1 2017-02-02 Jakub Jelinek <jakub@redhat.com>
4 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
5 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
6 conditions on a single line.
8 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
10 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
11 __S390_VX__ to __VX__.
13 2017-02-01 Andrew Pinski <apinski@cavium.com>
15 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
16 stmt_info to record_stmt_cost.
17 (vect_get_known_peeling_cost): Pass stmt_info if known to
19 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
20 cpu_vector_cost field into
21 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
22 field into vec_int_stmt_cost and vec_fp_stmt_cost.
23 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
24 splitting of scalar_stmt_cost and vec_stmt_cost.
25 (thunderx_vector_cost): Likewise.
26 (cortexa57_vector_cost): LIkewise.
27 (exynosm1_vector_cost): Likewise.
28 (xgene1_vector_cost): Likewise.
29 (thunderx2t99_vector_cost): Improve after the splitting of the two
31 (aarch64_builtin_vectorization_cost): Update for the splitting of
32 scalar_stmt_cost and vec_stmt_cost.
34 2017-02-01 Torvald Riegel <triegel@redhat.com>
35 Richard Henderson <rth@redhat.com>
37 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
38 conditional on existance of a fast atomic load.
39 * optabs-query.c (can_atomic_load_p): New function.
40 * optabs-query.h (can_atomic_load_p): Declare it.
41 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
42 no fast atomic load is available for the particular size of access.
43 (expand_atomic_compare_and_swap): Likewise.
44 (expand_atomic_load): Likewise.
45 (expand_atomic_store): Likewise.
46 (expand_atomic_fetch_op): Likewise.
47 * testsuite/lib/target-supports.exp
48 (check_effective_target_sync_int_128): Remove x86 because it provides
50 (check_effective_target_sync_int_128_runtime): Likewise.
52 2017-02-01 Richard Biener <rguenther@suse.de>
54 * graphite.c: Include tree-vectorizer.h for find_loop_location.
55 (graphite_transform_loops): Provide opt-info for optimized nests.
56 * tree-parloop.c (parallelize_loops): Provide opt-info for
59 2017-02-01 Richard Biener <rguenther@suse.de>
62 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
65 2017-02-01 Richard Biener <rguenther@suse.de>
67 PR tree-optimization/71824
68 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
69 Verify the loops are valid in the merged SESE region.
70 (scop_detection::can_represent_loop_1): Check analyzing the
71 evolution of the number of iterations in the region succeeds.
73 2017-01-31 Ian Lance Taylor <iant@golang.org>
75 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
76 REG_ARGS_SIZE note to 32-bit push insns and call insn.
78 2017-01-31 David Malcolm <dmalcolm@redhat.com>
81 * input.c (get_substring_ranges_for_loc): Replace line_width
82 assertion with error-handling.
84 2017-01-31 Richard Biener <rguenther@suse.de>
86 PR tree-optimization/77318
87 * graphite-sese-to-poly.c (extract_affine): Fix assert.
88 (create_pw_aff_from_tree): Take loop parameter.
89 (add_condition_to_pbb): Pass loop of the condition to
90 create_pw_aff_from_tree.
92 2017-01-31 Jakub Jelinek <jakub@redhat.com>
94 * config/s390/s390.c (s390_asan_shadow_offset): New function.
95 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
97 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
101 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
103 (convert_int_to_float128): Likewise.
104 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
105 (convert_int_to_float128): Likewise.
106 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
107 (UNSPEC_IEEE128_CONVERT): Likewise.
108 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
109 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
110 Use local variables for IBM extended format.
111 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
112 (fix_trunc<mode>si2_fprs): Likewise.
113 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
114 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
115 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
116 to know that we can now have integers of all sizes in vector
118 (fix<uns>_<mode>di2_hw): Likewise.
119 (float<uns>_<mode>si2_hw): Likewise.
120 (fix_<mode>si2_hw): Likewise.
121 (fixuns_<mode>si2_hw): Likewise.
122 (float<uns>_<mode>di2_hw): Likewise.
123 (float_<mode>di2_hw): Likewise.
124 (float_<mode>si2_hw): Likewise.
125 (floatuns_<mode>di2_hw): Likewise.
126 (floatuns_<mode>si2_hw): Likewise.
127 (xscvqp<su>wz_<mode>): Delete, no longer used.
128 (xscvqp<su>dz_<mode>): Likewise.
129 (xscv<su>dqp_<mode>): Likewise.
130 (ieee128_mfvsrd_64bit): Likewise.
131 (ieee128_mfvsrd_32bit): Likewise.
132 (ieee128_mfvsrwz): Likewise.
133 (ieee128_mtvsrw): Likewise.
134 (ieee128_mtvsrd_64bit): Likewise.
135 (ieee128_mtvsrd_32bit): Likewise.
