1 2017-02-02 Martin Sebor <msebor@redhat.com>
3 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
6 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
8 (-fdump-tree-@var): Add to index and document how to come up
9 with pass-specific option and dump file names.
10 (-fdump-passes): Clarify where to look for output.
12 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
15 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
16 statistics of the analyzed path; allow threading for speed when
17 any of BBs along the path are optimized for speed.
19 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
22 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
23 settings of the virtual registers.
26 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
28 * explow.c (get_dynamic_stack_size): Take known alignment of stack
29 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
32 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
34 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
35 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
37 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
39 * config/s390/s390.md: Add missing comments with the expanded
41 * config/s390/vector.md: Likewise.
42 * config/s390/vx-builtins.md: Likewise.
44 2017-02-02 Jakub Jelinek <jakub@redhat.com>
47 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
48 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
49 conditions on a single line.
51 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
53 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
54 __S390_VX__ to __VX__.
56 2017-02-01 Andrew Pinski <apinski@cavium.com>
58 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
59 stmt_info to record_stmt_cost.
60 (vect_get_known_peeling_cost): Pass stmt_info if known to
62 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
63 cpu_vector_cost field into
64 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
65 field into vec_int_stmt_cost and vec_fp_stmt_cost.
66 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
67 splitting of scalar_stmt_cost and vec_stmt_cost.
68 (thunderx_vector_cost): Likewise.
69 (cortexa57_vector_cost): LIkewise.
70 (exynosm1_vector_cost): Likewise.
71 (xgene1_vector_cost): Likewise.
72 (thunderx2t99_vector_cost): Improve after the splitting of the two
74 (aarch64_builtin_vectorization_cost): Update for the splitting of
75 scalar_stmt_cost and vec_stmt_cost.
77 2017-02-01 Torvald Riegel <triegel@redhat.com>
78 Richard Henderson <rth@redhat.com>
80 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
81 conditional on existance of a fast atomic load.
82 * optabs-query.c (can_atomic_load_p): New function.
83 * optabs-query.h (can_atomic_load_p): Declare it.
84 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
85 no fast atomic load is available for the particular size of access.
86 (expand_atomic_compare_and_swap): Likewise.
87 (expand_atomic_load): Likewise.
88 (expand_atomic_store): Likewise.
89 (expand_atomic_fetch_op): Likewise.
90 * testsuite/lib/target-supports.exp
91 (check_effective_target_sync_int_128): Remove x86 because it provides
93 (check_effective_target_sync_int_128_runtime): Likewise.
95 2017-02-01 Richard Biener <rguenther@suse.de>
97 * graphite.c: Include tree-vectorizer.h for find_loop_location.
98 (graphite_transform_loops): Provide opt-info for optimized nests.
99 * tree-parloop.c (parallelize_loops): Provide opt-info for
102 2017-02-01 Richard Biener <rguenther@suse.de>
105 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
108 2017-02-01 Richard Biener <rguenther@suse.de>
110 PR tree-optimization/71824
111 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
112 Verify the loops are valid in the merged SESE region.
113 (scop_detection::can_represent_loop_1): Check analyzing the
114 evolution of the number of iterations in the region succeeds.
116 2017-01-31 Ian Lance Taylor <iant@golang.org>
118 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
119 REG_ARGS_SIZE note to 32-bit push insns and call insn.
121 2017-01-31 David Malcolm <dmalcolm@redhat.com>
123 PR preprocessor/79210
124 * input.c (get_substring_ranges_for_loc): Replace line_width
125 assertion with error-handling.
127 2017-01-31 Richard Biener <rguenther@suse.de>
129 PR tree-optimization/77318
130 * graphite-sese-to-poly.c (extract_affine): Fix assert.
131 (create_pw_aff_from_tree): Take loop parameter.
132 (add_condition_to_pbb): Pass loop of the condition to
133 create_pw_aff_from_tree.
135 2017-01-31 Jakub Jelinek <jakub@redhat.com>
137 * config/s390/s390.c (s390_asan_shadow_offset): New function.
138 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
140 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
144 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
146 (convert_int_to_float128): Likewise.
147 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
148 (convert_int_to_float128): Likewise.
149 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
150 (UNSPEC_IEEE128_CONVERT): Likewise.
151 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
152 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
153 Use local variables for IBM extended format.
154 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
155 (fix_trunc<mode>si2_fprs): Likewise.
156 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
157 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
158 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
159 to know that we can now have integers of all sizes in vector
161 (fix<uns>_<mode>di2_hw): Likewise.
162 (float<uns>_<mode>si2_hw): Likewise.
163 (fix_<mode>si2_hw): Likewise.
164 (fixuns_<mode>si2_hw): Likewise.
165 (float<uns>_<mode>di2_hw): Likewise.
166 (float_<mode>di2_hw): Likewise.
167 (float_<mode>si2_hw): Likewise.
168 (floatuns_<mode>di2_hw): Likewise.
169 (floatuns_<mode>si2_hw): Likewise.
170 (xscvqp<su>wz_<mode>): Delete, no longer used.
171 (xscvqp<su>dz_<mode>): Likewise.
172 (xscv<su>dqp_<mode>): Likewise.
173 (ieee128_mfvsrd_64bit): Likewise.
174 (ieee128_mfvsrd_32bit): Likewise.
175 (ieee128_mfvsrwz): Likewise.
176 (ieee128_mtvsrw): Likewise.
177 (ieee128_mtvsrd_64bit): Likewise.
178 (ieee128_mtvsrd_32bit): Likewise.
180 2017-01-31 Martin Liska <mliska@suse.cz>
183 * ipa-prop.c (ipa_free_all_node_params): Call release method
184 instead of ~sumbol_summary to not to trigger double times
187 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
189 PR tree-optimization/71691
190 * bitmap.h (class auto_bitmap): New.
191 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
192 is_maybe_undefined instead of ssa_undefined_value_p.
194 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
196 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
197 __S390_ARCH_LEVEL__ to __ARCH__.
199 2017-01-31 Jakub Jelinek <jakub@redhat.com>
201 PR tree-optimization/79267
202 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
203 if should_remove_lhs_p is true.
205 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
208 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
209 (add_alignment_attribute): New.
210 (base_type_die): Add alignment attribute.
211 (subrange_type_die): Likewise.
212 (modified_type_die): Likewise.
213 (gen_array_type_die): Likewise.
214 (gen_descr_array_type_die: Likewise.
215 (gen_enumeration_type_die): Likewise.
216 (gen_subprogram_die): Likewise.
217 (gen_variable_die): Likewise.
218 (gen_field_die): Likewise.
219 (gen_ptr_to_mbr_type_die): Likewise.
220 (gen_struct_or_union_type_die): Likewise.
221 (gen_subroutine_type_die): Likewise.
222 (gen_typedef_die): Likewise.
223 (base_type_cmp): Compare alignment attribute.
225 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
228 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
229 (setb_unsigned) New pattern for setb with CCUNS.
230 * config/rs6000/rs6000.c (expand_block_compare): Use a different
231 subfc./subfe sequence to avoid overflow problems. Generate a
232 shorter sequence with cmpld/setb for power9.
233 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
234 for generating subfc. instruction.
235 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
236 now uses this instruction.
238 2017-01-30 Ian Lance Taylor <iant@google.com>
241 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
242 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
244 2017-01-30 Martin Sebor <msebor@redhat.com>
246 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
247 Move constant to the right of a relational operator.
248 (get_mpfr_format_length, format_character, format_string): Ditto.
249 (should_warn_p, maybe_warn): Same.
251 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
253 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
256 * asan.c (get_translation_unit_decl): Remove function.
257 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
259 2017-01-30 Martin Liska <mliska@suse.cz>
261 PR gcov-profile/79259
262 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
265 2017-01-30 Martin Liska <mliska@suse.cz>
268 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
269 Initialize variables with NULL value.
271 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
274 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
276 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
278 2017-01-30 Richard Biener <rguenther@suse.de>
280 PR tree-optimization/79276
281 * tree-vrp.c (process_assert_insertions): Properly adjust common
282 when removing a duplicate.
284 * gcc.dg/torture/pr79276.c: New testcase.
286 2017-01-30 Richard Biener <rguenther@suse.de>
288 PR tree-optimization/79256
289 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
290 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
292 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
294 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
297 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
298 ("*r<noxa>sbg_<mode>_sll_bitmask")
299 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
300 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
301 Use contiguous_bitmask_nowrap_operand.
303 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
306 * config/rs6000/altivec.h (vec_xl): Revise #define.
309 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
311 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
313 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
315 PR rtl-optimization/79194
316 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
317 traps before call to bypass_conditional_jumps.
319 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
321 PR tree-optimization/71374
322 * lra-constraints.c (check_conflict_input_operands): New.
323 (match_reload): Use it.
325 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
328 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
329 account to calculate conflict_set.
331 2017-01-27 Bin Cheng <bin.cheng@arm.com>
333 PR rtl-optimization/78559
334 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
335 other_insn in combine.
337 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
339 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
340 uint16_type_node for BT_UINT16.
342 2017-01-27 David Malcolm <dmalcolm@redhat.com>
344 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
346 (GIMPLE Tests): New node.
347 (RTL Tests): New node.
349 2017-01-27 Richard Biener <rguenther@suse.de>
351 PR tree-optimization/79245
352 * tree-loop-distribution.c (distribute_loop): Apply cost
353 modeling also to detected patterns.
355 2017-01-27 Richard Biener <rguenther@suse.de>
357 PR tree-optimization/71433
358 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
359 (compare_assert_loc): New function.
360 (process_assert_insertions): Sort and optimize assert locations
361 to remove duplicates and push down identical assertions on
362 edges to their destination block.
364 2017-01-27 Richard Biener <rguenther@suse.de>
366 PR tree-optimization/79244
367 * tree-vrp.c (remove_range_assertions): Forcefully propagate
368 out SSA names even if abnormal.
370 2017-01-27 Jakub Jelinek <jakub@redhat.com>
372 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
373 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
374 instead of MPFR_RNDN.
376 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
379 * arm.c (arm_option_override): Don't call build_target_option_node
380 until after doing all option overrides.
381 (arm_valid_target_attribute_tree): Likewise.
383 2017-01-27 Martin Liska <mliska@suse.cz>
385 * doc/invoke.texi (-fprofile-arcs): Document profiling support
386 for {cd}tors and C++ {cd}tors.
388 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
390 * config/s390/s390.md ("*setmem_long_and")
391 ("*setmem_long_and_31z"): Use zero_extend instead of and.
393 2017-01-26 Martin Sebor <msebor@redhat.com>
395 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
398 2017-01-26 Martin Sebor <msebor@redhat.com>
400 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
401 HAVE_DFmode before using XFmode or DFmode.
402 (parse_directive): Avoid using the z length modifier to avoid
403 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
406 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
407 to accept adjustment as an array.
408 (get_int_range): New function.
409 (struct directive): Make width and prec arrays.
410 (directive::set_width, directive::set_precision): Call get_int_range.
411 (format_integer, format_floating): Handle width and precision ranges.
412 (format_string, parse_directive): Same.
414 2017-01-26 Jakub Jelinek <jakub@redhat.com>
417 * dwarf2out.c (generate_skeleton_bottom_up): For children with
418 comdat_type_p set, just clone them, but keep the children in the
422 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
423 which have direct callers with -fvar-tracking-assignments enabled
425 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
426 inside of type units.
