1 2017-02-03 Martin Liska <mliska@suse.cz>
3 * cgraph.c (cgraph_node::dump): Dump function version info.
4 * symtab.c (symtab_node::dump_base): Add missing new line.
6 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
8 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
9 (ifcombine_ifandif): Use it.
11 2017-02-03 Martin Liska <mliska@suse.cz>
13 * doc/invoke.texi: Document default value for
14 use-after-scope-direct-emission-threshold.
16 2017-02-03 Martin Liska <mliska@suse.cz>
18 PR tree-optimization/79339
19 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
20 (format_floating): Likewise.
22 2017-02-03 Martin Liska <mliska@suse.cz>
25 * ipa-prop.c (ipa_node_params_t::insert): Remove current
27 (ipa_node_params_t::remove): Likewise.
28 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
29 initialization from removed ipa_node_params_t::insert.
30 (ipa_node_params::~ipa_node_params): Move from removed
31 ipa_node_params_t::release.
32 * symbol-summary.h (symbol_summary::m_released): New member.
33 Do not release a summary twice. Do not allow to call finalizer
34 for types of a summary that live in GGC memory.
36 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
38 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
41 2017-02-02 Martin Sebor <msebor@redhat.com>
44 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
45 (format_string): Tighten up the range of output for non-constant
46 strings and correct the expected range for wide non-constant strings.
48 2017-02-02 Martin Sebor <msebor@redhat.com>
50 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
53 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
55 (-fdump-tree-@var): Add to index and document how to come up
56 with pass-specific option and dump file names.
57 (-fdump-passes): Clarify where to look for output.
59 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
62 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
63 statistics of the analyzed path; allow threading for speed when
64 any of BBs along the path are optimized for speed.
66 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
69 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
70 settings of the virtual registers.
73 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
75 * explow.c (get_dynamic_stack_size): Take known alignment of stack
76 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
79 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
81 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
82 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
84 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
86 * config/s390/s390.md: Add missing comments with the expanded
88 * config/s390/vector.md: Likewise.
89 * config/s390/vx-builtins.md: Likewise.
91 2017-02-02 Jakub Jelinek <jakub@redhat.com>
94 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
95 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
96 conditions on a single line.
98 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
100 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
101 __S390_VX__ to __VX__.
103 2017-02-01 Andrew Pinski <apinski@cavium.com>
105 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
106 stmt_info to record_stmt_cost.
107 (vect_get_known_peeling_cost): Pass stmt_info if known to
109 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
110 cpu_vector_cost field into
111 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
112 field into vec_int_stmt_cost and vec_fp_stmt_cost.
113 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
114 splitting of scalar_stmt_cost and vec_stmt_cost.
115 (thunderx_vector_cost): Likewise.
116 (cortexa57_vector_cost): LIkewise.
117 (exynosm1_vector_cost): Likewise.
118 (xgene1_vector_cost): Likewise.
119 (thunderx2t99_vector_cost): Improve after the splitting of the two
121 (aarch64_builtin_vectorization_cost): Update for the splitting of
122 scalar_stmt_cost and vec_stmt_cost.
124 2017-02-01 Torvald Riegel <triegel@redhat.com>
125 Richard Henderson <rth@redhat.com>
127 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
128 conditional on existance of a fast atomic load.
129 * optabs-query.c (can_atomic_load_p): New function.
130 * optabs-query.h (can_atomic_load_p): Declare it.
131 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
132 no fast atomic load is available for the particular size of access.
133 (expand_atomic_compare_and_swap): Likewise.
134 (expand_atomic_load): Likewise.
135 (expand_atomic_store): Likewise.
136 (expand_atomic_fetch_op): Likewise.
137 * testsuite/lib/target-supports.exp
138 (check_effective_target_sync_int_128): Remove x86 because it provides
140 (check_effective_target_sync_int_128_runtime): Likewise.
142 2017-02-01 Richard Biener <rguenther@suse.de>
144 * graphite.c: Include tree-vectorizer.h for find_loop_location.
145 (graphite_transform_loops): Provide opt-info for optimized nests.
146 * tree-parloop.c (parallelize_loops): Provide opt-info for
149 2017-02-01 Richard Biener <rguenther@suse.de>
152 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
155 2017-02-01 Richard Biener <rguenther@suse.de>
157 PR tree-optimization/71824
158 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
159 Verify the loops are valid in the merged SESE region.
160 (scop_detection::can_represent_loop_1): Check analyzing the
161 evolution of the number of iterations in the region succeeds.
163 2017-01-31 Ian Lance Taylor <iant@golang.org>
165 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
166 REG_ARGS_SIZE note to 32-bit push insns and call insn.
168 2017-01-31 David Malcolm <dmalcolm@redhat.com>
170 PR preprocessor/79210
171 * input.c (get_substring_ranges_for_loc): Replace line_width
172 assertion with error-handling.
174 2017-01-31 Richard Biener <rguenther@suse.de>
176 PR tree-optimization/77318
177 * graphite-sese-to-poly.c (extract_affine): Fix assert.
178 (create_pw_aff_from_tree): Take loop parameter.
179 (add_condition_to_pbb): Pass loop of the condition to
180 create_pw_aff_from_tree.
182 2017-01-31 Jakub Jelinek <jakub@redhat.com>
184 * config/s390/s390.c (s390_asan_shadow_offset): New function.
185 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
187 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
191 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
193 (convert_int_to_float128): Likewise.
194 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
195 (convert_int_to_float128): Likewise.
196 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
197 (UNSPEC_IEEE128_CONVERT): Likewise.
198 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
199 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
200 Use local variables for IBM extended format.
201 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
202 (fix_trunc<mode>si2_fprs): Likewise.
203 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
204 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
205 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
206 to know that we can now have integers of all sizes in vector
208 (fix<uns>_<mode>di2_hw): Likewise.
209 (float<uns>_<mode>si2_hw): Likewise.
210 (fix_<mode>si2_hw): Likewise.
211 (fixuns_<mode>si2_hw): Likewise.
212 (float<uns>_<mode>di2_hw): Likewise.
213 (float_<mode>di2_hw): Likewise.
214 (float_<mode>si2_hw): Likewise.
215 (floatuns_<mode>di2_hw): Likewise.
216 (floatuns_<mode>si2_hw): Likewise.
217 (xscvqp<su>wz_<mode>): Delete, no longer used.
218 (xscvqp<su>dz_<mode>): Likewise.
219 (xscv<su>dqp_<mode>): Likewise.
220 (ieee128_mfvsrd_64bit): Likewise.
221 (ieee128_mfvsrd_32bit): Likewise.
222 (ieee128_mfvsrwz): Likewise.
223 (ieee128_mtvsrw): Likewise.
224 (ieee128_mtvsrd_64bit): Likewise.
225 (ieee128_mtvsrd_32bit): Likewise.
227 2017-01-31 Martin Liska <mliska@suse.cz>
230 * ipa-prop.c (ipa_free_all_node_params): Call release method
231 instead of ~sumbol_summary to not to trigger double times
234 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
236 PR tree-optimization/71691
237 * bitmap.h (class auto_bitmap): New.
238 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
239 is_maybe_undefined instead of ssa_undefined_value_p.
241 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
243 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
244 __S390_ARCH_LEVEL__ to __ARCH__.
246 2017-01-31 Jakub Jelinek <jakub@redhat.com>
248 PR tree-optimization/79267
249 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
250 if should_remove_lhs_p is true.
252 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
255 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
256 (add_alignment_attribute): New.
257 (base_type_die): Add alignment attribute.
258 (subrange_type_die): Likewise.
259 (modified_type_die): Likewise.
260 (gen_array_type_die): Likewise.
261 (gen_descr_array_type_die: Likewise.
262 (gen_enumeration_type_die): Likewise.
263 (gen_subprogram_die): Likewise.
264 (gen_variable_die): Likewise.
265 (gen_field_die): Likewise.
266 (gen_ptr_to_mbr_type_die): Likewise.
267 (gen_struct_or_union_type_die): Likewise.
268 (gen_subroutine_type_die): Likewise.
269 (gen_typedef_die): Likewise.
270 (base_type_cmp): Compare alignment attribute.
272 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
275 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
276 (setb_unsigned) New pattern for setb with CCUNS.
277 * config/rs6000/rs6000.c (expand_block_compare): Use a different
278 subfc./subfe sequence to avoid overflow problems. Generate a
279 shorter sequence with cmpld/setb for power9.
280 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
281 for generating subfc. instruction.
282 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
283 now uses this instruction.
285 2017-01-30 Ian Lance Taylor <iant@google.com>
288 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
289 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
291 2017-01-30 Martin Sebor <msebor@redhat.com>
293 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
294 Move constant to the right of a relational operator.
295 (get_mpfr_format_length, format_character, format_string): Ditto.
296 (should_warn_p, maybe_warn): Same.
298 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
300 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
303 * asan.c (get_translation_unit_decl): Remove function.
304 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
306 2017-01-30 Martin Liska <mliska@suse.cz>
308 PR gcov-profile/79259
309 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
312 2017-01-30 Martin Liska <mliska@suse.cz>
315 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
316 Initialize variables with NULL value.
318 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
321 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
323 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
325 2017-01-30 Richard Biener <rguenther@suse.de>
327 PR tree-optimization/79276
328 * tree-vrp.c (process_assert_insertions): Properly adjust common
329 when removing a duplicate.
331 * gcc.dg/torture/pr79276.c: New testcase.
333 2017-01-30 Richard Biener <rguenther@suse.de>
335 PR tree-optimization/79256
336 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
337 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
339 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
341 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
344 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
345 ("*r<noxa>sbg_<mode>_sll_bitmask")
346 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
347 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
348 Use contiguous_bitmask_nowrap_operand.
350 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
353 * config/rs6000/altivec.h (vec_xl): Revise #define.
356 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
358 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
360 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
362 PR rtl-optimization/79194
363 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
364 traps before call to bypass_conditional_jumps.
366 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
368 PR tree-optimization/71374
369 * lra-constraints.c (check_conflict_input_operands): New.
370 (match_reload): Use it.
372 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
375 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
376 account to calculate conflict_set.
378 2017-01-27 Bin Cheng <bin.cheng@arm.com>
380 PR rtl-optimization/78559
381 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
382 other_insn in combine.
384 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
386 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
387 uint16_type_node for BT_UINT16.
389 2017-01-27 David Malcolm <dmalcolm@redhat.com>
391 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
393 (GIMPLE Tests): New node.
394 (RTL Tests): New node.
396 2017-01-27 Richard Biener <rguenther@suse.de>
398 PR tree-optimization/79245
399 * tree-loop-distribution.c (distribute_loop): Apply cost
400 modeling also to detected patterns.
402 2017-01-27 Richard Biener <rguenther@suse.de>
404 PR tree-optimization/71433
405 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
406 (compare_assert_loc): New function.
407 (process_assert_insertions): Sort and optimize assert locations
408 to remove duplicates and push down identical assertions on
409 edges to their destination block.
411 2017-01-27 Richard Biener <rguenther@suse.de>
413 PR tree-optimization/79244
414 * tree-vrp.c (remove_range_assertions): Forcefully propagate
415 out SSA names even if abnormal.
417 2017-01-27 Jakub Jelinek <jakub@redhat.com>
419 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
420 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
421 instead of MPFR_RNDN.
423 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
426 * arm.c (arm_option_override): Don't call build_target_option_node
427 until after doing all option overrides.
428 (arm_valid_target_attribute_tree): Likewise.