137 2017-01-31 Martin Liska <mliska@suse.cz>
140 * ipa-prop.c (ipa_free_all_node_params): Call release method
141 instead of ~sumbol_summary to not to trigger double times
144 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
146 PR tree-optimization/71691
147 * bitmap.h (class auto_bitmap): New.
148 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
149 is_maybe_undefined instead of ssa_undefined_value_p.
151 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
153 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
154 __S390_ARCH_LEVEL__ to __ARCH__.
156 2017-01-31 Jakub Jelinek <jakub@redhat.com>
158 PR tree-optimization/79267
159 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
160 if should_remove_lhs_p is true.
162 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
165 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
166 (add_alignment_attribute): New.
167 (base_type_die): Add alignment attribute.
168 (subrange_type_die): Likewise.
169 (modified_type_die): Likewise.
170 (gen_array_type_die): Likewise.
171 (gen_descr_array_type_die: Likewise.
172 (gen_enumeration_type_die): Likewise.
173 (gen_subprogram_die): Likewise.
174 (gen_variable_die): Likewise.
175 (gen_field_die): Likewise.
176 (gen_ptr_to_mbr_type_die): Likewise.
177 (gen_struct_or_union_type_die): Likewise.
178 (gen_subroutine_type_die): Likewise.
179 (gen_typedef_die): Likewise.
180 (base_type_cmp): Compare alignment attribute.
182 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
185 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
186 (setb_unsigned) New pattern for setb with CCUNS.
187 * config/rs6000/rs6000.c (expand_block_compare): Use a different
188 subfc./subfe sequence to avoid overflow problems. Generate a
189 shorter sequence with cmpld/setb for power9.
190 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
191 for generating subfc. instruction.
192 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
193 now uses this instruction.
195 2017-01-30 Ian Lance Taylor <iant@google.com>
198 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
199 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
201 2017-01-30 Martin Sebor <msebor@redhat.com>
203 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
204 Move constant to the right of a relational operator.
205 (get_mpfr_format_length, format_character, format_string): Ditto.
206 (should_warn_p, maybe_warn): Same.
208 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
210 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
213 * asan.c (get_translation_unit_decl): Remove function.
214 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
216 2017-01-30 Martin Liska <mliska@suse.cz>
218 PR gcov-profile/79259
219 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
222 2017-01-30 Martin Liska <mliska@suse.cz>
225 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
226 Initialize variables with NULL value.
228 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
231 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
233 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
235 2017-01-30 Richard Biener <rguenther@suse.de>
237 PR tree-optimization/79276
238 * tree-vrp.c (process_assert_insertions): Properly adjust common
239 when removing a duplicate.
241 * gcc.dg/torture/pr79276.c: New testcase.
243 2017-01-30 Richard Biener <rguenther@suse.de>
245 PR tree-optimization/79256
246 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
247 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
249 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
251 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
254 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
255 ("*r<noxa>sbg_<mode>_sll_bitmask")
256 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
257 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
258 Use contiguous_bitmask_nowrap_operand.
260 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
263 * config/rs6000/altivec.h (vec_xl): Revise #define.
266 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
268 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
270 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
272 PR rtl-optimization/79194
273 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
274 traps before call to bypass_conditional_jumps.
276 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
278 PR tree-optimization/71374
279 * lra-constraints.c (check_conflict_input_operands): New.
280 (match_reload): Use it.
282 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
285 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
286 account to calculate conflict_set.
288 2017-01-27 Bin Cheng <bin.cheng@arm.com>
290 PR rtl-optimization/78559
291 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
292 other_insn in combine.
294 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
296 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
297 uint16_type_node for BT_UINT16.
299 2017-01-27 David Malcolm <dmalcolm@redhat.com>
301 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
303 (GIMPLE Tests): New node.
304 (RTL Tests): New node.
306 2017-01-27 Richard Biener <rguenther@suse.de>
308 PR tree-optimization/79245
309 * tree-loop-distribution.c (distribute_loop): Apply cost
310 modeling also to detected patterns.
312 2017-01-27 Richard Biener <rguenther@suse.de>
314 PR tree-optimization/71433
315 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
316 (compare_assert_loc): New function.
317 (process_assert_insertions): Sort and optimize assert locations
318 to remove duplicates and push down identical assertions on
319 edges to their destination block.
321 2017-01-27 Richard Biener <rguenther@suse.de>
323 PR tree-optimization/79244
324 * tree-vrp.c (remove_range_assertions): Forcefully propagate
325 out SSA names even if abnormal.
327 2017-01-27 Jakub Jelinek <jakub@redhat.com>
329 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
330 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
331 instead of MPFR_RNDN.
333 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
336 * arm.c (arm_option_override): Don't call build_target_option_node
337 until after doing all option overrides.
338 (arm_valid_target_attribute_tree): Likewise.
340 2017-01-27 Martin Liska <mliska@suse.cz>
342 * doc/invoke.texi (-fprofile-arcs): Document profiling support
343 for {cd}tors and C++ {cd}tors.