428 2017-01-26 Martin Sebor <msebor@redhat.com>
431 * gimple-ssa-sprintf.c (struct result_range): Add likely and
433 (struct format_result): Replace number_chars, number_chars_min,
434 and number_chars_max with a single member of struct result_range.
436 (format_result::operator+=): Adjust.
437 (struct fmtresult): Remove bounded. Handle likely and unlikely
439 (fmtresult::adjust_for_width_or_precision): New function.
440 (fmtresult:type_max_digits): New function.
441 (bytes_remaining): Handle likely and unlikely counters.
442 (min_bytes_remaining): Remove.
443 (format_percent): Simplify.
444 (format_integer, format_floating): Set likely and unlikely counters.
445 (get_string_length, format_character, format_string): Same.
446 (format_plain, should_warn_p): New function.
447 (maybe_warn): Call should_warn_p. Update diagnostic messages
448 and handle those for all directives, including plain strings.
449 (format_directive): Handle likely and unlikely counters.
450 Remove unnecessary quoting from diagnostics. Add an informational
453 (pass_sprintf_length::compute_format_length): Simplify.
454 (try_substitute_return_value): Handle likely and unlikely counters.
456 2017-01-26 Carl Love <cel@us.ibm.com>
458 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
459 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
461 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
464 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
465 endianess for subregs into account.
466 * lra-constraints.c (lra_constraints): Do risky transformations
467 always on the first iteration.
468 * lra-lives.c (check_pseudos_live_through_calls): Add arg
469 last_call_used_reg_set.
470 (process_bb_lives): Define and use last_call_used_reg_set.
471 * lra.c (lra): Always continue after lra_constraints on the first
474 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
476 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
478 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
480 2017-01-26 Jakub Jelinek <jakub@redhat.com>
482 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
483 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
484 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
485 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
486 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
487 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
488 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
489 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
490 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
492 2017-01-26 Marek Polacek <polacek@redhat.com>
495 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
496 for the third operand.
498 2017-01-26 Jakub Jelinek <jakub@redhat.com>
501 * omp-low.c (struct omp_context): Add simt_stmt field.
502 (scan_omp_for): Return omp_context *.
503 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
504 context to the _simt_ SIMD stmt.
505 (lower_omp_for): For combined SIMD with sibling _simt_
506 SIMD, make sure to use the same decls in _looptemp_
507 clauses as in the sibling.
509 2017-01-26 David Sherwood <david.sherwood@arm.com>
512 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
515 2017-01-26 Jakub Jelinek <jakub@redhat.com>
518 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
519 emit fld b; fld a; if possible.
521 * brig-builtins.def: Update copyright years.
522 * config/arm/arm_acle_builtins.def: Update copyright years.
524 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
527 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
528 constraint instead of o for the stxsd instruction.
530 2017-01-25 Carl Love <cel@us.ibm.com>
532 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
533 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
535 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
537 * doc/invoke.texi (C++ Dialect Options): Fix typo.
539 2017-01-25 Richard Biener <rguenther@suse.de>
541 PR tree-optimization/69264
542 * target.def (vector_alignment_reachable): Improve documentation.
543 * doc/tm.texi: Regenerate.
544 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
546 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
547 earlier changes with respect to TYPE_USER_ALIGN.
548 (vector_alignment_reachable_p): Likewise. Improve dumping.
550 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
553 * config/arm/arm.md (xordi3): Force constant operand into a register
556 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
558 * doc/invoke.texi (-fstore-merging): Correct default optimization
559 levels at which it is enabled.
560 (-O): Move -fstore-merging from list to...
563 2017-01-25 Richard Biener <rguenther@suse.de>
566 * omp-expand.c: Include debug.h.
567 (expand_omp_taskreg): Make sure to generate early debug before
568 outlining anything from a function.
569 (expand_omp_target): Likewise.
570 (grid_expand_target_grid_body): Likewise.
572 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
575 * asan.c (get_translation_unit_decl): New function.
576 (asan_add_global): Extract modules file name from globals
577 TRANSLATION_UNIT_DECL name.
579 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
582 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
583 for long calls with APCS frame and VFP.
585 2017-01-24 David Malcolm <dmalcolm@redhat.com>
587 * cfg.c (original_copy_tables_initialized_p): New function.
588 * cfg.h (original_copy_tables_initialized_p): New decl.
589 * cfgrtl.c (relink_block_chain): Guard the call to
590 free_original_copy_tables with a call to
591 original_copy_tables_initialized_p.
592 * cgraph.h (symtab_node::native_rtl_p): New decl.
593 * cgraphunit.c (symtab_node::native_rtl_p): New function.
594 (symtab_node::needed_p): Don't assert for early assembly output
596 (cgraph_node::finalize_function): Set "force_output" for __RTL
598 (cgraph_node::analyze): Bail out early for __RTL functions.
599 (analyze_functions): Update assertion to support __RTL functions.
600 (cgraph_node::expand): Bail out early for __RTL functions.
601 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
603 * function.h (struct function): Update comment for field
605 * gimple-expr.c: Include "tree-pass.h".
606 (gimple_has_body_p): Return false for __RTL functions.
607 * Makefile.in (OBJS): Add run-rtl-passes.o.
608 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
610 (gcc::pass_manager::get_clean_slate): New accessor.
611 * passes.c: Include "insn-addr.h".
612 (should_skip_pass_p): Add logging. Update logic for running
613 "expand" to be compatible with both __GIMPLE and __RTL. Guard
614 property-provider override so it is only done for gimple passes.
616 (skip_pass): New function.
617 (execute_one_pass): Call skip_pass when skipping passes.
618 * read-md.c (md_reader::read_char): Support filtering
619 the input to a subset of line numbers.
620 (md_reader::md_reader): Initialize fields
621 m_first_line and m_last_line.
622 (md_reader::read_file_fragment): New function.
623 * read-md.h (md_reader::read_file_fragment): New decl.
624 (md_reader::m_first_line): New field.
625 (md_reader::m_last_line): New field.
626 * read-rtl-function.c (function_reader::create_function): Only
627 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
628 curr_properties. Set DECL_INITIAL to a dummy block.
629 (read_rtl_function_body_from_file_range): New function.
630 * read-rtl-function.h (read_rtl_function_body_from_file_range):
632 * run-rtl-passes.c: New file.
633 * run-rtl-passes.h: New file.
635 2017-01-24 Jeff Law <law@redhat.com>
637 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
640 2017-01-24 Bin Cheng <bin.cheng@arm.com>
642 PR tree-optimization/79159
643 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
644 (record_nonwrapping_iv): Improve boundary using above function if no
645 value range information.
647 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
648 Martin Jambor <mjambor@suse.cz>
650 * brig-builtins.def: New file.
651 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
652 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
653 (DEF_HSAIL_SAT_BUILTIN): Likewise.
654 (DEF_HSAIL_INTR_BUILTIN): Likewise.
655 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
656 * builtin-types.def (BT_INT8): New.
657 (BT_INT16): Likewise.
658 (BT_UINT8): Likewise.
659 (BT_UINT16): Likewise.
660 (BT_FN_ULONG): Likewise.
661 (BT_FN_UINT_INT): Likewise.
662 (BT_FN_UINT_ULONG): Likewise.
663 (BT_FN_UINT_LONG): Likewise.
664 (BT_FN_UINT_PTR): Likewise.
665 (BT_FN_ULONG_PTR): Likewise.
666 (BT_FN_INT8_FLOAT): Likewise.
667 (BT_FN_INT16_FLOAT): Likewise.
668 (BT_FN_UINT32_FLOAT): Likewise.
669 (BT_FN_UINT16_FLOAT): Likewise.
670 (BT_FN_UINT8_FLOAT): Likewise.
671 (BT_FN_UINT64_FLOAT): Likewise.
672 (BT_FN_UINT16_UINT32): Likewise.
673 (BT_FN_UINT32_UINT16): Likewise.
674 (BT_FN_UINT16_UINT16_UINT16): Likewise.
675 (BT_FN_INT_PTR_INT): Likewise.
676 (BT_FN_UINT_PTR_UINT): Likewise.
677 (BT_FN_LONG_PTR_LONG): Likewise.
678 (BT_FN_ULONG_PTR_ULONG): Likewise.
679 (BT_FN_VOID_UINT64_UINT64): Likewise.
680 (BT_FN_UINT8_UINT8_UINT8): Likewise.
681 (BT_FN_INT8_INT8_INT8): Likewise.
682 (BT_FN_INT16_INT16_INT16): Likewise.
683 (BT_FN_INT_INT_INT): Likewise.
684 (BT_FN_UINT_FLOAT_UINT): Likewise.
685 (BT_FN_FLOAT_UINT_UINT): Likewise.
686 (BT_FN_ULONG_UINT_UINT): Likewise.
687 (BT_FN_ULONG_UINT_PTR): Likewise.
688 (BT_FN_ULONG_ULONG_ULONG): Likewise.
689 (BT_FN_UINT_UINT_UINT): Likewise.
690 (BT_FN_VOID_UINT_PTR): Likewise.
691 (BT_FN_UINT_UINT_PTR: Likewise.
692 (BT_FN_UINT32_UINT64_PTR): Likewise.
693 (BT_FN_INT_INT_UINT_UINT): Likewise.
694 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
695 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
696 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
697 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
698 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
699 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
700 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
701 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
702 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
703 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
704 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
705 * doc/frontends.texi: List BRIG FE.
706 * doc/install.texi (Testing): Add BRIG tesring requirements.
707 * doc/invoke.texi (Overall Options): Mention BRIG.
708 * doc/standards.texi (Standards): Doucment BRIG HSA version.
710 2017-01-24 Richard Biener <rguenther@suse.de>
713 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
715 2017-01-24 Martin Jambor <mjambor@suse.cz>
718 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
719 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
722 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
725 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
726 casts from signed to unsigned really don't have a range.
728 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
730 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
731 GMP_RNDx for compatiblity.
733 2017-01-24 Martin Liska <mliska@suse.cz>
736 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
737 that would prevent us to call alloca with -1 as argument.
739 2017-01-24 Jakub Jelinek <jakub@redhat.com>
741 * dwarf2out.c (output_compilation_unit_header, output_file_names):
742 Avoid -Wformat-security warning.
744 2017-01-23 Andrew Pinski <apinski@cavium.com>
746 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
749 2017-01-23 Martin Sebor <msebor@redhat.com>
752 * gimple-ssa-sprintf.c (warn_level): New global.
753 (format_integer): Use it here and throughout the rest of the file.
754 Use the same switch to compute sign as base.
755 (maybe_warn): New function.
756 (format_directive): Factor out warnings into maybe_warn.
757 Add debugging output. Use warn_level.
758 (add_bytes): Use warn_level.
759 (pass_sprintf_length::compute_format_length): Add debugging output.
760 (try_substitute_return_value): Same.
761 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
764 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
765 (struct fmtresult, format_integer, format_floating): Adjust.
766 (fmtresult::fmtresult): Set max correctly in two argument ctor.
767 (get_string_length, format_string,format_directive): Same.
768 (pass_sprintf_length::compute_format_length): Same.
769 (try_substitute_return_value): Simplify slightly.
772 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
773 (fmtresult::operator+=): Outlined.
774 (struct fmtresult): Add ctors.