430 2017-01-27 Martin Liska <mliska@suse.cz>
432 * doc/invoke.texi (-fprofile-arcs): Document profiling support
433 for {cd}tors and C++ {cd}tors.
435 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
437 * config/s390/s390.md ("*setmem_long_and")
438 ("*setmem_long_and_31z"): Use zero_extend instead of and.
440 2017-01-26 Martin Sebor <msebor@redhat.com>
442 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
445 2017-01-26 Martin Sebor <msebor@redhat.com>
447 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
448 HAVE_DFmode before using XFmode or DFmode.
449 (parse_directive): Avoid using the z length modifier to avoid
450 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
453 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
454 to accept adjustment as an array.
455 (get_int_range): New function.
456 (struct directive): Make width and prec arrays.
457 (directive::set_width, directive::set_precision): Call get_int_range.
458 (format_integer, format_floating): Handle width and precision ranges.
459 (format_string, parse_directive): Same.
461 2017-01-26 Jakub Jelinek <jakub@redhat.com>
464 * dwarf2out.c (generate_skeleton_bottom_up): For children with
465 comdat_type_p set, just clone them, but keep the children in the
469 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
470 which have direct callers with -fvar-tracking-assignments enabled
472 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
473 inside of type units.
475 2017-01-26 Martin Sebor <msebor@redhat.com>
478 * gimple-ssa-sprintf.c (struct result_range): Add likely and
480 (struct format_result): Replace number_chars, number_chars_min,
481 and number_chars_max with a single member of struct result_range.
483 (format_result::operator+=): Adjust.
484 (struct fmtresult): Remove bounded. Handle likely and unlikely
486 (fmtresult::adjust_for_width_or_precision): New function.
487 (fmtresult:type_max_digits): New function.
488 (bytes_remaining): Handle likely and unlikely counters.
489 (min_bytes_remaining): Remove.
490 (format_percent): Simplify.
491 (format_integer, format_floating): Set likely and unlikely counters.
492 (get_string_length, format_character, format_string): Same.
493 (format_plain, should_warn_p): New function.
494 (maybe_warn): Call should_warn_p. Update diagnostic messages
495 and handle those for all directives, including plain strings.
496 (format_directive): Handle likely and unlikely counters.
497 Remove unnecessary quoting from diagnostics. Add an informational
500 (pass_sprintf_length::compute_format_length): Simplify.
501 (try_substitute_return_value): Handle likely and unlikely counters.
503 2017-01-26 Carl Love <cel@us.ibm.com>
505 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
506 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
508 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
511 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
512 endianess for subregs into account.
513 * lra-constraints.c (lra_constraints): Do risky transformations
514 always on the first iteration.
515 * lra-lives.c (check_pseudos_live_through_calls): Add arg
516 last_call_used_reg_set.
517 (process_bb_lives): Define and use last_call_used_reg_set.
518 * lra.c (lra): Always continue after lra_constraints on the first
521 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
523 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
525 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
527 2017-01-26 Jakub Jelinek <jakub@redhat.com>
529 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
530 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
531 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
532 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
533 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
534 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
535 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
536 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
537 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
539 2017-01-26 Marek Polacek <polacek@redhat.com>
542 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
543 for the third operand.
545 2017-01-26 Jakub Jelinek <jakub@redhat.com>
548 * omp-low.c (struct omp_context): Add simt_stmt field.
549 (scan_omp_for): Return omp_context *.
550 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
551 context to the _simt_ SIMD stmt.
552 (lower_omp_for): For combined SIMD with sibling _simt_
553 SIMD, make sure to use the same decls in _looptemp_
554 clauses as in the sibling.
556 2017-01-26 David Sherwood <david.sherwood@arm.com>
559 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
562 2017-01-26 Jakub Jelinek <jakub@redhat.com>
565 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
566 emit fld b; fld a; if possible.
568 * brig-builtins.def: Update copyright years.
569 * config/arm/arm_acle_builtins.def: Update copyright years.
571 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
574 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
575 constraint instead of o for the stxsd instruction.
577 2017-01-25 Carl Love <cel@us.ibm.com>
579 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
580 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
582 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
584 * doc/invoke.texi (C++ Dialect Options): Fix typo.
586 2017-01-25 Richard Biener <rguenther@suse.de>
588 PR tree-optimization/69264
589 * target.def (vector_alignment_reachable): Improve documentation.
590 * doc/tm.texi: Regenerate.
591 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
593 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
594 earlier changes with respect to TYPE_USER_ALIGN.
595 (vector_alignment_reachable_p): Likewise. Improve dumping.
597 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
600 * config/arm/arm.md (xordi3): Force constant operand into a register
603 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
605 * doc/invoke.texi (-fstore-merging): Correct default optimization
606 levels at which it is enabled.
607 (-O): Move -fstore-merging from list to...
610 2017-01-25 Richard Biener <rguenther@suse.de>
613 * omp-expand.c: Include debug.h.
614 (expand_omp_taskreg): Make sure to generate early debug before
615 outlining anything from a function.
616 (expand_omp_target): Likewise.
617 (grid_expand_target_grid_body): Likewise.
619 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
622 * asan.c (get_translation_unit_decl): New function.
623 (asan_add_global): Extract modules file name from globals
624 TRANSLATION_UNIT_DECL name.
626 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
629 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
630 for long calls with APCS frame and VFP.
632 2017-01-24 David Malcolm <dmalcolm@redhat.com>
634 * cfg.c (original_copy_tables_initialized_p): New function.
635 * cfg.h (original_copy_tables_initialized_p): New decl.
636 * cfgrtl.c (relink_block_chain): Guard the call to
637 free_original_copy_tables with a call to
638 original_copy_tables_initialized_p.
639 * cgraph.h (symtab_node::native_rtl_p): New decl.
640 * cgraphunit.c (symtab_node::native_rtl_p): New function.
641 (symtab_node::needed_p): Don't assert for early assembly output
643 (cgraph_node::finalize_function): Set "force_output" for __RTL
645 (cgraph_node::analyze): Bail out early for __RTL functions.
646 (analyze_functions): Update assertion to support __RTL functions.
647 (cgraph_node::expand): Bail out early for __RTL functions.
648 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
650 * function.h (struct function): Update comment for field
652 * gimple-expr.c: Include "tree-pass.h".
653 (gimple_has_body_p): Return false for __RTL functions.
654 * Makefile.in (OBJS): Add run-rtl-passes.o.
655 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
657 (gcc::pass_manager::get_clean_slate): New accessor.
658 * passes.c: Include "insn-addr.h".
659 (should_skip_pass_p): Add logging. Update logic for running
660 "expand" to be compatible with both __GIMPLE and __RTL. Guard
661 property-provider override so it is only done for gimple passes.
663 (skip_pass): New function.
664 (execute_one_pass): Call skip_pass when skipping passes.
665 * read-md.c (md_reader::read_char): Support filtering
666 the input to a subset of line numbers.
667 (md_reader::md_reader): Initialize fields
668 m_first_line and m_last_line.
669 (md_reader::read_file_fragment): New function.
670 * read-md.h (md_reader::read_file_fragment): New decl.
671 (md_reader::m_first_line): New field.
672 (md_reader::m_last_line): New field.
673 * read-rtl-function.c (function_reader::create_function): Only
674 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
675 curr_properties. Set DECL_INITIAL to a dummy block.
676 (read_rtl_function_body_from_file_range): New function.
677 * read-rtl-function.h (read_rtl_function_body_from_file_range):
679 * run-rtl-passes.c: New file.
680 * run-rtl-passes.h: New file.
682 2017-01-24 Jeff Law <law@redhat.com>
684 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
687 2017-01-24 Bin Cheng <bin.cheng@arm.com>
689 PR tree-optimization/79159
690 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
691 (record_nonwrapping_iv): Improve boundary using above function if no
692 value range information.
694 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
695 Martin Jambor <mjambor@suse.cz>
697 * brig-builtins.def: New file.
698 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
699 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
700 (DEF_HSAIL_SAT_BUILTIN): Likewise.
701 (DEF_HSAIL_INTR_BUILTIN): Likewise.
702 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
703 * builtin-types.def (BT_INT8): New.
704 (BT_INT16): Likewise.
705 (BT_UINT8): Likewise.
706 (BT_UINT16): Likewise.
707 (BT_FN_ULONG): Likewise.
708 (BT_FN_UINT_INT): Likewise.
709 (BT_FN_UINT_ULONG): Likewise.
710 (BT_FN_UINT_LONG): Likewise.
711 (BT_FN_UINT_PTR): Likewise.
712 (BT_FN_ULONG_PTR): Likewise.
713 (BT_FN_INT8_FLOAT): Likewise.
714 (BT_FN_INT16_FLOAT): Likewise.
715 (BT_FN_UINT32_FLOAT): Likewise.
716 (BT_FN_UINT16_FLOAT): Likewise.
717 (BT_FN_UINT8_FLOAT): Likewise.
718 (BT_FN_UINT64_FLOAT): Likewise.
719 (BT_FN_UINT16_UINT32): Likewise.
720 (BT_FN_UINT32_UINT16): Likewise.
721 (BT_FN_UINT16_UINT16_UINT16): Likewise.
722 (BT_FN_INT_PTR_INT): Likewise.
723 (BT_FN_UINT_PTR_UINT): Likewise.
724 (BT_FN_LONG_PTR_LONG): Likewise.
725 (BT_FN_ULONG_PTR_ULONG): Likewise.
726 (BT_FN_VOID_UINT64_UINT64): Likewise.
727 (BT_FN_UINT8_UINT8_UINT8): Likewise.
728 (BT_FN_INT8_INT8_INT8): Likewise.
729 (BT_FN_INT16_INT16_INT16): Likewise.
730 (BT_FN_INT_INT_INT): Likewise.
731 (BT_FN_UINT_FLOAT_UINT): Likewise.
732 (BT_FN_FLOAT_UINT_UINT): Likewise.
733 (BT_FN_ULONG_UINT_UINT): Likewise.
734 (BT_FN_ULONG_UINT_PTR): Likewise.
735 (BT_FN_ULONG_ULONG_ULONG): Likewise.
736 (BT_FN_UINT_UINT_UINT): Likewise.
737 (BT_FN_VOID_UINT_PTR): Likewise.
738 (BT_FN_UINT_UINT_PTR: Likewise.
739 (BT_FN_UINT32_UINT64_PTR): Likewise.
740 (BT_FN_INT_INT_UINT_UINT): Likewise.
741 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
742 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
743 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
744 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
745 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
746 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
747 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
748 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
749 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
750 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
751 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
752 * doc/frontends.texi: List BRIG FE.
753 * doc/install.texi (Testing): Add BRIG tesring requirements.
754 * doc/invoke.texi (Overall Options): Mention BRIG.
755 * doc/standards.texi (Standards): Doucment BRIG HSA version.
757 2017-01-24 Richard Biener <rguenther@suse.de>
760 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
762 2017-01-24 Martin Jambor <mjambor@suse.cz>
765 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
766 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
769 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
772 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
773 casts from signed to unsigned really don't have a range.
775 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
777 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
778 GMP_RNDx for compatiblity.
780 2017-01-24 Martin Liska <mliska@suse.cz>
783 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
784 that would prevent us to call alloca with -1 as argument.