345 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
347 * config/s390/s390.md ("*setmem_long_and")
348 ("*setmem_long_and_31z"): Use zero_extend instead of and.
350 2017-01-26 Martin Sebor <msebor@redhat.com>
352 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
355 2017-01-26 Martin Sebor <msebor@redhat.com>
357 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
358 HAVE_DFmode before using XFmode or DFmode.
359 (parse_directive): Avoid using the z length modifier to avoid
360 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
363 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
364 to accept adjustment as an array.
365 (get_int_range): New function.
366 (struct directive): Make width and prec arrays.
367 (directive::set_width, directive::set_precision): Call get_int_range.
368 (format_integer, format_floating): Handle width and precision ranges.
369 (format_string, parse_directive): Same.
371 2017-01-26 Jakub Jelinek <jakub@redhat.com>
374 * dwarf2out.c (generate_skeleton_bottom_up): For children with
375 comdat_type_p set, just clone them, but keep the children in the
379 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
380 which have direct callers with -fvar-tracking-assignments enabled
382 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
383 inside of type units.
385 2017-01-26 Martin Sebor <msebor@redhat.com>
388 * gimple-ssa-sprintf.c (struct result_range): Add likely and
390 (struct format_result): Replace number_chars, number_chars_min,
391 and number_chars_max with a single member of struct result_range.
393 (format_result::operator+=): Adjust.
394 (struct fmtresult): Remove bounded. Handle likely and unlikely
396 (fmtresult::adjust_for_width_or_precision): New function.
397 (fmtresult:type_max_digits): New function.
398 (bytes_remaining): Handle likely and unlikely counters.
399 (min_bytes_remaining): Remove.
400 (format_percent): Simplify.
401 (format_integer, format_floating): Set likely and unlikely counters.
402 (get_string_length, format_character, format_string): Same.
403 (format_plain, should_warn_p): New function.
404 (maybe_warn): Call should_warn_p. Update diagnostic messages
405 and handle those for all directives, including plain strings.
406 (format_directive): Handle likely and unlikely counters.
407 Remove unnecessary quoting from diagnostics. Add an informational
410 (pass_sprintf_length::compute_format_length): Simplify.
411 (try_substitute_return_value): Handle likely and unlikely counters.
413 2017-01-26 Carl Love <cel@us.ibm.com>
415 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
416 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
418 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
421 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
422 endianess for subregs into account.
423 * lra-constraints.c (lra_constraints): Do risky transformations
424 always on the first iteration.
425 * lra-lives.c (check_pseudos_live_through_calls): Add arg
426 last_call_used_reg_set.
427 (process_bb_lives): Define and use last_call_used_reg_set.
428 * lra.c (lra): Always continue after lra_constraints on the first
431 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
433 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
435 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
437 2017-01-26 Jakub Jelinek <jakub@redhat.com>
439 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
440 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
441 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
442 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
443 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
444 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
445 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
446 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
447 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
449 2017-01-26 Marek Polacek <polacek@redhat.com>
452 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
453 for the third operand.
455 2017-01-26 Jakub Jelinek <jakub@redhat.com>
458 * omp-low.c (struct omp_context): Add simt_stmt field.
459 (scan_omp_for): Return omp_context *.
460 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
461 context to the _simt_ SIMD stmt.
462 (lower_omp_for): For combined SIMD with sibling _simt_
463 SIMD, make sure to use the same decls in _looptemp_
464 clauses as in the sibling.
466 2017-01-26 David Sherwood <david.sherwood@arm.com>
469 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
472 2017-01-26 Jakub Jelinek <jakub@redhat.com>
475 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
476 emit fld b; fld a; if possible.
478 * brig-builtins.def: Update copyright years.
479 * config/arm/arm_acle_builtins.def: Update copyright years.
481 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
484 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
485 constraint instead of o for the stxsd instruction.
487 2017-01-25 Carl Love <cel@us.ibm.com>
489 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
490 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
492 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
494 * doc/invoke.texi (C++ Dialect Options): Fix typo.
496 2017-01-25 Richard Biener <rguenther@suse.de>
498 PR tree-optimization/69264
499 * target.def (vector_alignment_reachable): Improve documentation.
500 * doc/tm.texi: Regenerate.
501 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
503 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
504 earlier changes with respect to TYPE_USER_ALIGN.
505 (vector_alignment_reachable_p): Likewise. Improve dumping.
507 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
510 * config/arm/arm.md (xordi3): Force constant operand into a register
513 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
515 * doc/invoke.texi (-fstore-merging): Correct default optimization
516 levels at which it is enabled.
517 (-O): Move -fstore-merging from list to...
520 2017-01-25 Richard Biener <rguenther@suse.de>
523 * omp-expand.c: Include debug.h.
524 (expand_omp_taskreg): Make sure to generate early debug before
525 outlining anything from a function.
526 (expand_omp_target): Likewise.