775 (struct conversion_spec): Rename...
776 (struct directive): ...to this. Add and remove data members.
777 (directive::set_width, directive::set_precision): New functions.
778 (format_percent): Use fmtresult ctor.
779 (get_width_and_precision): Remove.
780 (format_integer): Make naming changes. Avoid computing width and
782 (format_floating): Same. Adjust indentation.
783 (format_character, format_none): New functions.
784 (format_string): Moved character handling to format_character.
785 (format_directive): Remove arguments, change return type.
786 (parse_directive): New function.
787 (pass_sprintf_length::compute_format_length): Move directive
788 parsing to parse_directive.
790 2017-01-23 Jakub Jelinek <jakub@redhat.com>
792 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
793 (assign_assembler_name_if_needed): ... this.
794 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
795 (assign_assembler_name_if_needed): ... this.
796 (free_lang_data_in_cgraph): Adjust callers.
797 * cgraphunit.c (cgraph_node::analyze): Likewise.
798 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
801 2017-01-23 Richard Biener <rguenther@suse.de>
803 PR tree-optimization/79088
804 PR tree-optimization/79188
805 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
806 resetting loop bounds after last path deletion. Reset loop
807 bounds of the target loop, make code match the comments.
808 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
809 Make sure loops need no fixups.
811 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
813 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
814 exponent support with double type for first argument.
815 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
816 type returned by __builtin_vec_extract_sig,
817 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
818 functions from "vector int" to "vector unsigned int" or from
819 "vector long long int" to "vector unsigned long long int".
820 Changed type returned by __builtin_vec_extract_exp,
821 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
822 functions from "vector int" to "vector unsigned int" or from
823 "vector long long int" to "vector unsigned long long int".
824 Changed return type of __builtin_vec_test_data_class,
825 __builtin_vec_test_data_class_sp, and
826 __builtin_vec_test_data_class_dp from "vector int" to
827 "vector bool int" or from "vector long long int" to "vector bool
828 long long int" and changed second argument type from "unsigned
829 int" to "int". Added new overloaded function forms "vector float
830 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
831 "vector float __builtin_vec_insert_exp_sp (vector float, vector
832 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
833 double, vector unsigned long long int)" and "vector double
834 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
835 long int)". Changed return type of
836 __builtin_scalar_test_data_class and
837 __builtin_scalar_test_data_class_sp and
838 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
839 int" and changed second argument from "unsigned int" to "int".
840 Changed type returned by __builtin_scalar_test_neg,
841 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
842 from "int" to "bool int". Added new overloaded function form
843 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
844 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
845 exponent double-precision with floating point first argument.
846 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
847 documentation of scalar_test_data_class, scalar_test_neg,
848 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
849 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
850 vec_test_data_class built-in functions to reflect refinements in
851 their type signatures.
853 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
855 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
857 (aarch64_elf_asm_destructor): Likewise.
859 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
861 PR rtl-optimization/78634
862 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
863 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
864 * ifcvt.c (noce_try_cmove): Add missing cost check.
866 PR rtl-optimization/71724
867 * combine.c (if_then_else_cond): Look for situations where it is
868 beneficial to undo the work of one of the recursive calls.
870 2017-01-23 Bin Cheng <bin.cheng@arm.com>
872 PR tree-optimization/70754
873 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
874 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
875 combined stmt before it if not NULL.
876 (combine_chains): Process refs reversely and compute dominance point
879 2017-01-23 Martin Liska <mliska@suse.cz>
881 PR tree-optimization/79196
882 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
883 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
885 (strlen_optimize_stmt): Call the renamed function.
887 2017-01-23 Michael Matz <matz@suse.de>
889 PR tree-optimization/78384
890 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
892 2017-01-23 Richard Biener <rguenther@suse.de>
894 PR tree-optimization/79186
895 * tree-vrp.c (register_new_assert_for): Make sure we've seen
896 both incoming edges before moving an assert.
898 2017-01-23 Martin Jambor <mjambor@suse.cz>
900 * ipa-prop.c (load_from_param_1): Removed.
901 (load_from_unmodified_param): Bits from load_from_param_1 put back
903 (load_from_param): Removed.
904 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
905 with stmt. Reverted back to use of load_from_unmodified_param.
907 2017-01-23 Martin Jambor <mjambor@suse.cz>
910 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
911 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
912 field a pointer to garbage collected vector, mark lattices and
913 ipcp_orig_node with GTY((skip)).
914 (ipa_get_param_count): Adjust to descriptors being a pointer.
915 (ipa_get_param): Likewise.
916 (ipa_get_type): Likewise.
917 (ipa_get_param_move_cost): Likewise.
918 (ipa_set_param_used): Likewise.
919 (ipa_get_controlled_uses): Likewise.
920 (ipa_set_controlled_uses): Likewise.
921 (ipa_is_param_used): Likewise.
922 (ipa_node_params_t): Move into garbage collector. New methods insert
924 (ipa_node_params_sum): Annotate wth GTY(()).
925 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
927 (ipa_load_from_parm_agg): Adjust declaration.
928 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
929 * ipa-profile.c (ipa_profile): Likewise.
930 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
931 (ipa_populate_param_decls): Make descriptors parameter garbage
933 (ipa_dump_param): Adjust to descriptors being a pointer.
934 (ipa_alloc_node_params): Likewise.
935 (ipa_initialize_node_params): Likewise.
936 (load_from_param_1): Make descriptors parameter garbage collected.
937 (load_from_unmodified_param): Likewise.
938 (load_from_param): Likewise.
939 (ipa_load_from_parm_agg): Likewise.
940 (ipa_node_params::~ipa_node_params): Removed.
941 (ipa_free_all_node_params): Remove call to delete operator.
942 (ipa_node_params_t::insert): New.
943 (ipa_node_params_t::remove): Likewise.
944 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
945 copy known_csts and known_contexts vectors.
946 (ipa_read_node_info): Adjust to descriptors being a pointer.
947 (ipcp_modif_dom_walker): Make m_descriptors field garbage
949 (ipcp_transform_function): Make descriptors variable garbage
952 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
954 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
955 * config/i386/avx512dqintrin.h: Ditto.
956 * config/i386/avx512fintrin.h: Ditto.
957 * gcc/config/i386/i386.c: Handle new builtins.
958 * config/i386/i386-builtin.def: Add new builtins.
959 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
960 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
962 2017-01-23 Jakub Jelinek <jakub@redhat.com>
963 Martin Liska <mliska@suse.cz>
965 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
966 * asan.c (asan_expand_poison_ifn): Support stores and use
967 appropriate ASAN report function.
968 * internal-fn.c (expand_ASAN_POISON_USE): New function.
969 * internal-fn.def (ASAN_POISON_USE): Declare.
970 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
971 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
972 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
973 ASAN_POISON calls w/o LHS.
974 * tree-ssa.c (execute_update_addresses_taken): Create clobber
975 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
976 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
977 * gimplify.c (asan_poison_variables): Add attribute
978 use_after_scope_memory to variables that really needs to live
980 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
981 having the attribute.
983 2017-01-23 Martin Liska <mliska@suse.cz>
985 * asan.c (create_asan_shadow_var): New function.
986 (asan_expand_poison_ifn): Likewise.
987 * asan.h (asan_expand_poison_ifn): New declaration.
988 * internal-fn.c (expand_ASAN_POISON): Likewise.
989 * internal-fn.def (ASAN_POISON): New builtin.
990 * sanopt.c (pass_sanopt::execute): Expand
991 asan_expand_poison_ifn.
992 * tree-inline.c (copy_decl_for_dup_finish): Make function
994 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
995 * tree-ssa.c (is_asan_mark_p): New function.
996 (execute_update_addresses_taken): Rewrite local variables
997 (identified just by use-after-scope as addressable) into SSA.
999 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1001 * doc/install.texi (Specific): opensource.apple.com uses https
1002 now. Remove trailing slash.
1004 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1006 * README.Portability: Remove note on an Irix compatibility issue.
1008 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1010 * gcov.c (INCLUDE_ALGORITHM): Define.
1011 (INCLUDE_VECTOR): Define.
1012 No longer include <vector> and <algorithm> directly.
1014 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1016 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1018 * doc/invoke.texi (Code Gen Options): Ditto.
1020 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1023 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1025 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1027 rtl-optimization/79125
1028 * cprop.c (local_cprop_pass): Handle cases where we make an
1031 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1035 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1036 read from, for big endian.
1038 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1040 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1041 register pauth builtins for LP64 only.
1043 2017-01-20 Marek Polacek <polacek@redhat.com>
1046 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1049 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1051 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1053 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1054 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1055 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1057 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1060 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1061 in big-endian mode when they are not a single duplicated value.
1063 2017-01-20 Richard Biener <rguenther@suse.de>
1065 * BASE-VER: Bump to 7.0.1.
1067 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1069 * omp-low.c (omplow_simd_context): New struct. Use it...
1070 (lower_rec_simd_input_clauses): ...here and...
1071 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1072 references to idx, lane, max_vf, is_simt.
1074 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1076 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1079 2017-01-20 Martin Jambor <mjambor@suse.cz>
1081 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1082 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1084 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1085 (GTFILES): Rename hsa.c to hsa-common.c.
1086 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1087 * hsa-dump.c: Likewise.
1088 * hsa-gen.c: Likewise.
1089 * hsa-regalloc.c: Likewise.
1090 * ipa-hsa.c: Likewise.
1091 * omp-expand.c: Likewise.
1092 * omp-low.c: Likewise.
1093 * toplev.c: Likewise.
1095 2017-01-20 Marek Polacek <polacek@redhat.com>
1098 * doc/invoke.texi: Document -Wduplicated-branches.
1099 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1100 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1101 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1102 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1103 return 0 only when not OEP_LEXICOGRAPHIC.
1104 (fold_build_cleanup_point_expr): Use the expression
1105 location when building CLEANUP_POINT_EXPR.
1106 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1107 * tree.c (add_expr): Handle error_mark_node.
1109 2017-01-20 Martin Liska <mliska@suse.cz>
1112 * tree-profile.c (init_ic_make_global_vars): Do not call
1114 (gimple_init_gcov_profiler): Likewise.
1116 2017-01-20 Martin Liska <mliska@suse.cz>
1119 * cgraph.h (maybe_create_reference): Remove argument and
1121 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1123 * ipa-cp.c (create_specialized_node): Likewise.
1124 * symtab.c (symtab_node::maybe_create_reference): Handle
1125 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1127 2017-01-20 Martin Liska <mliska@suse.cz>
1129 * read-rtl-function.c (function_reader::create_function): Use
1130 build_decl instread of build_decl_stat.
1132 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1134 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1135 * config/i386/avx512dqintrin.h: Ditto.
1136 * config/i386/avx512fintrin.h: Ditto.
1137 * config/i386/i386-builtin-types.def: Add new types.
1138 * gcc/config/i386/i386.c: Handle new types.
1139 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1140 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1141 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1142 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1143 (__builtin_ia32_kshiftridi): New.
1144 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1146 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1150 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1151 define to rs6000_init_stack_protect_guard.
1152 (rs6000_init_stack_protect_guard): New function.
1154 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1155 Yunqiang Su <yunqiang.su@imgtec.com>
1157 * config.gcc (supported_defaults): Add madd4.
1158 (with_madd4): Add validation.
