 (arm_expand_neon_builtin_1): Rename to ...
4784 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
4785 values, arm_expand_builtin_args and add bool parameter NEON.
4786 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
4787 (arm_expand_vfp_builtin): Likewise.
4788 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
4790 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4793 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
4794 * predict.c (tree_estimate_probability_bb): Reverse direction of
4795 polymorphic call predictor.
4797 2017-01-06 David Malcolm <dmalcolm@redhat.com>
4799 * passes.c (execute_one_pass): Split out pass-skipping logic into...
4800 (determine_pass_name_match): ...this new function and...
4801 (should_skip_pass_p): ...this new function.
4803 2017-01-06 Nathan Sidwell <nathan@acm.org>
4805 * ipa-visibility.c (function_and_variable_visibility): Reformat
4806 comments and long lines. Remove extrneous if.
4807 * symtab.c (symtab_node::make_decl_local): Fix code format.
4808 (symtab_node::set_section_for_node): Fix comment typo.
4810 2017-01-06 Martin Liska <mliska@suse.cz>
4813 * lra-constraints.c: Rename invariant to lra_invariant.
4814 * predict.c (set_even_probabilities): Initialize e to NULL.
4816 2017-01-05 Martin Sebor <msebor@redhat.com>
4818 PR tree-optimization/78910
4819 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
4820 (format_integer): Correct off-by-one error in the handling
4821 of precision with negative numbers in signed conversions..
4823 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
4825 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
4827 2017-01-05 Jakub Jelinek <jakub@redhat.com>
4829 PR tree-optimization/71016
4830 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
4831 factor_out_conditional_conversion. Formatting fix.
4832 (factor_out_conditional_conversion): Add cond_stmt argument.
4833 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
4834 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
4837 2017-01-05 David Malcolm <dmalcolm@redhat.com>
4839 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
4840 read-rtl-function.o, and selftest-rtl.o.
4841 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
4842 (selftest::aarch64_test_loading_full_dump): New function.
4843 (selftest::aarch64_run_selftests): New function.
4844 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
4845 selftest::aarch64_run_selftests.
4846 * config/i386/i386.c
4847 (selftest::ix86_test_loading_dump_fragment_1): New function.
4848 (selftest::ix86_test_loading_call_insn): New function.
4849 (selftest::ix86_test_loading_full_dump): New function.
4850 (selftest::ix86_test_loading_unspec): New function.
4851 (selftest::ix86_run_selftests): Call the new functions.
4852 * emit-rtl.c (maybe_set_max_label_num): New function.
4853 * emit-rtl.h (maybe_set_max_label_num): New decl.
4854 * function.c (instantiate_decls): Guard call to
4855 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
4856 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
4858 * gensupport.c (gen_reader::gen_reader): Pass "false"
4859 for new "compact" param of rtx_reader.
4860 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
4861 rather than an empty string for NULL strings.
4862 * read-md.c: Potentially include config.h rather than bconfig.h.
4863 Wrap include of errors.h with #ifdef GENERATOR_FILE.
4864 (have_error): New global, copied from errors.c.
4865 (md_reader::read_name): Rename to...
4866 (md_reader::read_name_1): ...this, adding "out_loc" param,
4867 and converting "missing name or number" to returning false, rather
4869 (md_reader::read_name): Reimplement in terms of read_name_1.
4870 (md_reader::read_name_or_nil): New function.
4871 (md_reader::read_string): Handle "(nil)" by returning NULL.
4872 (md_reader::md_reader): Add new param "compact".
4873 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
4874 (md_reader::read_file): New method.
4875 * read-md.h (md_reader::md_reader): Add new param "compact".
4876 (md_reader::read_file): New method.
4877 (md_reader::is_compact): New accessor.
4878 (md_reader::read_name): Convert return type from void to file_location.
4879 (md_reader::read_name_or_nil): New decl.
4880 (md_reader::read_name_1): New decl.
4881 (md_reader::m_compact): New field.
4882 (noop_reader::noop_reader): Pass "false" for new "compact" param
4884 (rtx_reader::rtx_reader): Add new "compact" param.
4885 (rtx_reader::read_rtx_operand): Make virtual and convert return
4886 type from void to rtx.