527 (grid_expand_target_grid_body): Likewise.
529 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
532 * asan.c (get_translation_unit_decl): New function.
533 (asan_add_global): Extract modules file name from globals
534 TRANSLATION_UNIT_DECL name.
536 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
539 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
540 for long calls with APCS frame and VFP.
542 2017-01-24 David Malcolm <dmalcolm@redhat.com>
544 * cfg.c (original_copy_tables_initialized_p): New function.
545 * cfg.h (original_copy_tables_initialized_p): New decl.
546 * cfgrtl.c (relink_block_chain): Guard the call to
547 free_original_copy_tables with a call to
548 original_copy_tables_initialized_p.
549 * cgraph.h (symtab_node::native_rtl_p): New decl.
550 * cgraphunit.c (symtab_node::native_rtl_p): New function.
551 (symtab_node::needed_p): Don't assert for early assembly output
553 (cgraph_node::finalize_function): Set "force_output" for __RTL
555 (cgraph_node::analyze): Bail out early for __RTL functions.
556 (analyze_functions): Update assertion to support __RTL functions.
557 (cgraph_node::expand): Bail out early for __RTL functions.
558 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
560 * function.h (struct function): Update comment for field
562 * gimple-expr.c: Include "tree-pass.h".
563 (gimple_has_body_p): Return false for __RTL functions.
564 * Makefile.in (OBJS): Add run-rtl-passes.o.
565 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
567 (gcc::pass_manager::get_clean_slate): New accessor.
568 * passes.c: Include "insn-addr.h".
569 (should_skip_pass_p): Add logging. Update logic for running
570 "expand" to be compatible with both __GIMPLE and __RTL. Guard
571 property-provider override so it is only done for gimple passes.
573 (skip_pass): New function.
574 (execute_one_pass): Call skip_pass when skipping passes.
575 * read-md.c (md_reader::read_char): Support filtering
576 the input to a subset of line numbers.
577 (md_reader::md_reader): Initialize fields
578 m_first_line and m_last_line.
579 (md_reader::read_file_fragment): New function.
580 * read-md.h (md_reader::read_file_fragment): New decl.
581 (md_reader::m_first_line): New field.
582 (md_reader::m_last_line): New field.
583 * read-rtl-function.c (function_reader::create_function): Only
584 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
585 curr_properties. Set DECL_INITIAL to a dummy block.
586 (read_rtl_function_body_from_file_range): New function.
587 * read-rtl-function.h (read_rtl_function_body_from_file_range):
589 * run-rtl-passes.c: New file.
590 * run-rtl-passes.h: New file.
592 2017-01-24 Jeff Law <law@redhat.com>
594 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
597 2017-01-24 Bin Cheng <bin.cheng@arm.com>
599 PR tree-optimization/79159
600 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
601 (record_nonwrapping_iv): Improve boundary using above function if no
602 value range information.
604 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
605 Martin Jambor <mjambor@suse.cz>
607 * brig-builtins.def: New file.
608 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
609 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
610 (DEF_HSAIL_SAT_BUILTIN): Likewise.
611 (DEF_HSAIL_INTR_BUILTIN): Likewise.
612 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
613 * builtin-types.def (BT_INT8): New.
614 (BT_INT16): Likewise.
615 (BT_UINT8): Likewise.
616 (BT_UINT16): Likewise.
617 (BT_FN_ULONG): Likewise.
618 (BT_FN_UINT_INT): Likewise.
619 (BT_FN_UINT_ULONG): Likewise.
620 (BT_FN_UINT_LONG): Likewise.
621 (BT_FN_UINT_PTR): Likewise.
622 (BT_FN_ULONG_PTR): Likewise.
623 (BT_FN_INT8_FLOAT): Likewise.
624 (BT_FN_INT16_FLOAT): Likewise.
625 (BT_FN_UINT32_FLOAT): Likewise.
626 (BT_FN_UINT16_FLOAT): Likewise.
627 (BT_FN_UINT8_FLOAT): Likewise.
628 (BT_FN_UINT64_FLOAT): Likewise.
629 (BT_FN_UINT16_UINT32): Likewise.
630 (BT_FN_UINT32_UINT16): Likewise.
631 (BT_FN_UINT16_UINT16_UINT16): Likewise.
632 (BT_FN_INT_PTR_INT): Likewise.
633 (BT_FN_UINT_PTR_UINT): Likewise.
634 (BT_FN_LONG_PTR_LONG): Likewise.
635 (BT_FN_ULONG_PTR_ULONG): Likewise.
636 (BT_FN_VOID_UINT64_UINT64): Likewise.
637 (BT_FN_UINT8_UINT8_UINT8): Likewise.
638 (BT_FN_INT8_INT8_INT8): Likewise.
639 (BT_FN_INT16_INT16_INT16): Likewise.
640 (BT_FN_INT_INT_INT): Likewise.
641 (BT_FN_UINT_FLOAT_UINT): Likewise.