1159 (all_defaults): Add madd4.
1160 * config/mips/mips.opt (mmadd4): New option.
1161 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1163 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1165 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1166 (ISA_HAS_FUSED_MADD4): Likewise.
1167 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1168 * gcc/doc/install.texi (--with-madd4): Document the new option.
1170 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1172 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1173 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1174 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1175 (aarch64_init_pauth_hint_builtins): New.
1176 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1177 (aarch64_expand_builtin): Expand new builtins.
1179 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1181 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1182 * combine-stack-adj.c (no_unhandled_cfa): Handle
1183 REG_CFA_TOGGLE_RA_MANGLE.
1184 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1185 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1186 info for return address signing.
1187 (aarch64_expand_epilogue): Likewise.
1189 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1191 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1192 * config/aarch64/aarch64-protos.h
1193 (aarch64_return_address_signing_enabled): New declaration.
1194 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1196 (aarch64_expand_prologue): Sign return address before it's pushed onto
1198 (aarch64_expand_epilogue): Authenticate return address fetched from
1200 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1201 (aarch64_attributes): New function attributes for "sign-return-address".
1202 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1203 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1204 ("*do_return"): Generate combined instructions according to key index.
1205 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1206 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1208 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1209 * config/aarch64/aarch64.opt (msign-return-address=): New.
1210 * doc/extend.texi (AArch64 Function Attributes): Documents
1211 "sign-return-address=".
1212 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1214 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1216 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1217 overall option summary.
1219 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1221 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1222 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1223 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1224 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1226 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1228 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1229 -mpower9-minmax by default for -mcpu=power9.
1230 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1231 128-bit floating point.
1233 2017-01-20 Alan Modra <amodra@gmail.com>
1235 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1236 optimizing for size.
1238 2017-01-20 Alan Modra <amodra@gmail.com>
1241 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1242 for strcmp and strncmp from corresponding builtin decl.
1244 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1246 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1247 instead of i386/rtems-64.h.
1248 * config/i386/rtems-64.h: Remove.
1250 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1254 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1256 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1258 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1260 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1261 Change int to HOST_WIDE_INT.
1262 * config/aarch64/aarch64-protos.h
1263 (aarch64_simd_gen_const_vector_dup): Likewise.
1264 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1266 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1268 * langhooks-def.h (lhd_type_for_size): New decl.
1269 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1270 * langhooks.c (lhd_type_for_size): New function, taken from
1273 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1275 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1276 define_bypass for CR latency.
1277 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1278 (power9-alu2): Add define_bypass for CR latency.
1280 (power9-mul): Update insn latency.
1281 (power9-mul-compare): Update insn latency, bypass latency and remove
1284 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1286 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1288 * config/aarch64/aarch64.md
1289 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1290 aarch64_nopcrelative_literal_loads.
1291 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1293 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1295 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1297 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1299 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1302 * config.gcc (supported_defaults): Add lxc1-sxc1.
1303 (with_lxc1_sxc1): Add validation.
1304 (all_defaults): Add lxc1-sxc1.
1305 * config/mips/mips.opt (mlxc1-sxc1): New option.
1306 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1308 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1309 __mips_no_lxc1_sxc1.
1310 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1311 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1312 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1314 2017-01-19 Richard Biener <rguenther@suse.de>
1316 PR tree-optimization/72488
1317 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1318 sure to restore SSA info.
1319 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1321 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1323 PR rtl-optimization/79121
1324 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1325 of the inner type when shifting an extended value.
1327 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1330 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1331 interposable aliases.
1333 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1336 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1337 Use the evmergelohi instruction.
1338 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1339 (mov_sitf_e500_subreg8_2_be): Likewise.
1340 (mov_sitf_e500_subreg12_2_le): Likewise.
1341 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1342 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1343 (mov_sitf_e500_subreg8_2_le): Likewise.
1344 (mov_sitf_e500_subreg12_2_be): Likewise.
1346 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1348 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1349 attribute from vecsimple to vecperm.
1350 (altivec_vbpermq2): Likewise.
1352 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1355 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1357 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1358 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1359 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1360 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1361 case where N arg is SIZE_MAX.
1362 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1363 (cmpstrsi): Add pattern.
1365 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1367 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1368 __builtin_vec_revb builtins.
1369 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1370 built-in functions to support generation of the ISA 3.0 XXBR<x>
1371 vector byte reverse instructions.
1372 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1373 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1374 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1375 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1376 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1377 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1378 (P9V_BUILTIN_VEC_REVB): Likewise.
1379 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1380 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1381 (p9_xxbrq_v16qi): Likewise.
1382 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1383 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1384 (p9_xxbrh_v8hi): Likewise.
1385 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1386 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1387 vec_revb built-in functions.
1389 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1391 PR rtl-optimization/78952
1392 * config/i386/i386.md (any_extract): New code iterator.
1393 (*insvqi_2): Use any_extract for source operand.
1394 (*insvqi_3): Use any_shiftrt for source operand.
1396 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1398 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1400 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1402 2017-01-18 Matthias Klose <doko@ubuntu.com>
1404 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1406 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1408 * config/rs6000/altivec.h (vec_bperm): Change #define.
1409 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1410 (altivec_vbpermq2): New define_insn.
1411 (altivec_vbpermd): Likewise.
1412 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1414 (VBPERMD): Likewise.
1415 (VBPERM): New polymorphic function interface.
1416 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1417 Add entries for P9V_BUILTIN_VEC_VBPERM.
1418 * doc/extend.texi: Add interfaces for vec_bperm.
1420 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1422 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1423 first letter of error messages.
1424 (s390_resolve_overloaded_builtin): Likewise.
1425 * config/s390/s390.c (s390_expand_builtin): Likewise.
1426 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1427 (s390_valid_target_attribute_inner_p): Likewise.
1428 * config/s390/s390.md ("tabort"): Likewise.
1430 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1432 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1433 (ISA_AVOID_DIV_HILO): New macro.
1434 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1435 (ISA_HAS_DDIV): Likewise.
1437 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1439 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1441 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1443 * doc/invoke.texi (fabi-version): Spelling fix.
1445 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1448 * doc/invoke.texi (fabi-version): Mention mangling fix for
1451 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1454 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1456 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1459 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1460 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1462 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1464 (stack_protect_test): Ditto.
1465 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1466 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1468 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1469 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1470 -mstack-protector-guard-offset=.
1471 (RS/6000 and PowerPC Options): Ditto.
1473 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1475 * config/i386/i386.h (MASK_CLASS_P): New define.
1476 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1477 there are no registers from different register sets also when
1478 mask registers are used. Update function comment.
1479 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1480 to (*k/*r) and (*k/*km) alternatives.
1482 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1484 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1485 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1486 (EH_RETURN_HANDLER_RTX): New define.
1487 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1488 Force frame pointer in EH return functions.
1489 (aarch64_expand_epilogue): Add barrier for eh_return.
1490 (aarch64_final_eh_return_addr): Remove.
1491 (aarch64_eh_return_handler_rtx): New function.
1492 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1494 (aarch64_eh_return_handler_rtx): New prototype.
1496 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1498 * config/rs6000/altivec.h (vec_rlmi): New #define.
1499 (vec_vrlnm): Likewise.
1500 (vec_rlnm): Likewise.
1501 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1502 (UNSPEC_VRLNM): Likewise.
1503 (VIlong): New mode iterator.
1504 (altivec_vrl<VI_char>mi): New define_insn.
1505 (altivec_vrl<VI_char>nm): Likewise.
1506 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1509 (RLNM): New polymorphic function entry.
1510 (VRLWMI): New monomorphic function entry.
1512 (RLMI): New polymorphic function entry.
1513 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1514 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1515 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1518 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1521 * dwarf2out.c (field_byte_offset): Restore the
1522 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1523 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1524 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1527 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1530 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1532 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1535 * dwarf2out.c (add_data_member_location_attribute): For constant
1536 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1537 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1538 DW_AT_byte_size attributes.
1540 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1542 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1543 after forcing to constant memory when the code model is medium.
1545 2017-01-17 Julia Koval <julia.koval@intel.com>
1548 * config/i386/avx512fintrin.h
1549 (_mm512_i32gather_ps): Change __addr type to void const*.
1550 (_mm512_mask_i32gather_ps): Ditto.
1551 (_mm512_i32gather_pd): Ditto.
1552 (_mm512_mask_i32gather_pd): Ditto.
1553 (_mm512_i64gather_ps): Ditto.
1554 (_mm512_mask_i64gather_ps): Ditto.
1555 (_mm512_i64gather_pd): Ditto.
1556 (_mm512_mask_i64gather_pd): Ditto.
1557 (_mm512_i32gather_epi32): Ditto.
1558 (_mm512_mask_i32gather_epi32): Ditto.
1559 (_mm512_i32gather_epi64): Ditto.
1560 (_mm512_mask_i32gather_epi64): Ditto.
1561 (_mm512_i64gather_epi32): Ditto.
1562 (_mm512_mask_i64gather_epi32): Ditto.
1563 (_mm512_i64gather_epi64): Ditto.
1564 (_mm512_mask_i64gather_epi64): Ditto.
1565 (_mm512_i32scatter_ps): Change __addr type to void*.
1566 (_mm512_mask_i32scatter_ps): Ditto.
1567 (_mm512_i32scatter_pd): Ditto.
1568 (_mm512_mask_i32scatter_pd): Ditto.
1569 (_mm512_i64scatter_ps): Ditto.
1570 (_mm512_mask_i64scatter_ps): Ditto.
1571 (_mm512_i64scatter_pd): Ditto.
1572 (_mm512_mask_i64scatter_pd): Ditto.
1573 (_mm512_i32scatter_epi32): Ditto.
1574 (_mm512_mask_i32scatter_epi32): Ditto.
1575 (_mm512_i32scatter_epi64): Ditto.
1576 (_mm512_mask_i32scatter_epi64): Ditto.
1577 (_mm512_i64scatter_epi32): Ditto.
1578 (_mm512_mask_i64scatter_epi32): Ditto.
1579 (_mm512_i64scatter_epi64): Ditto.
1580 (_mm512_mask_i64scatter_epi64): Ditto.
1581 * config/i386/avx512pfintrin.h
1582 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1583 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1584 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1585 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1586 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1587 (_mm512_prefetch_i32scatter_ps): Ditto.
1588 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1589 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1590 (_mm512_prefetch_i64scatter_pd): Ditto.
1591 (_mm512_prefetch_i64scatter_ps): Ditto.
1592 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1593 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1594 * config/i386/avx512vlintrin.h
1595 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1596 (_mm_mmask_i32gather_ps): Ditto.
1597 (_mm256_mmask_i32gather_pd): Ditto.
1598 (_mm_mmask_i32gather_pd): Ditto.
1599 (_mm256_mmask_i64gather_ps): Ditto.
1600 (_mm_mmask_i64gather_ps): Ditto.
1601 (_mm256_mmask_i64gather_pd): Ditto.
1602 (_mm_mmask_i64gather_pd): Ditto.
1603 (_mm256_mmask_i32gather_epi32): Ditto.
1604 (_mm_mmask_i32gather_epi32): Ditto.
1605 (_mm256_mmask_i32gather_epi64): Ditto.
1606 (_mm_mmask_i32gather_epi64): Ditto.
1607 (_mm256_mmask_i64gather_epi32): Ditto.