786 2017-01-24 Jakub Jelinek <jakub@redhat.com>
788 * dwarf2out.c (output_compilation_unit_header, output_file_names):
789 Avoid -Wformat-security warning.
791 2017-01-23 Andrew Pinski <apinski@cavium.com>
793 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
796 2017-01-23 Martin Sebor <msebor@redhat.com>
799 * gimple-ssa-sprintf.c (warn_level): New global.
800 (format_integer): Use it here and throughout the rest of the file.
801 Use the same switch to compute sign as base.
802 (maybe_warn): New function.
803 (format_directive): Factor out warnings into maybe_warn.
804 Add debugging output. Use warn_level.
805 (add_bytes): Use warn_level.
806 (pass_sprintf_length::compute_format_length): Add debugging output.
807 (try_substitute_return_value): Same.
808 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
811 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
812 (struct fmtresult, format_integer, format_floating): Adjust.
813 (fmtresult::fmtresult): Set max correctly in two argument ctor.
814 (get_string_length, format_string,format_directive): Same.
815 (pass_sprintf_length::compute_format_length): Same.
816 (try_substitute_return_value): Simplify slightly.
819 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
820 (fmtresult::operator+=): Outlined.
821 (struct fmtresult): Add ctors.
822 (struct conversion_spec): Rename...
823 (struct directive): ...to this. Add and remove data members.
824 (directive::set_width, directive::set_precision): New functions.
825 (format_percent): Use fmtresult ctor.
826 (get_width_and_precision): Remove.
827 (format_integer): Make naming changes. Avoid computing width and
829 (format_floating): Same. Adjust indentation.
830 (format_character, format_none): New functions.
831 (format_string): Moved character handling to format_character.
832 (format_directive): Remove arguments, change return type.
833 (parse_directive): New function.
834 (pass_sprintf_length::compute_format_length): Move directive
835 parsing to parse_directive.
837 2017-01-23 Jakub Jelinek <jakub@redhat.com>
839 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
840 (assign_assembler_name_if_needed): ... this.
841 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
842 (assign_assembler_name_if_needed): ... this.
843 (free_lang_data_in_cgraph): Adjust callers.
844 * cgraphunit.c (cgraph_node::analyze): Likewise.
845 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
848 2017-01-23 Richard Biener <rguenther@suse.de>
850 PR tree-optimization/79088
851 PR tree-optimization/79188
852 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
853 resetting loop bounds after last path deletion. Reset loop
854 bounds of the target loop, make code match the comments.
855 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
856 Make sure loops need no fixups.
858 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
860 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
861 exponent support with double type for first argument.
862 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
863 type returned by __builtin_vec_extract_sig,
864 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
865 functions from "vector int" to "vector unsigned int" or from
866 "vector long long int" to "vector unsigned long long int".
867 Changed type returned by __builtin_vec_extract_exp,
868 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
869 functions from "vector int" to "vector unsigned int" or from
870 "vector long long int" to "vector unsigned long long int".
871 Changed return type of __builtin_vec_test_data_class,
872 __builtin_vec_test_data_class_sp, and
873 __builtin_vec_test_data_class_dp from "vector int" to
874 "vector bool int" or from "vector long long int" to "vector bool
875 long long int" and changed second argument type from "unsigned
876 int" to "int". Added new overloaded function forms "vector float
877 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
878 "vector float __builtin_vec_insert_exp_sp (vector float, vector
879 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
880 double, vector unsigned long long int)" and "vector double
881 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
882 long int)". Changed return type of
883 __builtin_scalar_test_data_class and
884 __builtin_scalar_test_data_class_sp and
885 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
886 int" and changed second argument from "unsigned int" to "int".
887 Changed type returned by __builtin_scalar_test_neg,
888 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
889 from "int" to "bool int". Added new overloaded function form
890 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
891 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
892 exponent double-precision with floating point first argument.
893 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
894 documentation of scalar_test_data_class, scalar_test_neg,
895 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
896 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
897 vec_test_data_class built-in functions to reflect refinements in
898 their type signatures.
900 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
902 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
904 (aarch64_elf_asm_destructor): Likewise.
906 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
908 PR rtl-optimization/78634
909 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
910 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
911 * ifcvt.c (noce_try_cmove): Add missing cost check.
913 PR rtl-optimization/71724
914 * combine.c (if_then_else_cond): Look for situations where it is
915 beneficial to undo the work of one of the recursive calls.
917 2017-01-23 Bin Cheng <bin.cheng@arm.com>
919 PR tree-optimization/70754
920 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
921 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
922 combined stmt before it if not NULL.
923 (combine_chains): Process refs reversely and compute dominance point
926 2017-01-23 Martin Liska <mliska@suse.cz>
928 PR tree-optimization/79196
929 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
930 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
932 (strlen_optimize_stmt): Call the renamed function.
934 2017-01-23 Michael Matz <matz@suse.de>
936 PR tree-optimization/78384
937 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
939 2017-01-23 Richard Biener <rguenther@suse.de>
941 PR tree-optimization/79186
942 * tree-vrp.c (register_new_assert_for): Make sure we've seen
943 both incoming edges before moving an assert.
945 2017-01-23 Martin Jambor <mjambor@suse.cz>
947 * ipa-prop.c (load_from_param_1): Removed.
948 (load_from_unmodified_param): Bits from load_from_param_1 put back
950 (load_from_param): Removed.
951 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
952 with stmt. Reverted back to use of load_from_unmodified_param.
954 2017-01-23 Martin Jambor <mjambor@suse.cz>
957 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
958 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
959 field a pointer to garbage collected vector, mark lattices and
960 ipcp_orig_node with GTY((skip)).
961 (ipa_get_param_count): Adjust to descriptors being a pointer.
962 (ipa_get_param): Likewise.
963 (ipa_get_type): Likewise.
964 (ipa_get_param_move_cost): Likewise.
965 (ipa_set_param_used): Likewise.
966 (ipa_get_controlled_uses): Likewise.
967 (ipa_set_controlled_uses): Likewise.
968 (ipa_is_param_used): Likewise.
969 (ipa_node_params_t): Move into garbage collector. New methods insert
971 (ipa_node_params_sum): Annotate wth GTY(()).
972 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
974 (ipa_load_from_parm_agg): Adjust declaration.
975 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
976 * ipa-profile.c (ipa_profile): Likewise.
977 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
978 (ipa_populate_param_decls): Make descriptors parameter garbage
980 (ipa_dump_param): Adjust to descriptors being a pointer.
981 (ipa_alloc_node_params): Likewise.
982 (ipa_initialize_node_params): Likewise.
983 (load_from_param_1): Make descriptors parameter garbage collected.
984 (load_from_unmodified_param): Likewise.
985 (load_from_param): Likewise.
986 (ipa_load_from_parm_agg): Likewise.
987 (ipa_node_params::~ipa_node_params): Removed.
988 (ipa_free_all_node_params): Remove call to delete operator.
989 (ipa_node_params_t::insert): New.
990 (ipa_node_params_t::remove): Likewise.
991 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
992 copy known_csts and known_contexts vectors.
993 (ipa_read_node_info): Adjust to descriptors being a pointer.
994 (ipcp_modif_dom_walker): Make m_descriptors field garbage
996 (ipcp_transform_function): Make descriptors variable garbage
999 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1001 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1002 * config/i386/avx512dqintrin.h: Ditto.
1003 * config/i386/avx512fintrin.h: Ditto.
1004 * gcc/config/i386/i386.c: Handle new builtins.
1005 * config/i386/i386-builtin.def: Add new builtins.
1006 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1007 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1009 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1010 Martin Liska <mliska@suse.cz>
1012 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1013 * asan.c (asan_expand_poison_ifn): Support stores and use
1014 appropriate ASAN report function.
1015 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1016 * internal-fn.def (ASAN_POISON_USE): Declare.
1017 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1018 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1019 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1020 ASAN_POISON calls w/o LHS.
1021 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1022 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1023 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1024 * gimplify.c (asan_poison_variables): Add attribute
1025 use_after_scope_memory to variables that really needs to live
1027 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1028 having the attribute.
1030 2017-01-23 Martin Liska <mliska@suse.cz>
1032 * asan.c (create_asan_shadow_var): New function.
1033 (asan_expand_poison_ifn): Likewise.
1034 * asan.h (asan_expand_poison_ifn): New declaration.
1035 * internal-fn.c (expand_ASAN_POISON): Likewise.
1036 * internal-fn.def (ASAN_POISON): New builtin.
1037 * sanopt.c (pass_sanopt::execute): Expand
1038 asan_expand_poison_ifn.
1039 * tree-inline.c (copy_decl_for_dup_finish): Make function
1041 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1042 * tree-ssa.c (is_asan_mark_p): New function.
1043 (execute_update_addresses_taken): Rewrite local variables
1044 (identified just by use-after-scope as addressable) into SSA.
1046 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1048 * doc/install.texi (Specific): opensource.apple.com uses https
1049 now. Remove trailing slash.
1051 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1053 * README.Portability: Remove note on an Irix compatibility issue.
1055 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1057 * gcov.c (INCLUDE_ALGORITHM): Define.
1058 (INCLUDE_VECTOR): Define.
1059 No longer include <vector> and <algorithm> directly.
1061 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1063 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1065 * doc/invoke.texi (Code Gen Options): Ditto.
1067 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1070 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1072 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1074 rtl-optimization/79125
1075 * cprop.c (local_cprop_pass): Handle cases where we make an
1078 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1082 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1083 read from, for big endian.
1085 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1087 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1088 register pauth builtins for LP64 only.
1090 2017-01-20 Marek Polacek <polacek@redhat.com>
1093 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1096 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1098 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1100 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1101 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1102 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1104 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1107 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1108 in big-endian mode when they are not a single duplicated value.
1110 2017-01-20 Richard Biener <rguenther@suse.de>
1112 * BASE-VER: Bump to 7.0.1.
1114 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1116 * omp-low.c (omplow_simd_context): New struct. Use it...
1117 (lower_rec_simd_input_clauses): ...here and...
1118 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1119 references to idx, lane, max_vf, is_simt.
1121 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1123 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1126 2017-01-20 Martin Jambor <mjambor@suse.cz>
1128 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1129 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1131 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1132 (GTFILES): Rename hsa.c to hsa-common.c.
1133 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1134 * hsa-dump.c: Likewise.
1135 * hsa-gen.c: Likewise.
1136 * hsa-regalloc.c: Likewise.
1137 * ipa-hsa.c: Likewise.
1138 * omp-expand.c: Likewise.
1139 * omp-low.c: Likewise.
1140 * toplev.c: Likewise.
1142 2017-01-20 Marek Polacek <polacek@redhat.com>
1145 * doc/invoke.texi: Document -Wduplicated-branches.
1146 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1147 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1148 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1149 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1150 return 0 only when not OEP_LEXICOGRAPHIC.
1151 (fold_build_cleanup_point_expr): Use the expression
1152 location when building CLEANUP_POINT_EXPR.
1153 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1154 * tree.c (add_expr): Handle error_mark_node.
1156 2017-01-20 Martin Liska <mliska@suse.cz>
1159 * tree-profile.c (init_ic_make_global_vars): Do not call
1161 (gimple_init_gcov_profiler): Likewise.
1163 2017-01-20 Martin Liska <mliska@suse.cz>
1166 * cgraph.h (maybe_create_reference): Remove argument and
1168 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1170 * ipa-cp.c (create_specialized_node): Likewise.
1171 * symtab.c (symtab_node::maybe_create_reference): Handle
1172 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1174 2017-01-20 Martin Liska <mliska@suse.cz>
1176 * read-rtl-function.c (function_reader::create_function): Use
1177 build_decl instread of build_decl_stat.