642 (BT_FN_FLOAT_UINT_UINT): Likewise.
643 (BT_FN_ULONG_UINT_UINT): Likewise.
644 (BT_FN_ULONG_UINT_PTR): Likewise.
645 (BT_FN_ULONG_ULONG_ULONG): Likewise.
646 (BT_FN_UINT_UINT_UINT): Likewise.
647 (BT_FN_VOID_UINT_PTR): Likewise.
648 (BT_FN_UINT_UINT_PTR: Likewise.
649 (BT_FN_UINT32_UINT64_PTR): Likewise.
650 (BT_FN_INT_INT_UINT_UINT): Likewise.
651 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
652 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
653 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
654 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
655 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
656 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
657 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
658 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
659 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
660 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
661 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
662 * doc/frontends.texi: List BRIG FE.
663 * doc/install.texi (Testing): Add BRIG tesring requirements.
664 * doc/invoke.texi (Overall Options): Mention BRIG.
665 * doc/standards.texi (Standards): Doucment BRIG HSA version.
667 2017-01-24 Richard Biener <rguenther@suse.de>
670 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
672 2017-01-24 Martin Jambor <mjambor@suse.cz>
675 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
676 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
679 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
682 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
683 casts from signed to unsigned really don't have a range.
685 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
687 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
688 GMP_RNDx for compatiblity.
690 2017-01-24 Martin Liska <mliska@suse.cz>
693 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
694 that would prevent us to call alloca with -1 as argument.
696 2017-01-24 Jakub Jelinek <jakub@redhat.com>
698 * dwarf2out.c (output_compilation_unit_header, output_file_names):
699 Avoid -Wformat-security warning.
701 2017-01-23 Andrew Pinski <apinski@cavium.com>
703 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
706 2017-01-23 Martin Sebor <msebor@redhat.com>
709 * gimple-ssa-sprintf.c (warn_level): New global.
710 (format_integer): Use it here and throughout the rest of the file.
711 Use the same switch to compute sign as base.
712 (maybe_warn): New function.
713 (format_directive): Factor out warnings into maybe_warn.
714 Add debugging output. Use warn_level.
715 (add_bytes): Use warn_level.
716 (pass_sprintf_length::compute_format_length): Add debugging output.
717 (try_substitute_return_value): Same.
718 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
721 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
722 (struct fmtresult, format_integer, format_floating): Adjust.
723 (fmtresult::fmtresult): Set max correctly in two argument ctor.
724 (get_string_length, format_string,format_directive): Same.
725 (pass_sprintf_length::compute_format_length): Same.
726 (try_substitute_return_value): Simplify slightly.
729 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
730 (fmtresult::operator+=): Outlined.
731 (struct fmtresult): Add ctors.
732 (struct conversion_spec): Rename...
733 (struct directive): ...to this. Add and remove data members.
734 (directive::set_width, directive::set_precision): New functions.
735 (format_percent): Use fmtresult ctor.
736 (get_width_and_precision): Remove.
737 (format_integer): Make naming changes. Avoid computing width and
739 (format_floating): Same. Adjust indentation.
740 (format_character, format_none): New functions.
741 (format_string): Moved character handling to format_character.
742 (format_directive): Remove arguments, change return type.
743 (parse_directive): New function.
744 (pass_sprintf_length::compute_format_length): Move directive
745 parsing to parse_directive.
747 2017-01-23 Jakub Jelinek <jakub@redhat.com>
749 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
750 (assign_assembler_name_if_needed): ... this.
751 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
752 (assign_assembler_name_if_needed): ... this.
753 (free_lang_data_in_cgraph): Adjust callers.
754 * cgraphunit.c (cgraph_node::analyze): Likewise.
755 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
758 2017-01-23 Richard Biener <rguenther@suse.de>
760 PR tree-optimization/79088
761 PR tree-optimization/79188
762 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
763 resetting loop bounds after last path deletion. Reset loop
764 bounds of the target loop, make code match the comments.
765 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
766 Make sure loops need no fixups.
768 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
770 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
771 exponent support with double type for first argument.
772 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
773 type returned by __builtin_vec_extract_sig,
774 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
775 functions from "vector int" to "vector unsigned int" or from
776 "vector long long int" to "vector unsigned long long int".
777 Changed type returned by __builtin_vec_extract_exp,
778 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
779 functions from "vector int" to "vector unsigned int" or from
780 "vector long long int" to "vector unsigned long long int".
781 Changed return type of __builtin_vec_test_data_class,
782 __builtin_vec_test_data_class_sp, and
783 __builtin_vec_test_data_class_dp from "vector int" to
784 "vector bool int" or from "vector long long int" to "vector bool
785 long long int" and changed second argument type from "unsigned
786 int" to "int". Added new overloaded function forms "vector float
787 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
788 "vector float __builtin_vec_insert_exp_sp (vector float, vector
789 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
790 double, vector unsigned long long int)" and "vector double
791 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
792 long int)". Changed return type of
793 __builtin_scalar_test_data_class and
794 __builtin_scalar_test_data_class_sp and
795 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
796 int" and changed second argument from "unsigned int" to "int".