1608 (_mm_mmask_i64gather_epi32): Ditto.
1609 (_mm256_mmask_i64gather_epi64): Ditto.
1610 (_mm_mmask_i64gather_epi64): Ditto.
1611 (_mm256_i32scatter_ps): Change __addr type to void*.
1612 (_mm256_mask_i32scatter_ps): Ditto.
1613 (_mm_i32scatter_ps): Ditto.
1614 (_mm_mask_i32scatter_ps): Ditto.
1615 (_mm256_i32scatter_pd): Ditto.
1616 (_mm256_mask_i32scatter_pd): Ditto.
1617 (_mm_i32scatter_pd): Ditto.
1618 (_mm_mask_i32scatter_pd): Ditto.
1619 (_mm256_i64scatter_ps): Ditto.
1620 (_mm256_mask_i64scatter_ps): Ditto.
1621 (_mm_i64scatter_ps): Ditto.
1622 (_mm_mask_i64scatter_ps): Ditto.
1623 (_mm256_i64scatter_pd): Ditto.
1624 (_mm256_mask_i64scatter_pd): Ditto.
1625 (_mm_i64scatter_pd): Ditto.
1626 (_mm_mask_i64scatter_pd): Ditto.
1627 (_mm256_i32scatter_epi32): Ditto.
1628 (_mm256_mask_i32scatter_epi32): Ditto.
1629 (_mm_i32scatter_epi32): Ditto.
1630 (_mm_mask_i32scatter_epi32): Ditto.
1631 (_mm256_i32scatter_epi64): Ditto.
1632 (_mm256_mask_i32scatter_epi64): Ditto.
1633 (_mm_i32scatter_epi64): Ditto.
1634 (_mm_mask_i32scatter_epi64): Ditto.
1635 (_mm256_i64scatter_epi32): Ditto.
1636 (_mm256_mask_i64scatter_epi32): Ditto.
1637 (_mm_i64scatter_epi32): Ditto.
1638 (_mm_mask_i64scatter_epi32): Ditto.
1639 (_mm256_i64scatter_epi64): Ditto.
1640 (_mm256_mask_i64scatter_epi64): Ditto.
1641 (_mm_i64scatter_epi64): Ditto.
1642 (_mm_mask_i64scatter_epi64): Ditto.
1643 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1644 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1645 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1646 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1647 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1648 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1649 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1650 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1651 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1652 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1653 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1654 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1655 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1656 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1657 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1658 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1659 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1660 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1661 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1662 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1663 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1664 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1665 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1666 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1667 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1668 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1669 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1670 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1671 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1672 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1673 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1674 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1675 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1676 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1677 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1678 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1679 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1680 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1681 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1682 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1683 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1684 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1685 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1686 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1687 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1688 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1689 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1690 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1691 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1692 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1693 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1694 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1695 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1696 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1697 definitions accordingly.
1699 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1700 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1703 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1706 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1709 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1710 hard regs for inner regno.
1712 2017-01-17 Martin Liska <mliska@suse.cz>
1715 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1716 assumption and add comment.
1718 2017-01-17 Nathan Sidwell <nathan@acm.org>
1720 * ipa-visibility.c (localize_node): New function, broken out of ...
1721 (function_and_variable_visibility): ... here. Call it.
1723 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1726 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1727 correctly set frequency of oudgoing edge.
1728 (duplicate_thread_path): Fix profile updating.
1730 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1733 * configure.ac: Add GCC_BASE_VER.
1734 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1735 version from BASE-VER file.
1736 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1737 (gcc.o): Depend on $(BASEVER).
1738 * common.opt (dumpfullversion): New option.
1739 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1740 * doc/invoke.texi: Document -dumpfullversion.
1741 * doc/install.texi: Document --with-gcc-major-version-only.
1742 * configure: Regenerated.
1744 2017-01-17 Richard Biener <rguenther@suse.de>
1746 PR tree-optimization/71433
1747 * tree-vrp.c (register_new_assert_for): Merge same asserts
1748 on all incoming edges.
1749 (process_assert_insertions_for): Handle insertions at the
1752 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1754 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1755 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1757 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1760 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1763 2017-01-17 Alan Modra <amodra@gmail.com>
1766 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1767 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1768 symbolic stack limit when pic.
1770 2017-01-16 Martin Sebor <msebor@redhat.com>
1772 PR tree-optimization/78608
1773 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1775 2017-01-16 Jeff Law <law@redhat.com>
1778 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1779 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1780 for several include directories that may be relative to sysroot.
1781 * config/i386/x-mingw32 (gplus_includedir): Define.
1782 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1783 (native_system_includedir): Likewise.
1784 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1785 override if TARGET_SYSTEM_ROOT is defined.
1786 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1788 PR tree-optimization/79090
1789 PR tree-optimization/33562
1790 PR tree-optimization/61912
1791 PR tree-optimization/77485
1792 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1793 and computed trims into the dump file.
1795 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1797 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1799 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1802 * gimplify.c (gimplify_init_constructor): If want_value and
1803 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1807 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1808 sequence. Formatting fixes.
1809 (doloop_optimize): Formatting fixes.
1812 * gcc.c (debug_level_greater_than_spec_func): New function.
1813 (static_spec_functions): Add debug-level-gt spec function.
1814 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1816 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1817 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1818 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1819 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1820 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1821 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1823 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1825 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1826 QImode fixups to general and mask registers only.
1828 2017-01-16 Carl Love <cel@us.ibm.com>
1830 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1831 for built-in functions
1832 vector signed char vec_nabs (vector signed char)
1833 vector signed short vec_nabs (vector signed short)
1834 vector signed int vec_nabs (vector signed int)
1835 vector signed long long vec_nabs (vector signed long long)
1836 vector float vec_nabs (vector float)
1837 vector double vec_nabs (vector double)
1838 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1840 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1841 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1842 * doc/extend.texi: Update the documentation file for the new built-in
1845 2017-01-16 Martin Sebor <msebor@redhat.com>
1847 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1850 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1852 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1853 UNSPEC_VSX__XXSPLTD to require special splat handling.
1855 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1858 * system.h: Poison strndup.
1860 2017-01-16 Alan Modra <amodra@gmail.com>
1863 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1866 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1868 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1870 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1872 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1873 call recog here. Assert that INSN_CODE (insn) is non-negative.
1875 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1878 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1880 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1881 in the currently scheduled RTL fragment.
1883 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1885 PR rtl-optimization/78751
1886 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1889 2017-01-14 Jeff Law <law@redhat.com>
1891 PR tree-optimization/79090
1892 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1893 variable length stores.
1894 (compute_trims): Delete dead assignment to *trim_tail.
1895 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1898 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1900 PR rtl-optimization/78626
1901 PR rtl-optimization/78727
1902 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1903 of a block, and split such blocks after everything else is finished.
1905 2017-01-14 Alan Modra <amodra@gmail.com>
1908 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1909 target legitimate_combined_insn.
1910 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1911 (rs6000_legitimate_combined_insn): New function.
1912 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1914 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1915 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1916 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1918 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1920 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1922 2017-01-13 Jeff Law <law@redhat.com>
1924 PR tree-optimization/33562
1925 PR tree-optimization/61912
1926 PR tree-optimization/77485
1927 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1929 (delete_dead_assignment): Likewise.
1930 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1931 statement to delete_dead_call and delete_dead_assignment.
1933 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1936 * substring-locations.c (format_warning_va): Strengthen case 1 so
1937 that both endpoints of the substring must be within the format
1938 range for just the substring to be printed.
1940 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1942 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1943 * config/i386/i386.c (ix86_target_string): Add missing options
1944 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1945 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1946 flags_other and ix86_target_other to flags2_other. Display unknown
1948 (ix86_valid_target_attribute_inner_p): Add missing options and
1949 reorder options by implied ISAs, as in ix86_target_string.
1951 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1953 * hash-table.h (hash_table::too_empty_p): New function.
1954 (hash_table::expand): Use it.
1955 (hash_table::traverse): Likewise.
1956 (hash_table::empty_slot): Use sizeof (value_type) instead of
1957 sizeof (PTR) to convert bytes to elements. Shrink the table
1958 if the current size is excessive for the current number of
1961 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1963 * ira-costs.c (record_reg_classes): Break from the inner loop
1964 early once alt_fail is known to be true. Update outer loop
1965 handling accordingly.
1967 2017-01-13 Jeff Law <law@redhat.com>
1969 * tree-ssa-dse.c (decrement_count): New function.
1970 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1971 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1972 when we know the partially dead statement is a mem* function.
1974 PR tree-optimization/61912
1975 PR tree-optimization/77485
1976 * tree-ssa-dse.c: Include expr.h.
1977 (maybe_trim_constructor_store): New function.
1978 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1980 PR tree-optimization/33562
1981 PR tree-optimization/61912
1982 PR tree-optimization/77485
1983 * doc/invoke.texi: Document new dse-max-object-size param.
1984 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1985 * tree-ssa-dse.c: Include params.h.
1986 (dse_store_status): New enum.
1987 (initialize_ao_ref_for_dse): New, partially extracted from
1989 (valid_ao_ref_for_dse, normalize_ref): New.
1990 (setup_live_bytes_from_ref, compute_trims): Likewise.
1991 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1992 (maybe_trim_partially_dead_store): Likewise.
1993 (maybe_trim_complex_store): Likewise.
1994 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1995 Track what bytes live from the original store. Return tri-state
1996 for dead, partially dead or live.
1997 (dse_dom_walker): Add constructor, destructor and new private members.
1998 (delete_dead_call, delete_dead_assignment): New extracted from
2000 (dse_optimize_stmt): Make a member of dse_dom_walker.
2001 Use initialize_ao_ref_for_dse.
2003 PR tree-optimization/33562
2004 PR tree-optimization/61912
2005 PR tree-optimization/77485
2006 * sbitmap.h (bitmap_count_bits): Prototype.
2007 (bitmap_clear_range, bitmap_set_range): Likewise.
2008 * sbitmap.c (bitmap_clear_range): New function.
2009 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2011 2017-01-13 Martin Liska <mliska@suse.cz>
2014 * function.c (set_cfun): Add new argument force.
2015 * function.h (set_cfun): Likewise.
2016 * ipa-inline-transform.c (inline_call): Use the function when
2017 strict alising from is dropped for function we inline to.
2019 2017-01-13 Richard Biener <rguenther@suse.de>
2021 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2022 for dumping GIMPLE INTEGER_CSTs.
2024 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2026 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2027 to 201112L since C++17.
2029 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2032 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2033 if -fsanitize=kernel-address is present.
2035 2017-01-13 Richard Biener <rguenther@suse.de>
2037 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2038 as _Literal ( type ) number in case usual suffixes do not
2039 preserve all information.
2041 2017-01-13 Richard Biener <rguenther@suse.de>
2043 PR tree-optimization/77283
2044 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2045 and ssa-iterators.h.
2046 (is_feasible_trace): Implement a cost model based on joiner
2049 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2052 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2053 char or short to __float128/_Float128 directly.
2055 2017-01-12 Martin Sebor <msebor@redhat.com>
2057 to -Wformat-overflow.
2058 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2059 (min_bytes_remaining): Same.
2060 (get_string_length): Same.
2061 (format_string): Same.
2062 (format_directive): Same.