1179 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1181 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1182 * config/i386/avx512dqintrin.h: Ditto.
1183 * config/i386/avx512fintrin.h: Ditto.
1184 * config/i386/i386-builtin-types.def: Add new types.
1185 * gcc/config/i386/i386.c: Handle new types.
1186 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1187 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1188 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1189 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1190 (__builtin_ia32_kshiftridi): New.
1191 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1193 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1197 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1198 define to rs6000_init_stack_protect_guard.
1199 (rs6000_init_stack_protect_guard): New function.
1201 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1202 Yunqiang Su <yunqiang.su@imgtec.com>
1204 * config.gcc (supported_defaults): Add madd4.
1205 (with_madd4): Add validation.
1206 (all_defaults): Add madd4.
1207 * config/mips/mips.opt (mmadd4): New option.
1208 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1210 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1212 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1213 (ISA_HAS_FUSED_MADD4): Likewise.
1214 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1215 * gcc/doc/install.texi (--with-madd4): Document the new option.
1217 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1219 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1220 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1221 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1222 (aarch64_init_pauth_hint_builtins): New.
1223 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1224 (aarch64_expand_builtin): Expand new builtins.
1226 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1228 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1229 * combine-stack-adj.c (no_unhandled_cfa): Handle
1230 REG_CFA_TOGGLE_RA_MANGLE.
1231 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1232 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1233 info for return address signing.
1234 (aarch64_expand_epilogue): Likewise.
1236 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1238 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1239 * config/aarch64/aarch64-protos.h
1240 (aarch64_return_address_signing_enabled): New declaration.
1241 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1243 (aarch64_expand_prologue): Sign return address before it's pushed onto
1245 (aarch64_expand_epilogue): Authenticate return address fetched from
1247 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1248 (aarch64_attributes): New function attributes for "sign-return-address".
1249 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1250 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1251 ("*do_return"): Generate combined instructions according to key index.
1252 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1253 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1255 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1256 * config/aarch64/aarch64.opt (msign-return-address=): New.
1257 * doc/extend.texi (AArch64 Function Attributes): Documents
1258 "sign-return-address=".
1259 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1261 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1263 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1264 overall option summary.
1266 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1268 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1269 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1270 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1271 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1273 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1275 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1276 -mpower9-minmax by default for -mcpu=power9.
1277 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1278 128-bit floating point.
1280 2017-01-20 Alan Modra <amodra@gmail.com>
1282 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1283 optimizing for size.
1285 2017-01-20 Alan Modra <amodra@gmail.com>
1288 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1289 for strcmp and strncmp from corresponding builtin decl.
1291 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1293 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1294 instead of i386/rtems-64.h.
1295 * config/i386/rtems-64.h: Remove.
1297 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1301 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1303 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1305 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1307 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1308 Change int to HOST_WIDE_INT.
1309 * config/aarch64/aarch64-protos.h
1310 (aarch64_simd_gen_const_vector_dup): Likewise.
1311 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1313 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1315 * langhooks-def.h (lhd_type_for_size): New decl.
1316 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1317 * langhooks.c (lhd_type_for_size): New function, taken from
1320 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1322 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1323 define_bypass for CR latency.
1324 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1325 (power9-alu2): Add define_bypass for CR latency.
1327 (power9-mul): Update insn latency.
1328 (power9-mul-compare): Update insn latency, bypass latency and remove
1331 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1333 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1335 * config/aarch64/aarch64.md
1336 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1337 aarch64_nopcrelative_literal_loads.
1338 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1340 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1342 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1344 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1346 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1349 * config.gcc (supported_defaults): Add lxc1-sxc1.
1350 (with_lxc1_sxc1): Add validation.
1351 (all_defaults): Add lxc1-sxc1.
1352 * config/mips/mips.opt (mlxc1-sxc1): New option.
1353 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1355 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1356 __mips_no_lxc1_sxc1.
1357 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1358 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1359 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1361 2017-01-19 Richard Biener <rguenther@suse.de>
1363 PR tree-optimization/72488
1364 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1365 sure to restore SSA info.
1366 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1368 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1370 PR rtl-optimization/79121
1371 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1372 of the inner type when shifting an extended value.
1374 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1377 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1378 interposable aliases.
1380 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1383 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1384 Use the evmergelohi instruction.
1385 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1386 (mov_sitf_e500_subreg8_2_be): Likewise.
1387 (mov_sitf_e500_subreg12_2_le): Likewise.
1388 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1389 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1390 (mov_sitf_e500_subreg8_2_le): Likewise.
1391 (mov_sitf_e500_subreg12_2_be): Likewise.
1393 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1395 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1396 attribute from vecsimple to vecperm.
1397 (altivec_vbpermq2): Likewise.
1399 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1402 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1404 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1405 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1406 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1407 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1408 case where N arg is SIZE_MAX.
1409 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1410 (cmpstrsi): Add pattern.
1412 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1414 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1415 __builtin_vec_revb builtins.
1416 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1417 built-in functions to support generation of the ISA 3.0 XXBR<x>
1418 vector byte reverse instructions.
1419 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1420 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1421 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1422 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1423 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1424 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1425 (P9V_BUILTIN_VEC_REVB): Likewise.
1426 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1427 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1428 (p9_xxbrq_v16qi): Likewise.
1429 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1430 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1431 (p9_xxbrh_v8hi): Likewise.
1432 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1433 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1434 vec_revb built-in functions.
1436 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1438 PR rtl-optimization/78952
1439 * config/i386/i386.md (any_extract): New code iterator.
1440 (*insvqi_2): Use any_extract for source operand.
1441 (*insvqi_3): Use any_shiftrt for source operand.
1443 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1445 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1447 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1449 2017-01-18 Matthias Klose <doko@ubuntu.com>
1451 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1453 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1455 * config/rs6000/altivec.h (vec_bperm): Change #define.
1456 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1457 (altivec_vbpermq2): New define_insn.
1458 (altivec_vbpermd): Likewise.
1459 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1461 (VBPERMD): Likewise.
1462 (VBPERM): New polymorphic function interface.
1463 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1464 Add entries for P9V_BUILTIN_VEC_VBPERM.
1465 * doc/extend.texi: Add interfaces for vec_bperm.
1467 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1469 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1470 first letter of error messages.
1471 (s390_resolve_overloaded_builtin): Likewise.
1472 * config/s390/s390.c (s390_expand_builtin): Likewise.
1473 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1474 (s390_valid_target_attribute_inner_p): Likewise.
1475 * config/s390/s390.md ("tabort"): Likewise.
1477 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1479 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1480 (ISA_AVOID_DIV_HILO): New macro.
1481 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1482 (ISA_HAS_DDIV): Likewise.
1484 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1486 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1488 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1490 * doc/invoke.texi (fabi-version): Spelling fix.
1492 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1495 * doc/invoke.texi (fabi-version): Mention mangling fix for
1498 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1501 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1503 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1506 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1507 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1509 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1511 (stack_protect_test): Ditto.
1512 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1513 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1515 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1516 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1517 -mstack-protector-guard-offset=.
1518 (RS/6000 and PowerPC Options): Ditto.
1520 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1522 * config/i386/i386.h (MASK_CLASS_P): New define.
1523 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1524 there are no registers from different register sets also when
1525 mask registers are used. Update function comment.
1526 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1527 to (*k/*r) and (*k/*km) alternatives.
1529 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1531 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1532 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1533 (EH_RETURN_HANDLER_RTX): New define.
1534 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1535 Force frame pointer in EH return functions.
1536 (aarch64_expand_epilogue): Add barrier for eh_return.
1537 (aarch64_final_eh_return_addr): Remove.
1538 (aarch64_eh_return_handler_rtx): New function.
1539 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1541 (aarch64_eh_return_handler_rtx): New prototype.
1543 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1545 * config/rs6000/altivec.h (vec_rlmi): New #define.
1546 (vec_vrlnm): Likewise.
1547 (vec_rlnm): Likewise.
1548 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1549 (UNSPEC_VRLNM): Likewise.
1550 (VIlong): New mode iterator.
1551 (altivec_vrl<VI_char>mi): New define_insn.
1552 (altivec_vrl<VI_char>nm): Likewise.
1553 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1556 (RLNM): New polymorphic function entry.
1557 (VRLWMI): New monomorphic function entry.
1559 (RLMI): New polymorphic function entry.
1560 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1561 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1562 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1565 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1568 * dwarf2out.c (field_byte_offset): Restore the
1569 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1570 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1571 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1574 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1577 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1579 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1582 * dwarf2out.c (add_data_member_location_attribute): For constant
1583 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1584 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1585 DW_AT_byte_size attributes.
1587 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1589 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1590 after forcing to constant memory when the code model is medium.
1592 2017-01-17 Julia Koval <julia.koval@intel.com>
1595 * config/i386/avx512fintrin.h
1596 (_mm512_i32gather_ps): Change __addr type to void const*.
1597 (_mm512_mask_i32gather_ps): Ditto.
1598 (_mm512_i32gather_pd): Ditto.
1599 (_mm512_mask_i32gather_pd): Ditto.
1600 (_mm512_i64gather_ps): Ditto.
1601 (_mm512_mask_i64gather_ps): Ditto.
1602 (_mm512_i64gather_pd): Ditto.
1603 (_mm512_mask_i64gather_pd): Ditto.
1604 (_mm512_i32gather_epi32): Ditto.
1605 (_mm512_mask_i32gather_epi32): Ditto.
1606 (_mm512_i32gather_epi64): Ditto.
1607 (_mm512_mask_i32gather_epi64): Ditto.
1608 (_mm512_i64gather_epi32): Ditto.
1609 (_mm512_mask_i64gather_epi32): Ditto.
1610 (_mm512_i64gather_epi64): Ditto.
1611 (_mm512_mask_i64gather_epi64): Ditto.
1612 (_mm512_i32scatter_ps): Change __addr type to void*.
1613 (_mm512_mask_i32scatter_ps): Ditto.
1614 (_mm512_i32scatter_pd): Ditto.
1615 (_mm512_mask_i32scatter_pd): Ditto.
1616 (_mm512_i64scatter_ps): Ditto.
1617 (_mm512_mask_i64scatter_ps): Ditto.
1618 (_mm512_i64scatter_pd): Ditto.
1619 (_mm512_mask_i64scatter_pd): Ditto.
1620 (_mm512_i32scatter_epi32): Ditto.
1621 (_mm512_mask_i32scatter_epi32): Ditto.
1622 (_mm512_i32scatter_epi64): Ditto.
1623 (_mm512_mask_i32scatter_epi64): Ditto.
1624 (_mm512_i64scatter_epi32): Ditto.
1625 (_mm512_mask_i64scatter_epi32): Ditto.
1626 (_mm512_i64scatter_epi64): Ditto.
1627 (_mm512_mask_i64scatter_epi64): Ditto.
1628 * config/i386/avx512pfintrin.h
1629 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1630 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1631 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1632 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1633 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1634 (_mm512_prefetch_i32scatter_ps): Ditto.
1635 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1636 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1637 (_mm512_prefetch_i64scatter_pd): Ditto.
1638 (_mm512_prefetch_i64scatter_ps): Ditto.
1639 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1640 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1641 * config/i386/avx512vlintrin.h
1642 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1643 (_mm_mmask_i32gather_ps): Ditto.
1644 (_mm256_mmask_i32gather_pd): Ditto.
1645 (_mm_mmask_i32gather_pd): Ditto.