797 Changed type returned by __builtin_scalar_test_neg,
798 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
799 from "int" to "bool int". Added new overloaded function form
800 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
801 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
802 exponent double-precision with floating point first argument.
803 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
804 documentation of scalar_test_data_class, scalar_test_neg,
805 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
806 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
807 vec_test_data_class built-in functions to reflect refinements in
808 their type signatures.
810 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
812 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
814 (aarch64_elf_asm_destructor): Likewise.
816 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
818 PR rtl-optimization/78634
819 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
820 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
821 * ifcvt.c (noce_try_cmove): Add missing cost check.
823 PR rtl-optimization/71724
824 * combine.c (if_then_else_cond): Look for situations where it is
825 beneficial to undo the work of one of the recursive calls.
827 2017-01-23 Bin Cheng <bin.cheng@arm.com>
829 PR tree-optimization/70754
830 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
831 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
832 combined stmt before it if not NULL.
833 (combine_chains): Process refs reversely and compute dominance point
836 2017-01-23 Martin Liska <mliska@suse.cz>
838 PR tree-optimization/79196
839 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
840 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
842 (strlen_optimize_stmt): Call the renamed function.
844 2017-01-23 Michael Matz <matz@suse.de>
846 PR tree-optimization/78384
847 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
849 2017-01-23 Richard Biener <rguenther@suse.de>
851 PR tree-optimization/79186
852 * tree-vrp.c (register_new_assert_for): Make sure we've seen
853 both incoming edges before moving an assert.
855 2017-01-23 Martin Jambor <mjambor@suse.cz>
857 * ipa-prop.c (load_from_param_1): Removed.
858 (load_from_unmodified_param): Bits from load_from_param_1 put back
860 (load_from_param): Removed.
861 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
862 with stmt. Reverted back to use of load_from_unmodified_param.
864 2017-01-23 Martin Jambor <mjambor@suse.cz>
867 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
868 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
869 field a pointer to garbage collected vector, mark lattices and
870 ipcp_orig_node with GTY((skip)).
871 (ipa_get_param_count): Adjust to descriptors being a pointer.
872 (ipa_get_param): Likewise.
873 (ipa_get_type): Likewise.
874 (ipa_get_param_move_cost): Likewise.
875 (ipa_set_param_used): Likewise.
876 (ipa_get_controlled_uses): Likewise.
877 (ipa_set_controlled_uses): Likewise.
878 (ipa_is_param_used): Likewise.
879 (ipa_node_params_t): Move into garbage collector. New methods insert
881 (ipa_node_params_sum): Annotate wth GTY(()).
882 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
884 (ipa_load_from_parm_agg): Adjust declaration.
885 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
886 * ipa-profile.c (ipa_profile): Likewise.
887 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
888 (ipa_populate_param_decls): Make descriptors parameter garbage
890 (ipa_dump_param): Adjust to descriptors being a pointer.
891 (ipa_alloc_node_params): Likewise.
892 (ipa_initialize_node_params): Likewise.
893 (load_from_param_1): Make descriptors parameter garbage collected.
894 (load_from_unmodified_param): Likewise.
895 (load_from_param): Likewise.
896 (ipa_load_from_parm_agg): Likewise.
897 (ipa_node_params::~ipa_node_params): Removed.
898 (ipa_free_all_node_params): Remove call to delete operator.
899 (ipa_node_params_t::insert): New.
900 (ipa_node_params_t::remove): Likewise.
901 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
902 copy known_csts and known_contexts vectors.
903 (ipa_read_node_info): Adjust to descriptors being a pointer.
904 (ipcp_modif_dom_walker): Make m_descriptors field garbage
906 (ipcp_transform_function): Make descriptors variable garbage
909 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
911 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
912 * config/i386/avx512dqintrin.h: Ditto.
913 * config/i386/avx512fintrin.h: Ditto.
914 * gcc/config/i386/i386.c: Handle new builtins.
915 * config/i386/i386-builtin.def: Add new builtins.
916 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
917 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
919 2017-01-23 Jakub Jelinek <jakub@redhat.com>
920 Martin Liska <mliska@suse.cz>
922 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
923 * asan.c (asan_expand_poison_ifn): Support stores and use
924 appropriate ASAN report function.
925 * internal-fn.c (expand_ASAN_POISON_USE): New function.
926 * internal-fn.def (ASAN_POISON_USE): Declare.
927 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
928 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
929 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
930 ASAN_POISON calls w/o LHS.
931 * tree-ssa.c (execute_update_addresses_taken): Create clobber
932 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
933 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
934 * gimplify.c (asan_poison_variables): Add attribute
935 use_after_scope_memory to variables that really needs to live
937 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
938 having the attribute.