2064 (pass_sprintf_length::handle_gimple_call): Same.
2066 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2068 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2069 info.nowrite calls with no lhs that can't throw. Return bool
2070 whether gsi_remove has been called or not.
2071 (pass_sprintf_length::handle_gimple_call): Return bool whether
2072 try_substitute_return_value called gsi_remove. Formatting fix.
2073 (pass_sprintf_length::execute): Don't use gsi_remove if
2074 handle_gimple_call returned true.
2077 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2078 be removed due to side-effects, don't remove following barrier nor
2079 turn the successor edge into fallthru edge.
2081 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2084 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2085 element-reversing loads and stores as not swappable.
2087 2017-01-12 Nathan Sidwell <nathan@acm.org>
2088 Nicolai Stange <nicstange@gmail.com>
2090 * combine.c (try_combine): Don't ignore result of overlap checking
2091 loop. Combine overlap & asm check into single loop.
2093 2017-01-12 Richard Biener <rguenther@suse.de>
2095 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2096 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2098 2017-01-12 Richard Biener <rguenther@suse.de>
2100 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2101 and TS_TARGET_OPTION directly derive from TS_BASE.
2102 * tree-core.h (tree_optimization_option): Derive from tree_base.
2103 (tree_target_option): Likewise.
2105 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2107 * config/i386/i386.c (memory_address_length): Increase len
2108 only when rip_relative_addr_p returns false.
2110 2017-01-11 Julia Koval <julia.koval@intel.com>
2112 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2113 (OPTION_MASK_ISA_SGX_SET): New.
2114 (ix86_handle_option): Handle OPT_msgx.
2115 * config.gcc: Added sgxintrin.h.
2116 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2117 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2118 * config/i386/i386.c (ix86_target_string): Add -msgx.
2120 (ix86_option_override_internal): Handle new options.
2121 (ix86_valid_target_attribute_inner_p): Add sgx.
2122 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2123 * config/i386/i386.opt: Add msgx.
2124 * config/i386/sgxintrin.h: New file.
2125 * config/i386/x86intrin.h: Add sgxintrin.h.
2127 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2130 * fold-const.c (maybe_nonzero_address): Return 1 for function
2132 (tree_single_nonzero_warnv_p): Don't handle function local objects
2136 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2139 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2142 * opts.c: Include "spellcheck.h"
2143 (struct string_fragment): New struct.
2144 (struct edit_distance_traits<const string_fragment &>): New
2146 (get_closest_sanitizer_option): New function.
2147 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2149 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2151 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2153 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2154 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2155 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2156 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2157 for initial die_offset if dwarf_split_debug_info.
2158 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2159 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2161 (output_skeleton_debug_sections): Formatting fix. Use
2162 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2163 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2165 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2167 * config/arm/cortex-a53.md: Add bypasses for
2169 (cortex_a53_r2f): Only use for transfers.
2170 (cortex_a53_f2r): Likewise.
2171 (cortex_a53_r2f_cvt): Add reservation for conversions.
2172 (cortex_a53_f2r_cvt): Likewise.
2174 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2176 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2177 to all inlined functions, change static to extern.
2179 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2182 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2184 (arm_assemble_integer): Likewise.
2186 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2188 * config.gcc: Use new awk script to check CPU, FPU and architecture
2189 parameters for --with-... options.
2190 * config/arm/parsecpu.awk: New file
2191 * config/arm/arm-cpus.in: New file.
2192 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2194 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2196 * config/arm/t-arm: Update dependency rules.
2197 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2198 of processing .def files.
2199 * config/arm/genopt.sh: Deleted.
2200 * config/arm/gentune.sh: Deleted.
2201 * config/arm/arm-cores.def: Deleted.
2202 * config/arm/arm-arches.def: Deleted.
2203 * config/arm/arm-fpus.def: Deleted.
2204 * config/arm/arm-tune.md: Regenerated.
2205 * config/arm/arm-tables.opt: Regenerated.
2206 * config/arm/arm-cpu.h: New generated file.
2207 * config/arm/arm-cpu-data.h: New generated file.
2208 * config/arm/arm-cpu-cdata.h: New generated file.
2210 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2213 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2215 (input_varpool_node): Unpack dynamically_initialized bit.
2217 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2219 PR rtl-optimization/79032
2220 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2221 the alignment of the adjusted memory reference against that of MODE,
2222 instead of the alignment of the original memory reference.
2224 2017-01-11 Martin Jambor <mjambor@suse.cz>
2226 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2228 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2229 decorated functions.
2231 2017-01-11 Richard Biener <rguenther@suse.de>
2233 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2234 set range/nonnull info for PHI results. Do not set it on
2235 stmts marked for removal.
2237 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2239 * expr.c (store_field): In the bitfield case, fetch the return value
2240 from the registers before applying a single big-endian adjustment.
2241 Always do a final load for a BLKmode value not larger than a word.
2243 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2246 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2247 that we correctly handle column numbers greater than
2248 LINE_MAP_MAX_COLUMN_NUMBER.
2250 2017-01-10 Martin Sebor <msebor@redhat.com>
2253 * gimple-ssa-sprintf.c (get_destination_size): Call
2254 {init,fini}object_sizes.
2255 * tree-object-size.c (addr_object_size): Adjust.
2256 (pass_through_call): Adjust.
2257 (pass_object_sizes::execute): Adjust.
2258 * tree-object-size.h (fini_object_sizes): Declare.
2260 2017-01-10 Martin Sebor <msebor@redhat.com>
2262 PR tree-optimization/78775
2263 * builtins.c (get_size_range): Move...
2264 * calls.c: ...to here.
2265 (alloc_max_size): Accept zero argument.
2266 (operand_signed_p): Remove.
2267 (maybe_warn_alloc_args_overflow): Call get_size_range.
2268 * calls.h (get_size_range): Declare.
2270 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2272 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2273 from TI's devices.csv file as of September 2016.
2274 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2276 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2278 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2279 * doc/invoke.texi: Likewise.
2280 * doc/md.texi: Likewise.
2281 * doc/objc.texi: Likewise.
2283 2017-01-10 Joshua Conner <joshconner@google.com>
2285 * config/arm/fuchsia-elf.h: New file.
2286 * config/fuchsia.h: New file.
2287 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2288 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2290 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2292 2016-01-10 Richard Biener <rguenther@suse.de>
2294 PR tree-optimization/79034
2295 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2296 Propagate out degenerate PHIs in the joiner.
2298 2017-01-10 Martin Liska <mliska@suse.cz>
2300 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2301 (sort_congruence_classes_by_decl_uid): Likewise.
2302 (sort_congruence_class_groups_by_decl_uid): Likewise.
2303 (sem_item_optimizer::merge_classes): Sort class, groups in these
2304 classes and members in the groups by DECL_UID of declarations.
2305 This would make merge operations stable.
2307 2017-01-10 Martin Liska <mliska@suse.cz>
2309 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2310 usage of m_classes_vec.
2311 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2312 (sem_item_optimizer::get_group_by_hash): Likewise.
2313 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2314 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2315 (sem_item_optimizer::verify_classes): Likewise.
2316 (sem_item_optimizer::process_cong_reduction): Likewise.
2317 (sem_item_optimizer::dump_cong_classes): Likewise.
2318 (sem_item_optimizer::merge_classes): Likewise.
2319 * ipa-icf.h (congruence_class_hash): Rename from
2320 congruence_class_group_hash. Remove declaration of m_classes_vec.
2322 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2324 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2325 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2326 * config.gcc: Add avx512vpopcntdqintrin.h.
2327 * config/i386/avx512vpopcntdqintrin.h: New.
2328 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2329 * config/i386/i386-builtin-types.def: Add new types.
2330 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2331 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2332 __builtin_ia32_vpopcountq_v8di_mask): New.
2333 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2334 __AVX512VPOPCNTDQ__.
2335 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2336 (PTA_AVX512VPOPCNTDQ): Define.
2337 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2338 TARGET_AVX512VPOPCNTDQ_P): Define.
2339 * config/i386/i386.opt: Add mavx512vpopcntdq.
2340 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2341 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2343 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2346 * predict.def (PRED_CALL): Set to 67.
2348 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2350 * expr.c (store_field): In the bitfield case, if the value comes from
2351 a function call and is of an aggregate type returned in registers, do
2352 not modify the field mode; extract the value in all cases if the mode
2353 is BLKmode and the size is not larger than a word.
2355 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2358 * config/i386/cpuid.h: Fix undefined behavior.
2360 2017-01-04 Jeff Law <law@redhat.com>
2362 PR tree-optimization/79007
2363 PR tree-optimization/67955
2364 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2365 conservative for pt.null when flag_non_call_exceptions is on.
2367 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2369 PR translation/79019
2370 PR translation/79020
2371 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2372 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2373 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2375 * config/avr/avr.opt (maccumulate-args): Likewise.
2376 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2377 * common.opt (freport-bug): Likewise.
2378 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2379 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2380 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2381 translatable string.
2382 * config/i386/i386.c (function_value_32): Likewise.
2383 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2384 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2386 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2387 * common/config/msp430/msp430-common.c (msp430_handle_option):
2389 * symtab.c (symtab_node::verify_base): Likewise.
2390 * opts.c (set_debug_level): Likewise.
2391 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2392 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2393 missing whitespace to translatable strings.
2394 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2395 * config/sh/superh.h: Likewise.
2396 * config/i386/xopintrin.h: Likewise.
2397 * config/i386/znver1.md: Likewise.
2398 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2399 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2400 * double-int.h (struct double_int): Likewise.
2401 * double-int.c (div_and_round_double): Likewise.
2402 * wide-int.cc: Likewise.
2403 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2404 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2405 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2406 (crossjumps_occurred): ... this.
2407 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2410 PR tree-optimization/78899
2411 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2412 returning bool return struct loop *, NULL for failure and the new
2414 (versionable_outer_loop_p): Don't version outer loop if it has
2415 dont_vectorized bit set.
2416 (tree_if_conversion): When versioning outer loop, ensure
2417 tree_if_conversion is performed also on the inner loop of the
2418 non-vectorizable outer loop copy.
2419 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2420 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2421 prevent vectorization of it.
2422 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2423 the outer loop vectorization of the non-scalar version is attempted
2424 before vectorization of the inner loop in scalar version. If
2425 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2426 vectorization of its inner loop.
2427 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2428 has 2 inner loops, rename also on edges from bb whose single pred
2429 is outer_loop->header. Fix typo in function comment.
2431 2017-01-09 Martin Sebor <msebor@redhat.com>
2434 * asan.c (asan_emit_stack_protection): Increase local buffer size
2435 to avoid snprintf truncation warning.
2437 2017-01-09 Andrew Pinski <apinski@cavium.com>
2439 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2440 to reference thunderx2t99 for the tuning structure
2441 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2443 (thunderx2t99_extra_costs): This.
2444 * config/aarch64/aarch64-tune.md: Regenerate.
2445 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2446 (vulcan_addrcost_table): This.
2447 (vulcan_regmove_cost): Rename to ...
2448 (thunderx2t99_regmove_cost): This.
2449 (vulcan_vector_cost): Rename to ...
2450 (thunderx2t99_vector_cost): this.
2451 (vulcan_branch_cost): Rename to ...
2452 (thunderx2t99_branch_cost): This.
2453 (vulcan_tunings): Rename to ...