1646 (_mm256_mmask_i64gather_ps): Ditto.
1647 (_mm_mmask_i64gather_ps): Ditto.
1648 (_mm256_mmask_i64gather_pd): Ditto.
1649 (_mm_mmask_i64gather_pd): Ditto.
1650 (_mm256_mmask_i32gather_epi32): Ditto.
1651 (_mm_mmask_i32gather_epi32): Ditto.
1652 (_mm256_mmask_i32gather_epi64): Ditto.
1653 (_mm_mmask_i32gather_epi64): Ditto.
1654 (_mm256_mmask_i64gather_epi32): Ditto.
1655 (_mm_mmask_i64gather_epi32): Ditto.
1656 (_mm256_mmask_i64gather_epi64): Ditto.
1657 (_mm_mmask_i64gather_epi64): Ditto.
1658 (_mm256_i32scatter_ps): Change __addr type to void*.
1659 (_mm256_mask_i32scatter_ps): Ditto.
1660 (_mm_i32scatter_ps): Ditto.
1661 (_mm_mask_i32scatter_ps): Ditto.
1662 (_mm256_i32scatter_pd): Ditto.
1663 (_mm256_mask_i32scatter_pd): Ditto.
1664 (_mm_i32scatter_pd): Ditto.
1665 (_mm_mask_i32scatter_pd): Ditto.
1666 (_mm256_i64scatter_ps): Ditto.
1667 (_mm256_mask_i64scatter_ps): Ditto.
1668 (_mm_i64scatter_ps): Ditto.
1669 (_mm_mask_i64scatter_ps): Ditto.
1670 (_mm256_i64scatter_pd): Ditto.
1671 (_mm256_mask_i64scatter_pd): Ditto.
1672 (_mm_i64scatter_pd): Ditto.
1673 (_mm_mask_i64scatter_pd): Ditto.
1674 (_mm256_i32scatter_epi32): Ditto.
1675 (_mm256_mask_i32scatter_epi32): Ditto.
1676 (_mm_i32scatter_epi32): Ditto.
1677 (_mm_mask_i32scatter_epi32): Ditto.
1678 (_mm256_i32scatter_epi64): Ditto.
1679 (_mm256_mask_i32scatter_epi64): Ditto.
1680 (_mm_i32scatter_epi64): Ditto.
1681 (_mm_mask_i32scatter_epi64): Ditto.
1682 (_mm256_i64scatter_epi32): Ditto.
1683 (_mm256_mask_i64scatter_epi32): Ditto.
1684 (_mm_i64scatter_epi32): Ditto.
1685 (_mm_mask_i64scatter_epi32): Ditto.
1686 (_mm256_i64scatter_epi64): Ditto.
1687 (_mm256_mask_i64scatter_epi64): Ditto.
1688 (_mm_i64scatter_epi64): Ditto.
1689 (_mm_mask_i64scatter_epi64): Ditto.
1690 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1691 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1692 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1693 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1694 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1695 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1696 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1697 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1698 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1699 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1700 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1701 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1702 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1703 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1704 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1705 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1706 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1707 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1708 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1709 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1710 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1711 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1712 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1713 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1714 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1715 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1716 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1717 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1718 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1719 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1720 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1721 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1722 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1723 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1724 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1725 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1726 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1727 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1728 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1729 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1730 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1731 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1732 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1733 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1734 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1735 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1736 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1737 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1738 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1739 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1740 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1741 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1742 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1743 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1744 definitions accordingly.
1746 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1747 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1750 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1753 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1756 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1757 hard regs for inner regno.
1759 2017-01-17 Martin Liska <mliska@suse.cz>
1762 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1763 assumption and add comment.
1765 2017-01-17 Nathan Sidwell <nathan@acm.org>
1767 * ipa-visibility.c (localize_node): New function, broken out of ...
1768 (function_and_variable_visibility): ... here. Call it.
1770 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1773 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1774 correctly set frequency of oudgoing edge.
1775 (duplicate_thread_path): Fix profile updating.
1777 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1780 * configure.ac: Add GCC_BASE_VER.
1781 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1782 version from BASE-VER file.
1783 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1784 (gcc.o): Depend on $(BASEVER).
1785 * common.opt (dumpfullversion): New option.
1786 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1787 * doc/invoke.texi: Document -dumpfullversion.
1788 * doc/install.texi: Document --with-gcc-major-version-only.
1789 * configure: Regenerated.
1791 2017-01-17 Richard Biener <rguenther@suse.de>
1793 PR tree-optimization/71433
1794 * tree-vrp.c (register_new_assert_for): Merge same asserts
1795 on all incoming edges.
1796 (process_assert_insertions_for): Handle insertions at the
1799 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1801 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1802 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1804 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1807 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1810 2017-01-17 Alan Modra <amodra@gmail.com>
1813 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1814 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1815 symbolic stack limit when pic.
1817 2017-01-16 Martin Sebor <msebor@redhat.com>
1819 PR tree-optimization/78608
1820 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1822 2017-01-16 Jeff Law <law@redhat.com>
1825 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1826 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1827 for several include directories that may be relative to sysroot.
1828 * config/i386/x-mingw32 (gplus_includedir): Define.
1829 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1830 (native_system_includedir): Likewise.
1831 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1832 override if TARGET_SYSTEM_ROOT is defined.
1833 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1835 PR tree-optimization/79090
1836 PR tree-optimization/33562
1837 PR tree-optimization/61912
1838 PR tree-optimization/77485
1839 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1840 and computed trims into the dump file.
1842 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1844 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1846 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1849 * gimplify.c (gimplify_init_constructor): If want_value and
1850 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1854 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1855 sequence. Formatting fixes.
1856 (doloop_optimize): Formatting fixes.
1859 * gcc.c (debug_level_greater_than_spec_func): New function.
1860 (static_spec_functions): Add debug-level-gt spec function.
1861 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1863 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1864 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1865 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1866 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1867 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1868 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1870 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1872 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1873 QImode fixups to general and mask registers only.
1875 2017-01-16 Carl Love <cel@us.ibm.com>
1877 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1878 for built-in functions
1879 vector signed char vec_nabs (vector signed char)
1880 vector signed short vec_nabs (vector signed short)
1881 vector signed int vec_nabs (vector signed int)
1882 vector signed long long vec_nabs (vector signed long long)
1883 vector float vec_nabs (vector float)
1884 vector double vec_nabs (vector double)
1885 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1887 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1888 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1889 * doc/extend.texi: Update the documentation file for the new built-in
1892 2017-01-16 Martin Sebor <msebor@redhat.com>
1894 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1897 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1899 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1900 UNSPEC_VSX__XXSPLTD to require special splat handling.
1902 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1905 * system.h: Poison strndup.
1907 2017-01-16 Alan Modra <amodra@gmail.com>
1910 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1913 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1915 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1917 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1919 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1920 call recog here. Assert that INSN_CODE (insn) is non-negative.
1922 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1925 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1927 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1928 in the currently scheduled RTL fragment.
1930 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1932 PR rtl-optimization/78751
1933 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1936 2017-01-14 Jeff Law <law@redhat.com>
1938 PR tree-optimization/79090
1939 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1940 variable length stores.
1941 (compute_trims): Delete dead assignment to *trim_tail.
1942 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1945 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1947 PR rtl-optimization/78626
1948 PR rtl-optimization/78727
1949 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1950 of a block, and split such blocks after everything else is finished.
1952 2017-01-14 Alan Modra <amodra@gmail.com>
1955 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1956 target legitimate_combined_insn.
1957 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1958 (rs6000_legitimate_combined_insn): New function.
1959 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1961 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1962 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1963 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1965 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1967 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1969 2017-01-13 Jeff Law <law@redhat.com>
1971 PR tree-optimization/33562
1972 PR tree-optimization/61912
1973 PR tree-optimization/77485
1974 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1976 (delete_dead_assignment): Likewise.
1977 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1978 statement to delete_dead_call and delete_dead_assignment.
1980 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1983 * substring-locations.c (format_warning_va): Strengthen case 1 so
1984 that both endpoints of the substring must be within the format
1985 range for just the substring to be printed.
1987 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1989 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1990 * config/i386/i386.c (ix86_target_string): Add missing options
1991 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1992 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1993 flags_other and ix86_target_other to flags2_other. Display unknown
1995 (ix86_valid_target_attribute_inner_p): Add missing options and
1996 reorder options by implied ISAs, as in ix86_target_string.
1998 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2000 * hash-table.h (hash_table::too_empty_p): New function.
2001 (hash_table::expand): Use it.
2002 (hash_table::traverse): Likewise.
2003 (hash_table::empty_slot): Use sizeof (value_type) instead of
2004 sizeof (PTR) to convert bytes to elements. Shrink the table
2005 if the current size is excessive for the current number of
2008 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2010 * ira-costs.c (record_reg_classes): Break from the inner loop
2011 early once alt_fail is known to be true. Update outer loop
2012 handling accordingly.
2014 2017-01-13 Jeff Law <law@redhat.com>
2016 * tree-ssa-dse.c (decrement_count): New function.
2017 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2018 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2019 when we know the partially dead statement is a mem* function.
2021 PR tree-optimization/61912
2022 PR tree-optimization/77485
2023 * tree-ssa-dse.c: Include expr.h.
2024 (maybe_trim_constructor_store): New function.
2025 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2027 PR tree-optimization/33562
2028 PR tree-optimization/61912
2029 PR tree-optimization/77485
2030 * doc/invoke.texi: Document new dse-max-object-size param.
2031 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2032 * tree-ssa-dse.c: Include params.h.
2033 (dse_store_status): New enum.
2034 (initialize_ao_ref_for_dse): New, partially extracted from
2036 (valid_ao_ref_for_dse, normalize_ref): New.
2037 (setup_live_bytes_from_ref, compute_trims): Likewise.
2038 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2039 (maybe_trim_partially_dead_store): Likewise.
2040 (maybe_trim_complex_store): Likewise.
2041 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2042 Track what bytes live from the original store. Return tri-state
2043 for dead, partially dead or live.
2044 (dse_dom_walker): Add constructor, destructor and new private members.
2045 (delete_dead_call, delete_dead_assignment): New extracted from
2047 (dse_optimize_stmt): Make a member of dse_dom_walker.
2048 Use initialize_ao_ref_for_dse.
2050 PR tree-optimization/33562
2051 PR tree-optimization/61912
2052 PR tree-optimization/77485
2053 * sbitmap.h (bitmap_count_bits): Prototype.
2054 (bitmap_clear_range, bitmap_set_range): Likewise.
2055 * sbitmap.c (bitmap_clear_range): New function.
2056 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2058 2017-01-13 Martin Liska <mliska@suse.cz>
2061 * function.c (set_cfun): Add new argument force.
2062 * function.h (set_cfun): Likewise.
2063 * ipa-inline-transform.c (inline_call): Use the function when
2064 strict alising from is dropped for function we inline to.
2066 2017-01-13 Richard Biener <rguenther@suse.de>
2068 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2069 for dumping GIMPLE INTEGER_CSTs.
2071 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2073 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2074 to 201112L since C++17.
2076 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2079 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2080 if -fsanitize=kernel-address is present.
2082 2017-01-13 Richard Biener <rguenther@suse.de>
2084 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2085 as _Literal ( type ) number in case usual suffixes do not
2086 preserve all information.
2088 2017-01-13 Richard Biener <rguenther@suse.de>
2090 PR tree-optimization/77283
2091 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2092 and ssa-iterators.h.