940 2017-01-23 Martin Liska <mliska@suse.cz>
942 * asan.c (create_asan_shadow_var): New function.
943 (asan_expand_poison_ifn): Likewise.
944 * asan.h (asan_expand_poison_ifn): New declaration.
945 * internal-fn.c (expand_ASAN_POISON): Likewise.
946 * internal-fn.def (ASAN_POISON): New builtin.
947 * sanopt.c (pass_sanopt::execute): Expand
948 asan_expand_poison_ifn.
949 * tree-inline.c (copy_decl_for_dup_finish): Make function
951 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
952 * tree-ssa.c (is_asan_mark_p): New function.
953 (execute_update_addresses_taken): Rewrite local variables
954 (identified just by use-after-scope as addressable) into SSA.
956 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
958 * doc/install.texi (Specific): opensource.apple.com uses https
959 now. Remove trailing slash.
961 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
963 * README.Portability: Remove note on an Irix compatibility issue.
965 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
967 * gcov.c (INCLUDE_ALGORITHM): Define.
968 (INCLUDE_VECTOR): Define.
969 No longer include <vector> and <algorithm> directly.
971 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
973 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
975 * doc/invoke.texi (Code Gen Options): Ditto.
977 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
980 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
982 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
984 rtl-optimization/79125
985 * cprop.c (local_cprop_pass): Handle cases where we make an
988 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
992 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
993 read from, for big endian.
995 2017-01-20 Jiong Wang <jiong.wang@arm.com>
997 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
998 register pauth builtins for LP64 only.
1000 2017-01-20 Marek Polacek <polacek@redhat.com>
1003 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1006 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1008 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1010 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1011 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1012 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1014 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1017 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1018 in big-endian mode when they are not a single duplicated value.
1020 2017-01-20 Richard Biener <rguenther@suse.de>
1022 * BASE-VER: Bump to 7.0.1.
1024 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1026 * omp-low.c (omplow_simd_context): New struct. Use it...
1027 (lower_rec_simd_input_clauses): ...here and...
1028 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1029 references to idx, lane, max_vf, is_simt.
1031 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1033 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1036 2017-01-20 Martin Jambor <mjambor@suse.cz>
1038 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1039 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1041 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1042 (GTFILES): Rename hsa.c to hsa-common.c.
1043 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1044 * hsa-dump.c: Likewise.
1045 * hsa-gen.c: Likewise.
1046 * hsa-regalloc.c: Likewise.
1047 * ipa-hsa.c: Likewise.
1048 * omp-expand.c: Likewise.
1049 * omp-low.c: Likewise.
1050 * toplev.c: Likewise.
1052 2017-01-20 Marek Polacek <polacek@redhat.com>
1055 * doc/invoke.texi: Document -Wduplicated-branches.
1056 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1057 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1058 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1059 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1060 return 0 only when not OEP_LEXICOGRAPHIC.
1061 (fold_build_cleanup_point_expr): Use the expression
1062 location when building CLEANUP_POINT_EXPR.
1063 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1064 * tree.c (add_expr): Handle error_mark_node.
1066 2017-01-20 Martin Liska <mliska@suse.cz>
1069 * tree-profile.c (init_ic_make_global_vars): Do not call
1071 (gimple_init_gcov_profiler): Likewise.
1073 2017-01-20 Martin Liska <mliska@suse.cz>
1076 * cgraph.h (maybe_create_reference): Remove argument and
1078 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1080 * ipa-cp.c (create_specialized_node): Likewise.
1081 * symtab.c (symtab_node::maybe_create_reference): Handle
1082 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1084 2017-01-20 Martin Liska <mliska@suse.cz>
1086 * read-rtl-function.c (function_reader::create_function): Use
1087 build_decl instread of build_decl_stat.
1089 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1091 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1092 * config/i386/avx512dqintrin.h: Ditto.
1093 * config/i386/avx512fintrin.h: Ditto.
1094 * config/i386/i386-builtin-types.def: Add new types.
1095 * gcc/config/i386/i386.c: Handle new types.
1096 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1097 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1098 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1099 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1100 (__builtin_ia32_kshiftridi): New.
1101 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1103 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1107 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1108 define to rs6000_init_stack_protect_guard.
1109 (rs6000_init_stack_protect_guard): New function.
1111 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1112 Yunqiang Su <yunqiang.su@imgtec.com>
1114 * config.gcc (supported_defaults): Add madd4.
1115 (with_madd4): Add validation.
1116 (all_defaults): Add madd4.
1117 * config/mips/mips.opt (mmadd4): New option.
1118 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1120 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1122 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1123 (ISA_HAS_FUSED_MADD4): Likewise.
1124 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1125 * gcc/doc/install.texi (--with-madd4): Document the new option.
1127 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1129 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1130 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1131 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1132 (aarch64_init_pauth_hint_builtins): New.