2454 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2455 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2457 2017-01-09 Martin Jambor <mjambor@suse.cz>
2461 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2462 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2463 (propagate_vr_accross_jump_function): Use the above function for all
2464 value range computations for pass-through jump functions and type
2465 converasion from explicit value range values.
2466 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2467 parameters from TYPE_ARG_TYPES.
2468 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2469 (ipa_write_node_info): Stream type of the actual argument.
2470 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2472 2017-01-09 Martin Liska <mliska@suse.cz>
2475 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2476 (lookup_compiler): Do not show error message with have_E.
2478 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2480 PR tree-optimization/78938
2481 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2482 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2483 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2484 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2487 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2489 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2492 2017-01-09 Richard Biener <rguenther@suse.de>
2494 PR tree-optimization/78997
2495 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2496 name condition properly.
2498 2017-01-09 Richard Biener <rguenther@suse.de>
2501 * dwarf2out.c (is_cxx): New overload with context.
2502 (is_naming_typedef_decl): Use it.
2504 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2506 * invoke.texi (Option Summary): Correct spacing in option lists
2507 and add line breaks to fix over-long lines.
2509 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2513 * extend.texi (Common Variable Attributes): Add xref to GCC
2514 Internals manual to explain mode attribute keywords.
2516 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2519 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2520 and Preprocessor Options.
2521 (Options for Linking): Document -pthread here....
2522 (RS/6000 and PowerPC Options): ...not here.
2523 (Solaris 2 Options): ...or here.
2524 * doc/cppopts.texi: Document -pthread.
2526 2017-01-08 Martin Sebor <msebor@redhat.com>
2529 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2530 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2531 New member functions.
2532 (format_directive): Used them.
2534 (pass_sprintf_length::handle_gimple_call): Same.
2535 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2536 to avoid truncation for any argument.
2537 (extract_affine_mul): Same.
2538 * tree.c (get_file_function_name): Same.
2540 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2543 * predict.def (PRED_INDIR_CALL): Set to 86.
2545 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2547 PR preprocessor/54124
2548 * doc/cppopts.texi: Reformat -d subtable to list the full name
2549 of the options. Add cross-reference to the docs for the general
2550 compiler -d options.
2551 * doc/invoke.texi (Developer Options): Add cross-reference to the
2552 preprocessor-specific -d option documentation.
2554 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2556 PR preprocessor/13498
2557 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2558 redudant material, and reflect new command-line options.
2559 (System Headers): Likewise.
2561 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2563 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2564 -isystem, and -idirafter. Copy-edit.
2565 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2566 default for -ftrack-macro-expansion. Delete obsolete and
2567 badly-formatted implementation details about -fdebug-cpp output.
2568 * doc/cppwarnopts.texi: Copy-edit.
2570 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2573 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2574 that the transition from a max line width >= 1<<10 to narrower
2575 lines works correctly.
2577 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2579 * doc/options.texi (PerFunction): New.
2580 * opt-functions.awk (switch_flags): Map both Optimization and
2581 PerFunction to CL_OPTIMIZATION.
2582 * opth-gen.awk: Test for PerFunction flag along with
2584 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2585 it only when the latter is present. Skip those that don't in
2586 the hash function generator.
2587 * common.opt (fvar-tracking): Mark as PerFunction instead of
2589 (fvar-tracking-assignments): Likewise.
2590 (fvar-tracking-assignments-toggle): Likewise.
2591 (fvar-tracking-uninit): Likewise.
2593 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2595 PR translation/79018
2596 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2599 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2602 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2603 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2604 TARGET_LONG_JUMP_TABLE_OFFSETS.
2605 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2606 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2607 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2608 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2609 * config/m68k/m68k.md (tablejump expander): Likewise.
2610 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2611 TARGET_LONG_JUMP_TABLE_OFFSETS.
2612 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2613 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2615 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2616 David Holsgrove <david.holsgrove@xilinx.com>
2618 * common/config/microblaze/microblaze-common.c
2619 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2620 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2622 * config/microblaze/microblaze.c (microblaze_must_save_register)
2623 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2625 (microblaze_eh_return): New function.
2626 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2627 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2628 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2629 * config/microblaze/microblaze.md (eh_return): New pattern.
2631 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2633 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2634 GCC_DIAGNOSTIC_STRINGIFY): Define.
2636 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2638 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2640 * config/arm/arm.md (<mcrr>): New.
2642 * config/arm/arm.c (arm_arch5te): New.
2643 (arm_option_override): Set arm_arch5te.
2644 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2646 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2647 (arm_mcrr_qualifiers): ... this. New.
2648 (MRRC_QUALIFIERS): Define to...
2649 (arm_mrrc_qualifiers): ... this. New.
2650 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2652 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2653 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2654 (MRRCI, mrrc, MRRC): New.
2655 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2656 VUNSPEC_MRRC2): New.
2658 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2660 * config/arm/arm.md (<mcr>): New.
2662 * config/arm/arm.c (arm_coproc_builtin_available): Add
2663 support for mcr, mrc, mcr2 and mrc2.
2664 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2665 (arm_mcr_qualifiers): ... this. New.
2666 (MRC_QUALIFIERS): Define to ...
2667 (arm_mrc_qualifiers): ... this. New.
2668 (MCR_QUALIFIERS): Define to ...
2669 (arm_mcr_qualifiers): ... this. New.
2670 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2672 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2673 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2674 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2677 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2679 * config/arm/arm.md (*ldc): New.
2683 * config/arm/arm.c (arm_coproc_builtin_available): Add
2684 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2685 (arm_coproc_ldc_stc_legitimate_address): New.
2686 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2687 'qualifier_const_pointer'.
2688 (LDC_QUALIFIERS): Define to...
2689 (arm_ldc_qualifiers): ... this. New.
2690 (STC_QUALIFIERS): Define to...
2691 (arm_stc_qualifiers): ... this. New.
2692 * config/arm/arm-protos.h
2693 (arm_coproc_ldc_stc_legitimate_address): New.
2694 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2695 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2696 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2697 stc2, stcl, stc2l): New.
2698 * config/arm/constraints.md (Uz): New.
2699 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2700 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2701 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2702 VUNSPEC_STC2L): New.
2704 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2706 * config/arm/arm.md (<cdp>): New.
2707 * config/arm/arm.c (neon_const_bounds): Rename this ...
2708 (arm_const_bounds): ... this.
2709 (arm_coproc_builtin_available): New.
2710 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2711 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2712 (CDP_QUALIFIERS): Define to...
2713 (arm_cdp_qualifiers): ... this. New.
2715 (arm_expand_builtin_args): Add case for 6 arguments.
2716 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2717 (arm_const_bounds): ... this.
2718 (arm_coproc_builtin_available): New.
2719 * config/arm/arm_acle.h (__arm_cdp): New.
2721 * config/arm/arm_acle_builtins.def (cdp): New.
2723 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2724 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2726 * config/arm/types.md (coproc): New.
2727 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2728 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2729 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2730 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2732 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2734 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2735 (UBINOP_QUALIFIERS): New.
2737 (acle_builtin_data): New. Change comment.
2738 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2739 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2740 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2741 arm_acle_builtins.def.
2742 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2743 (arm_init_acle_builtins): New.
2744 (CRC32_BUILTIN): Remove.
2745 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2746 crc32cb, crc32ch and crc32cw.
2747 (arm_init_crc32_builtins): Remove.
2748 (arm_init_builtins): Use arm_init_acle_builtins rather
2749 than arm_init_crc32_builtins.
2750 (arm_expand_acle_builtin): New.
2751 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2752 * config/arm/arm_acle_builtins.def: New.
2754 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2756 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2757 (arm_builtin_datum): ... this.
2758 (arm_init_neon_builtin): Rename to ...
2759 (arm_init_builtin): ... this. Add a new parameters PREFIX
2760 and USE_SIG_IN_NAME.
2761 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2762 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2763 'arm_builtin_datum'.
2764 (arm_init_vfp_builtins): Likewise.
2765 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2766 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2767 (arm_expand_neon_args): Rename to ...
2768 (arm_expand_builtin_args): ... this. Rename builtin_arg
2769 enum values and differentiate between ARG_BUILTIN_MEMORY
2770 and ARG_BUILTIN_NEON_MEMORY.
2771 (arm_expand_neon_builtin_1): Rename to ...
2772 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2773 values, arm_expand_builtin_args and add bool parameter NEON.
2774 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2775 (arm_expand_vfp_builtin): Likewise.
2776 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2778 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2781 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2782 * predict.c (tree_estimate_probability_bb): Reverse direction of
2783 polymorphic call predictor.
2785 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2787 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2788 (determine_pass_name_match): ...this new function and...
2789 (should_skip_pass_p): ...this new function.
2791 2017-01-06 Nathan Sidwell <nathan@acm.org>
2793 * ipa-visibility.c (function_and_variable_visibility): Reformat
2794 comments and long lines. Remove extrneous if.
2795 * symtab.c (symtab_node::make_decl_local): Fix code format.
2796 (symtab_node::set_section_for_node): Fix comment typo.
2798 2017-01-06 Martin Liska <mliska@suse.cz>
2801 * lra-constraints.c: Rename invariant to lra_invariant.
2802 * predict.c (set_even_probabilities): Initialize e to NULL.
2804 2017-01-05 Martin Sebor <msebor@redhat.com>
2806 PR tree-optimization/78910
2807 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2808 (format_integer): Correct off-by-one error in the handling
2809 of precision with negative numbers in signed conversions..
2811 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2813 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2815 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2817 PR tree-optimization/71016
2818 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2819 factor_out_conditional_conversion. Formatting fix.
2820 (factor_out_conditional_conversion): Add cond_stmt argument.
2821 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2822 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2825 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2827 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2828 read-rtl-function.o, and selftest-rtl.o.
2829 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2830 (selftest::aarch64_test_loading_full_dump): New function.
2831 (selftest::aarch64_run_selftests): New function.
2832 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2833 selftest::aarch64_run_selftests.
2834 * config/i386/i386.c
2835 (selftest::ix86_test_loading_dump_fragment_1): New function.
2836 (selftest::ix86_test_loading_call_insn): New function.
2837 (selftest::ix86_test_loading_full_dump): New function.
2838 (selftest::ix86_test_loading_unspec): New function.
2839 (selftest::ix86_run_selftests): Call the new functions.
2840 * emit-rtl.c (maybe_set_max_label_num): New function.
2841 * emit-rtl.h (maybe_set_max_label_num): New decl.
2842 * function.c (instantiate_decls): Guard call to
2843 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2844 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2846 * gensupport.c (gen_reader::gen_reader): Pass "false"
2847 for new "compact" param of rtx_reader.
2848 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2849 rather than an empty string for NULL strings.
2850 * read-md.c: Potentially include config.h rather than bconfig.h.
2851 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2852 (have_error): New global, copied from errors.c.
2853 (md_reader::read_name): Rename to...
2854 (md_reader::read_name_1): ...this, adding "out_loc" param,
2855 and converting "missing name or number" to returning false, rather
2857 (md_reader::read_name): Reimplement in terms of read_name_1.
2858 (md_reader::read_name_or_nil): New function.
2859 (md_reader::read_string): Handle "(nil)" by returning NULL.
2860 (md_reader::md_reader): Add new param "compact".
2861 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2862 (md_reader::read_file): New method.