2093 (is_feasible_trace): Implement a cost model based on joiner
2096 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2099 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2100 char or short to __float128/_Float128 directly.
2102 2017-01-12 Martin Sebor <msebor@redhat.com>
2104 to -Wformat-overflow.
2105 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2106 (min_bytes_remaining): Same.
2107 (get_string_length): Same.
2108 (format_string): Same.
2109 (format_directive): Same.
2111 (pass_sprintf_length::handle_gimple_call): Same.
2113 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2115 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2116 info.nowrite calls with no lhs that can't throw. Return bool
2117 whether gsi_remove has been called or not.
2118 (pass_sprintf_length::handle_gimple_call): Return bool whether
2119 try_substitute_return_value called gsi_remove. Formatting fix.
2120 (pass_sprintf_length::execute): Don't use gsi_remove if
2121 handle_gimple_call returned true.
2124 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2125 be removed due to side-effects, don't remove following barrier nor
2126 turn the successor edge into fallthru edge.
2128 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2131 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2132 element-reversing loads and stores as not swappable.
2134 2017-01-12 Nathan Sidwell <nathan@acm.org>
2135 Nicolai Stange <nicstange@gmail.com>
2137 * combine.c (try_combine): Don't ignore result of overlap checking
2138 loop. Combine overlap & asm check into single loop.
2140 2017-01-12 Richard Biener <rguenther@suse.de>
2142 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2143 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2145 2017-01-12 Richard Biener <rguenther@suse.de>
2147 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2148 and TS_TARGET_OPTION directly derive from TS_BASE.
2149 * tree-core.h (tree_optimization_option): Derive from tree_base.
2150 (tree_target_option): Likewise.
2152 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2154 * config/i386/i386.c (memory_address_length): Increase len
2155 only when rip_relative_addr_p returns false.
2157 2017-01-11 Julia Koval <julia.koval@intel.com>
2159 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2160 (OPTION_MASK_ISA_SGX_SET): New.
2161 (ix86_handle_option): Handle OPT_msgx.
2162 * config.gcc: Added sgxintrin.h.
2163 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2164 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2165 * config/i386/i386.c (ix86_target_string): Add -msgx.
2167 (ix86_option_override_internal): Handle new options.
2168 (ix86_valid_target_attribute_inner_p): Add sgx.
2169 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2170 * config/i386/i386.opt: Add msgx.
2171 * config/i386/sgxintrin.h: New file.
2172 * config/i386/x86intrin.h: Add sgxintrin.h.
2174 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2177 * fold-const.c (maybe_nonzero_address): Return 1 for function
2179 (tree_single_nonzero_warnv_p): Don't handle function local objects
2183 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2186 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2189 * opts.c: Include "spellcheck.h"
2190 (struct string_fragment): New struct.
2191 (struct edit_distance_traits<const string_fragment &>): New
2193 (get_closest_sanitizer_option): New function.
2194 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2196 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2198 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2200 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2201 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2202 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2203 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2204 for initial die_offset if dwarf_split_debug_info.
2205 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2206 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2208 (output_skeleton_debug_sections): Formatting fix. Use
2209 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2210 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2212 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2214 * config/arm/cortex-a53.md: Add bypasses for
2216 (cortex_a53_r2f): Only use for transfers.
2217 (cortex_a53_f2r): Likewise.
2218 (cortex_a53_r2f_cvt): Add reservation for conversions.
2219 (cortex_a53_f2r_cvt): Likewise.
2221 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2223 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2224 to all inlined functions, change static to extern.
2226 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2229 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2231 (arm_assemble_integer): Likewise.
2233 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2235 * config.gcc: Use new awk script to check CPU, FPU and architecture
2236 parameters for --with-... options.
2237 * config/arm/parsecpu.awk: New file
2238 * config/arm/arm-cpus.in: New file.
2239 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2241 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2243 * config/arm/t-arm: Update dependency rules.
2244 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2245 of processing .def files.
2246 * config/arm/genopt.sh: Deleted.
2247 * config/arm/gentune.sh: Deleted.
2248 * config/arm/arm-cores.def: Deleted.
2249 * config/arm/arm-arches.def: Deleted.
2250 * config/arm/arm-fpus.def: Deleted.
2251 * config/arm/arm-tune.md: Regenerated.
2252 * config/arm/arm-tables.opt: Regenerated.
2253 * config/arm/arm-cpu.h: New generated file.
2254 * config/arm/arm-cpu-data.h: New generated file.
2255 * config/arm/arm-cpu-cdata.h: New generated file.
2257 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2260 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2262 (input_varpool_node): Unpack dynamically_initialized bit.
2264 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2266 PR rtl-optimization/79032
2267 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2268 the alignment of the adjusted memory reference against that of MODE,
2269 instead of the alignment of the original memory reference.
2271 2017-01-11 Martin Jambor <mjambor@suse.cz>
2273 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2275 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2276 decorated functions.
2278 2017-01-11 Richard Biener <rguenther@suse.de>
2280 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2281 set range/nonnull info for PHI results. Do not set it on
2282 stmts marked for removal.
2284 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2286 * expr.c (store_field): In the bitfield case, fetch the return value
2287 from the registers before applying a single big-endian adjustment.
2288 Always do a final load for a BLKmode value not larger than a word.
2290 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2293 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2294 that we correctly handle column numbers greater than
2295 LINE_MAP_MAX_COLUMN_NUMBER.
2297 2017-01-10 Martin Sebor <msebor@redhat.com>
2300 * gimple-ssa-sprintf.c (get_destination_size): Call
2301 {init,fini}object_sizes.
2302 * tree-object-size.c (addr_object_size): Adjust.
2303 (pass_through_call): Adjust.
2304 (pass_object_sizes::execute): Adjust.
2305 * tree-object-size.h (fini_object_sizes): Declare.
2307 2017-01-10 Martin Sebor <msebor@redhat.com>
2309 PR tree-optimization/78775
2310 * builtins.c (get_size_range): Move...
2311 * calls.c: ...to here.
2312 (alloc_max_size): Accept zero argument.
2313 (operand_signed_p): Remove.
2314 (maybe_warn_alloc_args_overflow): Call get_size_range.
2315 * calls.h (get_size_range): Declare.
2317 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2319 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2320 from TI's devices.csv file as of September 2016.
2321 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2323 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2325 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2326 * doc/invoke.texi: Likewise.
2327 * doc/md.texi: Likewise.
2328 * doc/objc.texi: Likewise.
2330 2017-01-10 Joshua Conner <joshconner@google.com>
2332 * config/arm/fuchsia-elf.h: New file.
2333 * config/fuchsia.h: New file.
2334 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2335 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2337 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2339 2016-01-10 Richard Biener <rguenther@suse.de>
2341 PR tree-optimization/79034
2342 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2343 Propagate out degenerate PHIs in the joiner.
2345 2017-01-10 Martin Liska <mliska@suse.cz>
2347 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2348 (sort_congruence_classes_by_decl_uid): Likewise.
2349 (sort_congruence_class_groups_by_decl_uid): Likewise.
2350 (sem_item_optimizer::merge_classes): Sort class, groups in these
2351 classes and members in the groups by DECL_UID of declarations.
2352 This would make merge operations stable.
2354 2017-01-10 Martin Liska <mliska@suse.cz>
2356 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2357 usage of m_classes_vec.
2358 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2359 (sem_item_optimizer::get_group_by_hash): Likewise.
2360 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2361 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2362 (sem_item_optimizer::verify_classes): Likewise.
2363 (sem_item_optimizer::process_cong_reduction): Likewise.
2364 (sem_item_optimizer::dump_cong_classes): Likewise.
2365 (sem_item_optimizer::merge_classes): Likewise.
2366 * ipa-icf.h (congruence_class_hash): Rename from
2367 congruence_class_group_hash. Remove declaration of m_classes_vec.
2369 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2371 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2372 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2373 * config.gcc: Add avx512vpopcntdqintrin.h.
2374 * config/i386/avx512vpopcntdqintrin.h: New.
2375 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2376 * config/i386/i386-builtin-types.def: Add new types.
2377 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2378 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2379 __builtin_ia32_vpopcountq_v8di_mask): New.
2380 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2381 __AVX512VPOPCNTDQ__.
2382 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2383 (PTA_AVX512VPOPCNTDQ): Define.
2384 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2385 TARGET_AVX512VPOPCNTDQ_P): Define.
2386 * config/i386/i386.opt: Add mavx512vpopcntdq.
2387 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2388 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2390 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2393 * predict.def (PRED_CALL): Set to 67.
2395 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2397 * expr.c (store_field): In the bitfield case, if the value comes from
2398 a function call and is of an aggregate type returned in registers, do
2399 not modify the field mode; extract the value in all cases if the mode
2400 is BLKmode and the size is not larger than a word.
2402 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2405 * config/i386/cpuid.h: Fix undefined behavior.
2407 2017-01-04 Jeff Law <law@redhat.com>
2409 PR tree-optimization/79007
2410 PR tree-optimization/67955
2411 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2412 conservative for pt.null when flag_non_call_exceptions is on.
2414 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2416 PR translation/79019
2417 PR translation/79020
2418 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2419 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2420 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2422 * config/avr/avr.opt (maccumulate-args): Likewise.
2423 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2424 * common.opt (freport-bug): Likewise.
2425 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2426 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2427 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2428 translatable string.
2429 * config/i386/i386.c (function_value_32): Likewise.
2430 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2431 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2433 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2434 * common/config/msp430/msp430-common.c (msp430_handle_option):
2436 * symtab.c (symtab_node::verify_base): Likewise.
2437 * opts.c (set_debug_level): Likewise.
2438 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2439 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2440 missing whitespace to translatable strings.
2441 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2442 * config/sh/superh.h: Likewise.
2443 * config/i386/xopintrin.h: Likewise.
2444 * config/i386/znver1.md: Likewise.
2445 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2446 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2447 * double-int.h (struct double_int): Likewise.
2448 * double-int.c (div_and_round_double): Likewise.
2449 * wide-int.cc: Likewise.
2450 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2451 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2452 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2453 (crossjumps_occurred): ... this.
2454 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2457 PR tree-optimization/78899
2458 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2459 returning bool return struct loop *, NULL for failure and the new
2461 (versionable_outer_loop_p): Don't version outer loop if it has
2462 dont_vectorized bit set.
2463 (tree_if_conversion): When versioning outer loop, ensure
2464 tree_if_conversion is performed also on the inner loop of the
2465 non-vectorizable outer loop copy.
2466 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2467 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2468 prevent vectorization of it.
2469 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2470 the outer loop vectorization of the non-scalar version is attempted
2471 before vectorization of the inner loop in scalar version. If
2472 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2473 vectorization of its inner loop.
2474 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2475 has 2 inner loops, rename also on edges from bb whose single pred
2476 is outer_loop->header. Fix typo in function comment.
2478 2017-01-09 Martin Sebor <msebor@redhat.com>
2481 * asan.c (asan_emit_stack_protection): Increase local buffer size
2482 to avoid snprintf truncation warning.
2484 2017-01-09 Andrew Pinski <apinski@cavium.com>
2486 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2487 to reference thunderx2t99 for the tuning structure
2488 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2490 (thunderx2t99_extra_costs): This.
2491 * config/aarch64/aarch64-tune.md: Regenerate.
2492 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2493 (vulcan_addrcost_table): This.
2494 (vulcan_regmove_cost): Rename to ...
2495 (thunderx2t99_regmove_cost): This.