4887 (rtx_reader::read_until): New decl.
4888 (rtx_reader::handle_any_trailing_information): New virtual function.
4889 (rtx_reader::postprocess): New virtual function.
4890 (rtx_reader::finalize_string): New virtual function.
4891 (rtx_reader::m_in_call_function_usage): New field.
4892 (rtx_reader::m_reuse_rtx_by_id): New field.
4893 * read-rtl-function.c: New file.
4894 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
4895 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
4896 (selftest::verify_three_block_rtl_cfg): New decl.
4897 * read-rtl-function.h: New file.
4898 * read-rtl.c: Potentially include config.h rather than bconfig.h.
4899 For host, include function.h, memmodel.h, and emit-rtl.h.
4900 (one_time_initialization): New function.
4901 (struct compact_insn_name): New struct.
4902 (compact_insn_names): New array.
4903 (find_code): Handle insn codes in compact dumps.
4904 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
4905 (bind_subst_iter_and_attr): Likewise.
4906 (add_condition_to_string): Likewise.
4907 (add_condition_to_rtx): Likewise.
4908 (apply_attribute_uses): Likewise.
4909 (add_current_iterators): Likewise.
4910 (apply_iterators): Likewise.
4911 (initialize_iterators): Guard usage of apply_subst_iterator with
4912 #ifdef GENERATOR_FILE.
4913 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
4914 (md_reader::read_mapping): Likewise.
4915 (add_define_attr_for_define_subst): Likewise.
4916 (add_define_subst_attr): Likewise.
4917 (read_subst_mapping): Likewise.
4918 (check_code_iterator): Likewise.
4919 (rtx_reader::read_rtx): Likewise. Move one-time initialization
4921 (one_time_initialization): New function.
4922 (rtx_reader::read_until): New method.
4923 (read_flags): New function.
4924 (parse_reg_note_name): New function.
4925 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
4926 Handle reuse_rtx ids.
4927 Wrap iterator lookup within #ifdef GENERATOR_FILE.
4928 Add parsing support for RTL dumps, mirroring the special-cases in
4929 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
4930 values, and calling handle_any_trailing_information.
4931 (rtx_reader::read_rtx_operand): Convert return type from void
4932 to rtx, returning return_rtx. Handle case 'e'. Call
4933 finalize_string on XSTR and XTMPL fields.
4934 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
4935 "(nil)" values were omitted. Call the postprocess vfunc on the
4937 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
4938 class ctor. Initialize m_in_call_function_usage. Call
4939 one_time_initialization.
4940 * rtl-tests.c (selftest::test_uncond_jump): Call
4941 set_new_first_and_last_insn.
4942 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
4943 * selftest-rtl.c: New file.
4944 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
4945 (selftest::get_insn_by_uid): New decl.
4946 * selftest-run-tests.c (selftest::run_tests): Call
4947 read_rtl_function_c_tests.
4948 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
4949 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
4952 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
4954 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
4955 operands in a special way. Assert that pos+len <= mode precision.
4957 2017-01-05 Jakub Jelinek <jakub@redhat.com>
4959 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
4960 3 argument Alias with unlimited for the negative form.
4961 (fno-vect-cost-model): Removed.
4963 2017-01-05 Martin Liska <mliska@suse.cz>
4965 * hsa-gen.c (gen_hsa_divmod): New function.
4966 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
4968 2017-01-05 Martin Liska <mliska@suse.cz>
4971 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
4974 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4976 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
4977 small constant length operands.
4979 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4981 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
4982 between loop iterations.
4984 2017-01-05 Martin Liska <mliska@suse.cz>
4987 * gimplify.c (gimplify_decl_expr): Compare to
4988 asan_poisoned_variables instread of checking flags.
4989 (gimplify_target_expr): Likewise.
4990 (gimplify_expr): Likewise.
4991 (gimplify_function_tree): Conditionally initialize
4992 asan_poisoned_variables.
4994 2017-01-04 Jeff Law <law@redhat.com>
4996 PR tree-optimizatin/78812
4997 * rtl.h (contains_mem_rtx_p): Prototype.
4998 * ifcvt.c (containts_mem_rtx_p): Move from here to...