2863 * read-md.h (md_reader::md_reader): Add new param "compact".
2864 (md_reader::read_file): New method.
2865 (md_reader::is_compact): New accessor.
2866 (md_reader::read_name): Convert return type from void to file_location.
2867 (md_reader::read_name_or_nil): New decl.
2868 (md_reader::read_name_1): New decl.
2869 (md_reader::m_compact): New field.
2870 (noop_reader::noop_reader): Pass "false" for new "compact" param
2872 (rtx_reader::rtx_reader): Add new "compact" param.
2873 (rtx_reader::read_rtx_operand): Make virtual and convert return
2874 type from void to rtx.
2875 (rtx_reader::read_until): New decl.
2876 (rtx_reader::handle_any_trailing_information): New virtual function.
2877 (rtx_reader::postprocess): New virtual function.
2878 (rtx_reader::finalize_string): New virtual function.
2879 (rtx_reader::m_in_call_function_usage): New field.
2880 (rtx_reader::m_reuse_rtx_by_id): New field.
2881 * read-rtl-function.c: New file.
2882 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2883 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2884 (selftest::verify_three_block_rtl_cfg): New decl.
2885 * read-rtl-function.h: New file.
2886 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2887 For host, include function.h, memmodel.h, and emit-rtl.h.
2888 (one_time_initialization): New function.
2889 (struct compact_insn_name): New struct.
2890 (compact_insn_names): New array.
2891 (find_code): Handle insn codes in compact dumps.
2892 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2893 (bind_subst_iter_and_attr): Likewise.
2894 (add_condition_to_string): Likewise.
2895 (add_condition_to_rtx): Likewise.
2896 (apply_attribute_uses): Likewise.
2897 (add_current_iterators): Likewise.
2898 (apply_iterators): Likewise.
2899 (initialize_iterators): Guard usage of apply_subst_iterator with
2900 #ifdef GENERATOR_FILE.
2901 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2902 (md_reader::read_mapping): Likewise.
2903 (add_define_attr_for_define_subst): Likewise.
2904 (add_define_subst_attr): Likewise.
2905 (read_subst_mapping): Likewise.
2906 (check_code_iterator): Likewise.
2907 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2909 (one_time_initialization): New function.
2910 (rtx_reader::read_until): New method.
2911 (read_flags): New function.
2912 (parse_reg_note_name): New function.
2913 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2914 Handle reuse_rtx ids.
2915 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2916 Add parsing support for RTL dumps, mirroring the special-cases in
2917 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2918 values, and calling handle_any_trailing_information.
2919 (rtx_reader::read_rtx_operand): Convert return type from void
2920 to rtx, returning return_rtx. Handle case 'e'. Call
2921 finalize_string on XSTR and XTMPL fields.
2922 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2923 "(nil)" values were omitted. Call the postprocess vfunc on the
2925 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2926 class ctor. Initialize m_in_call_function_usage. Call
2927 one_time_initialization.
2928 * rtl-tests.c (selftest::test_uncond_jump): Call
2929 set_new_first_and_last_insn.
2930 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2931 * selftest-rtl.c: New file.
2932 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2933 (selftest::get_insn_by_uid): New decl.
2934 * selftest-run-tests.c (selftest::run_tests): Call
2935 read_rtl_function_c_tests.
2936 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2937 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2940 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2942 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2943 operands in a special way. Assert that pos+len <= mode precision.
2945 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2947 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2948 3 argument Alias with unlimited for the negative form.
2949 (fno-vect-cost-model): Removed.
2951 2017-01-05 Martin Liska <mliska@suse.cz>
2953 * hsa-gen.c (gen_hsa_divmod): New function.
2954 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2956 2017-01-05 Martin Liska <mliska@suse.cz>
2959 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2962 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2964 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2965 small constant length operands.
2967 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2969 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2970 between loop iterations.
2972 2017-01-05 Martin Liska <mliska@suse.cz>
2975 * gimplify.c (gimplify_decl_expr): Compare to
2976 asan_poisoned_variables instread of checking flags.
2977 (gimplify_target_expr): Likewise.
2978 (gimplify_expr): Likewise.
2979 (gimplify_function_tree): Conditionally initialize
2980 asan_poisoned_variables.
2982 2017-01-04 Jeff Law <law@redhat.com>
2984 PR tree-optimizatin/78812
2985 * rtl.h (contains_mem_rtx_p): Prototype.
2986 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2987 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2988 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2989 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2990 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2992 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2994 * input.c (assert_char_at_range): Default-initialize actual_range.
2996 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2998 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2999 to match the caller.
3001 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3003 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3004 insns after final jump in test to emit dummy move.
3006 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3008 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3009 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3011 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3013 * multiple_target.c (create_dispatcher_calls): Init e_next.
3014 * tree-ssa-loop-split.c (split_loop): Init border.
3015 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3018 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3023 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3024 (altivec_register_operand): Do not return true if the operand
3025 contains a SUBREG mixing SImode and SFmode.
3026 (vsx_register_operand): Likewise.
3027 (vsx_reg_sfsubreg_ok): New predicate.
3028 (vfloat_operand): Do not return true if the operand contains a
3029 SUBREG mixing SImode and SFmode.
3030 (vint_operand): Likewise.
3031 (vlogical_operand): Likewise.
3032 (gpc_reg_operand): Likewise.
3033 (int_reg_operand): Likewise.
3034 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3035 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3036 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3038 (rs6000_emit_move_si_sf_subreg): New helper function.
3039 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3040 fixup SUBREGs involving SImode and SFmode.
3041 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3042 numbers for the new peephole2 optimization.
3043 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3044 the GLIBC math library that do AND/IOR/XOR operations on single
3045 precision floating point.
3046 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3047 target macros to say whether we need to avoid SUBREGs mixing
3049 (TARGET_ALLOW_SF_SUBREG): Likewise.
3050 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3051 (UNSPEC_SI_FROM_SF): Likewise.
3052 (iorxor): Change spacing.
3053 (and_ior_xor): New iterator for AND, IOR, and XOR.
3054 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3055 (movdi_from_sf_zero_ext): Likewise.
3056 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3057 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3058 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3059 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3060 (fms<mode>4): Likewise.
3061 (fnma<mode>4): Likewise.
3062 (fnms<mode>4): Likewise.
3063 (nfma<mode>4): Likewise.
3064 (nfms<mode>4): Likewise.
3066 2017-01-04 Marek Polacek <polacek@redhat.com>
3069 * doc/invoke.texi: Document -Wpointer-compare.
3071 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3073 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3076 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3077 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3080 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3083 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3084 documentation of the powerpc_popcntb_ok attribute.
3085 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3086 code to issue warning messages if a requested CPU configuration is
3087 not supported by the binary (assembler and loader) toolchain.
3088 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3089 made to define a built-in function that has been disabled.
3090 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3091 made to define a built-in function that has been disabled.
3092 (altivec_init_builtins): Add comment explaining why definition
3093 of the DST built-in functions is not preceded by an assertion
3094 check. Add assertions to prevent ICE if attempts are made to
3095 define an altivec predicate or an abs* built-in function that has
3097 (htm_init_builtins): Add comment explaining why definition of the
3098 htm built-in functions is not preceded by an assertion check.
3100 2017-01-04 Jeff Law <law@redhat.com>
3102 PR tree-optimizatin/67955
3103 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3104 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3105 the points-to solution does not include pt_null. Use DECL_PT_UID
3108 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3110 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3111 Use gen_int_mode instead of gen_lopwart for const_int operands.
3113 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3115 PR tree-optimization/71563
3116 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3117 out of range value - has low bits known to be zero.
3119 2017-01-04 Alan Modra <amodra@gmail.com>
3121 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3122 * configure: Regenerate.
3123 * config.in: Regenerate.
3125 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3128 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3129 a substring of the message, but strcmp with the whole message. Ifdef
3130 ENABLE_NLS, translate the message first using dgettext.
3132 2017-01-03 Jeff Law <law@redhat.com>
3134 PR tree-optimizatin/78856
3135 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3136 (mark_threaded_blocks): Remove code to truncate thread paths that
3137 cross multiple loop headers. Instead invalidate the cached loop
3138 iteration information and handle case of a thread path walking
3139 into an irreducible region.
3141 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3144 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3145 assertions. Add support for doing the signbit if the IEEE 128-bit
3146 floating point value is in a GPR.
3147 * config/rs6000/rs6000.md (Fsignbit): Delete.
3148 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3149 Update the length attribute if the value is in a GPR.
3150 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3151 the sign or zero extension instruction, since the value is always 0/1.
3152 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3155 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3156 extracting SImode to a GPR register so that we can generate a
3157 store, limit the vector to be in a traditional Altivec register
3158 for the vextuwrx instruction.
3160 2017-01-03 Ian Lance Taylor <iant@google.com>
3162 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3164 2017-01-03 Martin Sebor <msebor@redhat.com>
3166 PR tree-optimization/78696
3167 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3168 precision. Use MPFR for %f for greater fidelity. Correct handling
3170 (pass_sprintf_length::compute_format_length): Set width and precision
3171 specified by asrerisk to void_node for vararg functions.
3172 (try_substitute_return_value): Adjust dump output.
3174 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3176 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3178 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3180 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3181 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3182 -mlra/-mno-lra was passed to the compiler.
3184 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3186 PR rtl-optimization/65618
3187 * emit-rtl.c (try_split): Move initialization of "before" and
3188 "after" to just before the call to emit_insn_after_setloc.
3190 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3192 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3194 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3196 * dwarf2out.c (gen_enumeration_type_die): When
3197 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3199 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3201 PR tree-optimization/78965
3202 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3203 Change first argument from const call_info & to call_info &. For %n
3204 set info.nowrite to false.
3207 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3208 possibly throwing calls.
3210 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3211 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3212 and fns handling, rather than in a separate case SSA_NAME.
3214 2017-01-02 Jeff Law <law@redhat.com>
3216 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3217 fixes for first_period and second_period variables.
3219 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3222 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3223 (*insvqi_1): New insn pattern.
3224 (*insvqi_1_mem_rex64): Ditto.
3226 (*insvqi_3): Rename from *insvqi.
3228 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3230 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3232 * doc/cfg.texi (Edges): Remove reference to Java.
3233 (Maintaining the CFG): Ditto.
3235 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3238 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3239 transparent aliases.
3241 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3244 * predict.def (PRED_CALL): Update hitrate.
3245 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3246 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3247 into direct/indirect/polymorphic variants.
3249 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3251 Update copyright years.
3253 * gcc.c (process_command): Update copyright notice dates.
3254 * gcov-dump.c (print_version): Ditto.
3255 * gcov.c (print_version): Ditto.
3256 * gcov-tool.c (print_version): Ditto.
3257 * gengtype.c (create_file): Ditto.
3258 * doc/cpp.texi: Bump @copying's copyright year.
3259 * doc/cppinternals.texi: Ditto.
3260 * doc/gcc.texi: Ditto.
3261 * doc/gccint.texi: Ditto.
3262 * doc/gcov.texi: Ditto.
3263 * doc/install.texi: Ditto.
3264 * doc/invoke.texi: Ditto.
3266 Copyright (C) 2017 Free Software Foundation, Inc.
3268 Copying and distribution of this file, with or without modification,
3269 are permitted in any medium without royalty provided the copyright
3270 notice and this notice are preserved.