2496 (vulcan_vector_cost): Rename to ...
2497 (thunderx2t99_vector_cost): this.
2498 (vulcan_branch_cost): Rename to ...
2499 (thunderx2t99_branch_cost): This.
2500 (vulcan_tunings): Rename to ...
2501 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2502 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2504 2017-01-09 Martin Jambor <mjambor@suse.cz>
2508 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2509 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2510 (propagate_vr_accross_jump_function): Use the above function for all
2511 value range computations for pass-through jump functions and type
2512 converasion from explicit value range values.
2513 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2514 parameters from TYPE_ARG_TYPES.
2515 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2516 (ipa_write_node_info): Stream type of the actual argument.
2517 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2519 2017-01-09 Martin Liska <mliska@suse.cz>
2522 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2523 (lookup_compiler): Do not show error message with have_E.
2525 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2527 PR tree-optimization/78938
2528 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2529 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2530 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2531 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2534 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2536 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2539 2017-01-09 Richard Biener <rguenther@suse.de>
2541 PR tree-optimization/78997
2542 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2543 name condition properly.
2545 2017-01-09 Richard Biener <rguenther@suse.de>
2548 * dwarf2out.c (is_cxx): New overload with context.
2549 (is_naming_typedef_decl): Use it.
2551 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2553 * invoke.texi (Option Summary): Correct spacing in option lists
2554 and add line breaks to fix over-long lines.
2556 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2560 * extend.texi (Common Variable Attributes): Add xref to GCC
2561 Internals manual to explain mode attribute keywords.
2563 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2566 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2567 and Preprocessor Options.
2568 (Options for Linking): Document -pthread here....
2569 (RS/6000 and PowerPC Options): ...not here.
2570 (Solaris 2 Options): ...or here.
2571 * doc/cppopts.texi: Document -pthread.
2573 2017-01-08 Martin Sebor <msebor@redhat.com>
2576 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2577 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2578 New member functions.
2579 (format_directive): Used them.
2581 (pass_sprintf_length::handle_gimple_call): Same.
2582 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2583 to avoid truncation for any argument.
2584 (extract_affine_mul): Same.
2585 * tree.c (get_file_function_name): Same.
2587 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2590 * predict.def (PRED_INDIR_CALL): Set to 86.
2592 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2594 PR preprocessor/54124
2595 * doc/cppopts.texi: Reformat -d subtable to list the full name
2596 of the options. Add cross-reference to the docs for the general
2597 compiler -d options.
2598 * doc/invoke.texi (Developer Options): Add cross-reference to the
2599 preprocessor-specific -d option documentation.
2601 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2603 PR preprocessor/13498
2604 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2605 redudant material, and reflect new command-line options.
2606 (System Headers): Likewise.
2608 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2610 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2611 -isystem, and -idirafter. Copy-edit.
2612 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2613 default for -ftrack-macro-expansion. Delete obsolete and
2614 badly-formatted implementation details about -fdebug-cpp output.
2615 * doc/cppwarnopts.texi: Copy-edit.
2617 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2620 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2621 that the transition from a max line width >= 1<<10 to narrower
2622 lines works correctly.
2624 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2626 * doc/options.texi (PerFunction): New.
2627 * opt-functions.awk (switch_flags): Map both Optimization and
2628 PerFunction to CL_OPTIMIZATION.
2629 * opth-gen.awk: Test for PerFunction flag along with
2631 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2632 it only when the latter is present. Skip those that don't in
2633 the hash function generator.
2634 * common.opt (fvar-tracking): Mark as PerFunction instead of
2636 (fvar-tracking-assignments): Likewise.
2637 (fvar-tracking-assignments-toggle): Likewise.
2638 (fvar-tracking-uninit): Likewise.
2640 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2642 PR translation/79018
2643 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2646 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2649 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2650 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2651 TARGET_LONG_JUMP_TABLE_OFFSETS.
2652 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2653 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2654 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2655 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2656 * config/m68k/m68k.md (tablejump expander): Likewise.
2657 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2658 TARGET_LONG_JUMP_TABLE_OFFSETS.
2659 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2660 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2662 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2663 David Holsgrove <david.holsgrove@xilinx.com>
2665 * common/config/microblaze/microblaze-common.c
2666 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2667 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2669 * config/microblaze/microblaze.c (microblaze_must_save_register)
2670 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2672 (microblaze_eh_return): New function.
2673 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2674 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2675 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2676 * config/microblaze/microblaze.md (eh_return): New pattern.
2678 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2680 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2681 GCC_DIAGNOSTIC_STRINGIFY): Define.
2683 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2685 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2687 * config/arm/arm.md (<mcrr>): New.
2689 * config/arm/arm.c (arm_arch5te): New.
2690 (arm_option_override): Set arm_arch5te.
2691 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2693 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2694 (arm_mcrr_qualifiers): ... this. New.
2695 (MRRC_QUALIFIERS): Define to...
2696 (arm_mrrc_qualifiers): ... this. New.
2697 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2699 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2700 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2701 (MRRCI, mrrc, MRRC): New.
2702 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2703 VUNSPEC_MRRC2): New.
2705 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2707 * config/arm/arm.md (<mcr>): New.
2709 * config/arm/arm.c (arm_coproc_builtin_available): Add
2710 support for mcr, mrc, mcr2 and mrc2.
2711 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2712 (arm_mcr_qualifiers): ... this. New.
2713 (MRC_QUALIFIERS): Define to ...
2714 (arm_mrc_qualifiers): ... this. New.
2715 (MCR_QUALIFIERS): Define to ...
2716 (arm_mcr_qualifiers): ... this. New.
2717 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2719 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2720 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2721 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2724 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2726 * config/arm/arm.md (*ldc): New.
2730 * config/arm/arm.c (arm_coproc_builtin_available): Add
2731 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2732 (arm_coproc_ldc_stc_legitimate_address): New.
2733 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2734 'qualifier_const_pointer'.
2735 (LDC_QUALIFIERS): Define to...
2736 (arm_ldc_qualifiers): ... this. New.
2737 (STC_QUALIFIERS): Define to...
2738 (arm_stc_qualifiers): ... this. New.
2739 * config/arm/arm-protos.h
2740 (arm_coproc_ldc_stc_legitimate_address): New.
2741 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2742 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2743 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2744 stc2, stcl, stc2l): New.
2745 * config/arm/constraints.md (Uz): New.
2746 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2747 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2748 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2749 VUNSPEC_STC2L): New.
2751 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2753 * config/arm/arm.md (<cdp>): New.
2754 * config/arm/arm.c (neon_const_bounds): Rename this ...
2755 (arm_const_bounds): ... this.
2756 (arm_coproc_builtin_available): New.
2757 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2758 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2759 (CDP_QUALIFIERS): Define to...
2760 (arm_cdp_qualifiers): ... this. New.
2762 (arm_expand_builtin_args): Add case for 6 arguments.
2763 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2764 (arm_const_bounds): ... this.
2765 (arm_coproc_builtin_available): New.
2766 * config/arm/arm_acle.h (__arm_cdp): New.
2768 * config/arm/arm_acle_builtins.def (cdp): New.
2770 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2771 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2773 * config/arm/types.md (coproc): New.
2774 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2775 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2776 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2777 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2779 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2781 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2782 (UBINOP_QUALIFIERS): New.
2784 (acle_builtin_data): New. Change comment.
2785 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2786 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2787 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2788 arm_acle_builtins.def.
2789 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2790 (arm_init_acle_builtins): New.
2791 (CRC32_BUILTIN): Remove.
2792 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2793 crc32cb, crc32ch and crc32cw.
2794 (arm_init_crc32_builtins): Remove.
2795 (arm_init_builtins): Use arm_init_acle_builtins rather
2796 than arm_init_crc32_builtins.
2797 (arm_expand_acle_builtin): New.
2798 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2799 * config/arm/arm_acle_builtins.def: New.
2801 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2803 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2804 (arm_builtin_datum): ... this.
2805 (arm_init_neon_builtin): Rename to ...
2806 (arm_init_builtin): ... this. Add a new parameters PREFIX
2807 and USE_SIG_IN_NAME.
2808 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2809 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2810 'arm_builtin_datum'.
2811 (arm_init_vfp_builtins): Likewise.
2812 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2813 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2814 (arm_expand_neon_args): Rename to ...
2815 (arm_expand_builtin_args): ... this. Rename builtin_arg
2816 enum values and differentiate between ARG_BUILTIN_MEMORY
2817 and ARG_BUILTIN_NEON_MEMORY.
2818 (arm_expand_neon_builtin_1): Rename to ...
2819 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2820 values, arm_expand_builtin_args and add bool parameter NEON.
2821 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2822 (arm_expand_vfp_builtin): Likewise.
2823 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2825 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2828 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2829 * predict.c (tree_estimate_probability_bb): Reverse direction of
2830 polymorphic call predictor.
2832 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2834 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2835 (determine_pass_name_match): ...this new function and...
2836 (should_skip_pass_p): ...this new function.
2838 2017-01-06 Nathan Sidwell <nathan@acm.org>
2840 * ipa-visibility.c (function_and_variable_visibility): Reformat
2841 comments and long lines. Remove extrneous if.
2842 * symtab.c (symtab_node::make_decl_local): Fix code format.
2843 (symtab_node::set_section_for_node): Fix comment typo.
2845 2017-01-06 Martin Liska <mliska@suse.cz>
2848 * lra-constraints.c: Rename invariant to lra_invariant.
2849 * predict.c (set_even_probabilities): Initialize e to NULL.
2851 2017-01-05 Martin Sebor <msebor@redhat.com>
2853 PR tree-optimization/78910
2854 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2855 (format_integer): Correct off-by-one error in the handling
2856 of precision with negative numbers in signed conversions..
2858 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2860 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2862 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2864 PR tree-optimization/71016
2865 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2866 factor_out_conditional_conversion. Formatting fix.
2867 (factor_out_conditional_conversion): Add cond_stmt argument.
2868 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2869 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2872 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2874 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2875 read-rtl-function.o, and selftest-rtl.o.
2876 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2877 (selftest::aarch64_test_loading_full_dump): New function.
2878 (selftest::aarch64_run_selftests): New function.
2879 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2880 selftest::aarch64_run_selftests.
2881 * config/i386/i386.c
2882 (selftest::ix86_test_loading_dump_fragment_1): New function.
2883 (selftest::ix86_test_loading_call_insn): New function.
2884 (selftest::ix86_test_loading_full_dump): New function.
2885 (selftest::ix86_test_loading_unspec): New function.
2886 (selftest::ix86_run_selftests): Call the new functions.
2887 * emit-rtl.c (maybe_set_max_label_num): New function.
2888 * emit-rtl.h (maybe_set_max_label_num): New decl.
2889 * function.c (instantiate_decls): Guard call to
2890 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2891 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2893 * gensupport.c (gen_reader::gen_reader): Pass "false"
2894 for new "compact" param of rtx_reader.
2895 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2896 rather than an empty string for NULL strings.
2897 * read-md.c: Potentially include config.h rather than bconfig.h.
2898 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2899 (have_error): New global, copied from errors.c.
2900 (md_reader::read_name): Rename to...
2901 (md_reader::read_name_1): ...this, adding "out_loc" param,
2902 and converting "missing name or number" to returning false, rather
2904 (md_reader::read_name): Reimplement in terms of read_name_1.
2905 (md_reader::read_name_or_nil): New function.
2906 (md_reader::read_string): Handle "(nil)" by returning NULL.