4999 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
5000 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
5001 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
5002 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
5004 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
5006 * input.c (assert_char_at_range): Default-initialize actual_range.
5008 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
5010 * df-scan.c (df_ref_create_structure): Make regno unsigned,
5011 to match the caller.
5013 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
5015 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
5016 insns after final jump in test to emit dummy move.
5018 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
5020 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
5021 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
5023 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
5025 * multiple_target.c (create_dispatcher_calls): Init e_next.
5026 * tree-ssa-loop-split.c (split_loop): Init border.
5027 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
5030 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
5035 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
5036 (altivec_register_operand): Do not return true if the operand
5037 contains a SUBREG mixing SImode and SFmode.
5038 (vsx_register_operand): Likewise.
5039 (vsx_reg_sfsubreg_ok): New predicate.
5040 (vfloat_operand): Do not return true if the operand contains a
5041 SUBREG mixing SImode and SFmode.
5042 (vint_operand): Likewise.
5043 (vlogical_operand): Likewise.
5044 (gpc_reg_operand): Likewise.
5045 (int_reg_operand): Likewise.
5046 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
5047 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
5048 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
5050 (rs6000_emit_move_si_sf_subreg): New helper function.
5051 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
5052 fixup SUBREGs involving SImode and SFmode.
5053 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
5054 numbers for the new peephole2 optimization.
5055 (peephole2 for SFmode unions): New peephole2 to optimize cases in
5056 the GLIBC math library that do AND/IOR/XOR operations on single
5057 precision floating point.
5058 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
5059 target macros to say whether we need to avoid SUBREGs mixing
5061 (TARGET_ALLOW_SF_SUBREG): Likewise.
5062 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
5063 (UNSPEC_SI_FROM_SF): Likewise.
5064 (iorxor): Change spacing.
5065 (and_ior_xor): New iterator for AND, IOR, and XOR.
5066 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
5067 (movdi_from_sf_zero_ext): Likewise.
5068 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
5069 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
5070 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
5071 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
5072 (fms<mode>4): Likewise.
5073 (fnma<mode>4): Likewise.
5074 (fnms<mode>4): Likewise.
5075 (nfma<mode>4): Likewise.
5076 (nfms<mode>4): Likewise.
5078 2017-01-04 Marek Polacek <polacek@redhat.com>
5081 * doc/invoke.texi: Document -Wpointer-compare.
5083 2017-01-04 Jakub Jelinek <jakub@redhat.com>
5085 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
5088 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
5089 descriptions for -gdwarf-5 and emit them as uleb128 instead of
5092 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
5095 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
5096 documentation of the powerpc_popcntb_ok attribute.
5097 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
5098 code to issue warning messages if a requested CPU configuration is
5099 not supported by the binary (assembler and loader) toolchain.
5100 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
5101 made to define a built-in function that has been disabled.
5102 (paired_init_builtins): Add assertion to prevent ICE if attempt is
5103 made to define a built-in function that has been disabled.
5104 (altivec_init_builtins): Add comment explaining why definition
5105 of the DST built-in functions is not preceded by an assertion
5106 check. Add assertions to prevent ICE if attempts are made to
5107 define an altivec predicate or an abs* built-in function that has
5109 (htm_init_builtins): Add comment explaining why definition of the
5110 htm built-in functions is not preceded by an assertion check.
5112 2017-01-04 Jeff Law <law@redhat.com>
5114 PR tree-optimizatin/67955
5115 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
5116 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
5117 the points-to solution does not include pt_null. Use DECL_PT_UID
5120 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
5122 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
5123 Use gen_int_mode instead of gen_lopwart for const_int operands.
5125 2017-01-04 Jakub Jelinek <jakub@redhat.com>
5127 PR tree-optimization/71563
5128 * match.pd: Simplify X << Y into X if Y is known to be 0 or
5129 out of range value - has low bits known to be zero.
5131 2017-01-04 Alan Modra <amodra@gmail.com>
5133 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
5134 * configure: Regenerate.
5135 * config.in: Regenerate.