2907 (md_reader::md_reader): Add new param "compact".
2908 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2909 (md_reader::read_file): New method.
2910 * read-md.h (md_reader::md_reader): Add new param "compact".
2911 (md_reader::read_file): New method.
2912 (md_reader::is_compact): New accessor.
2913 (md_reader::read_name): Convert return type from void to file_location.
2914 (md_reader::read_name_or_nil): New decl.
2915 (md_reader::read_name_1): New decl.
2916 (md_reader::m_compact): New field.
2917 (noop_reader::noop_reader): Pass "false" for new "compact" param
2919 (rtx_reader::rtx_reader): Add new "compact" param.
2920 (rtx_reader::read_rtx_operand): Make virtual and convert return
2921 type from void to rtx.
2922 (rtx_reader::read_until): New decl.
2923 (rtx_reader::handle_any_trailing_information): New virtual function.
2924 (rtx_reader::postprocess): New virtual function.
2925 (rtx_reader::finalize_string): New virtual function.
2926 (rtx_reader::m_in_call_function_usage): New field.
2927 (rtx_reader::m_reuse_rtx_by_id): New field.
2928 * read-rtl-function.c: New file.
2929 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2930 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2931 (selftest::verify_three_block_rtl_cfg): New decl.
2932 * read-rtl-function.h: New file.
2933 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2934 For host, include function.h, memmodel.h, and emit-rtl.h.
2935 (one_time_initialization): New function.
2936 (struct compact_insn_name): New struct.
2937 (compact_insn_names): New array.
2938 (find_code): Handle insn codes in compact dumps.
2939 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2940 (bind_subst_iter_and_attr): Likewise.
2941 (add_condition_to_string): Likewise.
2942 (add_condition_to_rtx): Likewise.
2943 (apply_attribute_uses): Likewise.
2944 (add_current_iterators): Likewise.
2945 (apply_iterators): Likewise.
2946 (initialize_iterators): Guard usage of apply_subst_iterator with
2947 #ifdef GENERATOR_FILE.
2948 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2949 (md_reader::read_mapping): Likewise.
2950 (add_define_attr_for_define_subst): Likewise.
2951 (add_define_subst_attr): Likewise.
2952 (read_subst_mapping): Likewise.
2953 (check_code_iterator): Likewise.
2954 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2956 (one_time_initialization): New function.
2957 (rtx_reader::read_until): New method.
2958 (read_flags): New function.
2959 (parse_reg_note_name): New function.
2960 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2961 Handle reuse_rtx ids.
2962 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2963 Add parsing support for RTL dumps, mirroring the special-cases in
2964 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2965 values, and calling handle_any_trailing_information.
2966 (rtx_reader::read_rtx_operand): Convert return type from void
2967 to rtx, returning return_rtx. Handle case 'e'. Call
2968 finalize_string on XSTR and XTMPL fields.
2969 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2970 "(nil)" values were omitted. Call the postprocess vfunc on the
2972 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2973 class ctor. Initialize m_in_call_function_usage. Call
2974 one_time_initialization.
2975 * rtl-tests.c (selftest::test_uncond_jump): Call
2976 set_new_first_and_last_insn.
2977 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2978 * selftest-rtl.c: New file.
2979 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2980 (selftest::get_insn_by_uid): New decl.
2981 * selftest-run-tests.c (selftest::run_tests): Call
2982 read_rtl_function_c_tests.
2983 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2984 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2987 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2989 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2990 operands in a special way. Assert that pos+len <= mode precision.
2992 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2994 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2995 3 argument Alias with unlimited for the negative form.
2996 (fno-vect-cost-model): Removed.
2998 2017-01-05 Martin Liska <mliska@suse.cz>
3000 * hsa-gen.c (gen_hsa_divmod): New function.
3001 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3003 2017-01-05 Martin Liska <mliska@suse.cz>
3006 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3009 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3011 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3012 small constant length operands.
3014 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3016 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3017 between loop iterations.
3019 2017-01-05 Martin Liska <mliska@suse.cz>
3022 * gimplify.c (gimplify_decl_expr): Compare to
3023 asan_poisoned_variables instread of checking flags.
3024 (gimplify_target_expr): Likewise.
3025 (gimplify_expr): Likewise.
3026 (gimplify_function_tree): Conditionally initialize
3027 asan_poisoned_variables.
3029 2017-01-04 Jeff Law <law@redhat.com>
3031 PR tree-optimizatin/78812
3032 * rtl.h (contains_mem_rtx_p): Prototype.
3033 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3034 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3035 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3036 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3037 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3039 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3041 * input.c (assert_char_at_range): Default-initialize actual_range.
3043 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3045 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3046 to match the caller.
3048 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3050 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3051 insns after final jump in test to emit dummy move.
3053 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3055 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3056 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3058 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3060 * multiple_target.c (create_dispatcher_calls): Init e_next.
3061 * tree-ssa-loop-split.c (split_loop): Init border.
3062 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3065 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3070 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3071 (altivec_register_operand): Do not return true if the operand
3072 contains a SUBREG mixing SImode and SFmode.
3073 (vsx_register_operand): Likewise.
3074 (vsx_reg_sfsubreg_ok): New predicate.
3075 (vfloat_operand): Do not return true if the operand contains a
3076 SUBREG mixing SImode and SFmode.
3077 (vint_operand): Likewise.
3078 (vlogical_operand): Likewise.
3079 (gpc_reg_operand): Likewise.
3080 (int_reg_operand): Likewise.
3081 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3082 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3083 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3085 (rs6000_emit_move_si_sf_subreg): New helper function.
3086 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3087 fixup SUBREGs involving SImode and SFmode.
3088 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3089 numbers for the new peephole2 optimization.
3090 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3091 the GLIBC math library that do AND/IOR/XOR operations on single
3092 precision floating point.
3093 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3094 target macros to say whether we need to avoid SUBREGs mixing
3096 (TARGET_ALLOW_SF_SUBREG): Likewise.
3097 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3098 (UNSPEC_SI_FROM_SF): Likewise.
3099 (iorxor): Change spacing.
3100 (and_ior_xor): New iterator for AND, IOR, and XOR.
3101 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3102 (movdi_from_sf_zero_ext): Likewise.
3103 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3104 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3105 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3106 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3107 (fms<mode>4): Likewise.
3108 (fnma<mode>4): Likewise.
3109 (fnms<mode>4): Likewise.
3110 (nfma<mode>4): Likewise.
3111 (nfms<mode>4): Likewise.
3113 2017-01-04 Marek Polacek <polacek@redhat.com>
3116 * doc/invoke.texi: Document -Wpointer-compare.
3118 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3120 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3123 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3124 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3127 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3130 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3131 documentation of the powerpc_popcntb_ok attribute.
3132 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3133 code to issue warning messages if a requested CPU configuration is
3134 not supported by the binary (assembler and loader) toolchain.
3135 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3136 made to define a built-in function that has been disabled.
3137 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3138 made to define a built-in function that has been disabled.
3139 (altivec_init_builtins): Add comment explaining why definition
3140 of the DST built-in functions is not preceded by an assertion
3141 check. Add assertions to prevent ICE if attempts are made to
3142 define an altivec predicate or an abs* built-in function that has
3144 (htm_init_builtins): Add comment explaining why definition of the
3145 htm built-in functions is not preceded by an assertion check.
3147 2017-01-04 Jeff Law <law@redhat.com>
3149 PR tree-optimizatin/67955
3150 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3151 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3152 the points-to solution does not include pt_null. Use DECL_PT_UID
3155 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3157 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3158 Use gen_int_mode instead of gen_lopwart for const_int operands.
3160 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3162 PR tree-optimization/71563
3163 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3164 out of range value - has low bits known to be zero.
3166 2017-01-04 Alan Modra <amodra@gmail.com>
3168 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3169 * configure: Regenerate.
3170 * config.in: Regenerate.
3172 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3175 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3176 a substring of the message, but strcmp with the whole message. Ifdef
3177 ENABLE_NLS, translate the message first using dgettext.
3179 2017-01-03 Jeff Law <law@redhat.com>
3181 PR tree-optimizatin/78856
3182 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3183 (mark_threaded_blocks): Remove code to truncate thread paths that
3184 cross multiple loop headers. Instead invalidate the cached loop
3185 iteration information and handle case of a thread path walking
3186 into an irreducible region.
3188 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3191 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3192 assertions. Add support for doing the signbit if the IEEE 128-bit
3193 floating point value is in a GPR.
3194 * config/rs6000/rs6000.md (Fsignbit): Delete.
3195 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3196 Update the length attribute if the value is in a GPR.
3197 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3198 the sign or zero extension instruction, since the value is always 0/1.
3199 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3202 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3203 extracting SImode to a GPR register so that we can generate a
3204 store, limit the vector to be in a traditional Altivec register
3205 for the vextuwrx instruction.
3207 2017-01-03 Ian Lance Taylor <iant@google.com>
3209 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3211 2017-01-03 Martin Sebor <msebor@redhat.com>
3213 PR tree-optimization/78696
3214 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3215 precision. Use MPFR for %f for greater fidelity. Correct handling
3217 (pass_sprintf_length::compute_format_length): Set width and precision
3218 specified by asrerisk to void_node for vararg functions.
3219 (try_substitute_return_value): Adjust dump output.
3221 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3223 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3225 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3227 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3228 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3229 -mlra/-mno-lra was passed to the compiler.
3231 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3233 PR rtl-optimization/65618
3234 * emit-rtl.c (try_split): Move initialization of "before" and
3235 "after" to just before the call to emit_insn_after_setloc.
3237 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3239 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3241 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3243 * dwarf2out.c (gen_enumeration_type_die): When
3244 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3246 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3248 PR tree-optimization/78965
3249 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3250 Change first argument from const call_info & to call_info &. For %n
3251 set info.nowrite to false.
3254 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3255 possibly throwing calls.
3257 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3258 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3259 and fns handling, rather than in a separate case SSA_NAME.
3261 2017-01-02 Jeff Law <law@redhat.com>
3263 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3264 fixes for first_period and second_period variables.
3266 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3269 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3270 (*insvqi_1): New insn pattern.
3271 (*insvqi_1_mem_rex64): Ditto.
3273 (*insvqi_3): Rename from *insvqi.
3275 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3277 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3279 * doc/cfg.texi (Edges): Remove reference to Java.
3280 (Maintaining the CFG): Ditto.
3282 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3285 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3286 transparent aliases.
3288 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3291 * predict.def (PRED_CALL): Update hitrate.
3292 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3293 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3294 into direct/indirect/polymorphic variants.
3296 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3298 Update copyright years.
3300 * gcc.c (process_command): Update copyright notice dates.
3301 * gcov-dump.c (print_version): Ditto.
3302 * gcov.c (print_version): Ditto.
3303 * gcov-tool.c (print_version): Ditto.
3304 * gengtype.c (create_file): Ditto.
3305 * doc/cpp.texi: Bump @copying's copyright year.
3306 * doc/cppinternals.texi: Ditto.
3307 * doc/gcc.texi: Ditto.
3308 * doc/gccint.texi: Ditto.
3309 * doc/gcov.texi: Ditto.
3310 * doc/install.texi: Ditto.
3311 * doc/invoke.texi: Ditto.
3313 Copyright (C) 2017 Free Software Foundation, Inc.
3315 Copying and distribution of this file, with or without modification,
3316 are permitted in any medium without royalty provided the copyright
3317 notice and this notice are preserved.