5137 2017-01-04 Jakub Jelinek <jakub@redhat.com>
5140 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
5141 a substring of the message, but strcmp with the whole message. Ifdef
5142 ENABLE_NLS, translate the message first using dgettext.
5144 2017-01-03 Jeff Law <law@redhat.com>
5146 PR tree-optimizatin/78856
5147 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
5148 (mark_threaded_blocks): Remove code to truncate thread paths that
5149 cross multiple loop headers. Instead invalidate the cached loop
5150 iteration information and handle case of a thread path walking
5151 into an irreducible region.
5153 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5156 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
5157 assertions. Add support for doing the signbit if the IEEE 128-bit
5158 floating point value is in a GPR.
5159 * config/rs6000/rs6000.md (Fsignbit): Delete.
5160 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
5161 Update the length attribute if the value is in a GPR.
5162 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
5163 the sign or zero extension instruction, since the value is always 0/1.
5164 (signbit<mode>2_dm2): Delete using <Fsignbit>.
5167 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
5168 extracting SImode to a GPR register so that we can generate a
5169 store, limit the vector to be in a traditional Altivec register
5170 for the vextuwrx instruction.
5172 2017-01-03 Ian Lance Taylor <iant@google.com>
5174 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
5176 2017-01-03 Martin Sebor <msebor@redhat.com>
5178 PR tree-optimization/78696
5179 * gimple-ssa-sprintf.c (format_floating): Correct handling of
5180 precision. Use MPFR for %f for greater fidelity. Correct handling
5182 (pass_sprintf_length::compute_format_length): Set width and precision
5183 specified by asrerisk to void_node for vararg functions.
5184 (try_substitute_return_value): Adjust dump output.
5186 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
5188 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
5190 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
5192 * doc/invoke.texi (SPARC options): Document -mlra as the default.
5193 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
5194 -mlra/-mno-lra was passed to the compiler.
5196 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
5198 PR rtl-optimization/65618
5199 * emit-rtl.c (try_split): Move initialization of "before" and
5200 "after" to just before the call to emit_insn_after_setloc.
5202 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
5204 * doc/md.texi (Standard Names): Remove reference to Java frontend.
5206 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
5208 * dwarf2out.c (gen_enumeration_type_die): When
5209 -gno-strict-dwarf, add a DW_AT_encoding attribute.
5211 2017-01-03 Jakub Jelinek <jakub@redhat.com>
5213 PR tree-optimization/78965
5214 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
5215 Change first argument from const call_info & to call_info &. For %n
5216 set info.nowrite to false.
5219 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
5220 possibly throwing calls.
5222 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
5223 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
5224 and fns handling, rather than in a separate case SSA_NAME.
5226 2017-01-02 Jeff Law <law@redhat.com>
5228 * config/darwin-driver.c (darwin_driver_init): Const-correctness
5229 fixes for first_period and second_period variables.
5231 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
5234 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
5235 (*insvqi_1): New insn pattern.
5236 (*insvqi_1_mem_rex64): Ditto.
5238 (*insvqi_3): Rename from *insvqi.
5240 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
5242 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
5244 * doc/cfg.texi (Edges): Remove reference to Java.
5245 (Maintaining the CFG): Ditto.
5247 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
5250 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
5251 transparent aliases.
5253 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
5256 * predict.def (PRED_CALL): Update hitrate.
5257 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
5258 * predict.c (tree_estimate_probability_bb): Split CALL predictor
5259 into direct/indirect/polymorphic variants.
5261 2017-01-01 Jakub Jelinek <jakub@redhat.com>
5263 Update copyright years.
5265 * gcc.c (process_command): Update copyright notice dates.
5266 * gcov-dump.c (print_version): Ditto.
5267 * gcov.c (print_version): Ditto.
5268 * gcov-tool.c (print_version): Ditto.
5269 * gengtype.c (create_file): Ditto.
5270 * doc/cpp.texi: Bump @copying's copyright year.
5271 * doc/cppinternals.texi: Ditto.
5272 * doc/gcc.texi: Ditto.
5273 * doc/gccint.texi: Ditto.
5274 * doc/gcov.texi: Ditto.
5275 * doc/install.texi: Ditto.
5276 * doc/invoke.texi: Ditto.
5278 Copyright (C) 2017 Free Software Foundation, Inc.
5280 Copying and distribution of this file, with or without modification,
5281 are permitted in any medium without royalty provided the copyright
5282 notice and this notice are preserved.