1133 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1134 (aarch64_expand_builtin): Expand new builtins.
1136 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1138 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1139 * combine-stack-adj.c (no_unhandled_cfa): Handle
1140 REG_CFA_TOGGLE_RA_MANGLE.
1141 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1142 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1143 info for return address signing.
1144 (aarch64_expand_epilogue): Likewise.
1146 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1148 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1149 * config/aarch64/aarch64-protos.h
1150 (aarch64_return_address_signing_enabled): New declaration.
1151 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1153 (aarch64_expand_prologue): Sign return address before it's pushed onto
1155 (aarch64_expand_epilogue): Authenticate return address fetched from
1157 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1158 (aarch64_attributes): New function attributes for "sign-return-address".
1159 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1160 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1161 ("*do_return"): Generate combined instructions according to key index.
1162 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1163 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1165 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1166 * config/aarch64/aarch64.opt (msign-return-address=): New.
1167 * doc/extend.texi (AArch64 Function Attributes): Documents
1168 "sign-return-address=".
1169 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1171 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1173 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1174 overall option summary.
1176 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1178 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1179 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1180 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1181 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1183 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1185 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1186 -mpower9-minmax by default for -mcpu=power9.
1187 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1188 128-bit floating point.
1190 2017-01-20 Alan Modra <amodra@gmail.com>
1192 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1193 optimizing for size.
1195 2017-01-20 Alan Modra <amodra@gmail.com>
1198 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1199 for strcmp and strncmp from corresponding builtin decl.
1201 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1203 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1204 instead of i386/rtems-64.h.
1205 * config/i386/rtems-64.h: Remove.
1207 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1211 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1213 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1215 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1217 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1218 Change int to HOST_WIDE_INT.
1219 * config/aarch64/aarch64-protos.h
1220 (aarch64_simd_gen_const_vector_dup): Likewise.
1221 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1223 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1225 * langhooks-def.h (lhd_type_for_size): New decl.
1226 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1227 * langhooks.c (lhd_type_for_size): New function, taken from
1230 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1232 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1233 define_bypass for CR latency.
1234 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1235 (power9-alu2): Add define_bypass for CR latency.
1237 (power9-mul): Update insn latency.
1238 (power9-mul-compare): Update insn latency, bypass latency and remove
1241 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1243 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1245 * config/aarch64/aarch64.md
1246 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1247 aarch64_nopcrelative_literal_loads.
1248 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1250 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1252 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1254 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1256 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1259 * config.gcc (supported_defaults): Add lxc1-sxc1.
1260 (with_lxc1_sxc1): Add validation.
1261 (all_defaults): Add lxc1-sxc1.
1262 * config/mips/mips.opt (mlxc1-sxc1): New option.
1263 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1265 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1266 __mips_no_lxc1_sxc1.
1267 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1268 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1269 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1271 2017-01-19 Richard Biener <rguenther@suse.de>
1273 PR tree-optimization/72488
1274 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1275 sure to restore SSA info.
1276 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1278 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1280 PR rtl-optimization/79121
1281 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1282 of the inner type when shifting an extended value.
1284 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1287 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1288 interposable aliases.
1290 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1293 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1294 Use the evmergelohi instruction.
1295 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1296 (mov_sitf_e500_subreg8_2_be): Likewise.
1297 (mov_sitf_e500_subreg12_2_le): Likewise.
1298 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1299 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1300 (mov_sitf_e500_subreg8_2_le): Likewise.
1301 (mov_sitf_e500_subreg12_2_be): Likewise.
1303 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1305 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1306 attribute from vecsimple to vecperm.
1307 (altivec_vbpermq2): Likewise.
1309 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1312 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1314 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1315 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1316 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1317 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1318 case where N arg is SIZE_MAX.
1319 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1320 (cmpstrsi): Add pattern.
1322 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1324 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1325 __builtin_vec_revb builtins.
1326 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1327 built-in functions to support generation of the ISA 3.0 XXBR<x>
1328 vector byte reverse instructions.
1329 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1330 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1331 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1332 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1333 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1334 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1335 (P9V_BUILTIN_VEC_REVB): Likewise.
1336 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1337 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1338 (p9_xxbrq_v16qi): Likewise.
1339 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1340 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1341 (p9_xxbrh_v8hi): Likewise.
1342 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1343 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1344 vec_revb built-in functions.
1346 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1348 PR rtl-optimization/78952
1349 * config/i386/i386.md (any_extract): New code iterator.
1350 (*insvqi_2): Use any_extract for source operand.
1351 (*insvqi_3): Use any_shiftrt for source operand.
1353 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1355 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1357 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1359 2017-01-18 Matthias Klose <doko@ubuntu.com>
1361 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1363 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1365 * config/rs6000/altivec.h (vec_bperm): Change #define.
1366 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1367 (altivec_vbpermq2): New define_insn.
1368 (altivec_vbpermd): Likewise.
1369 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1371 (VBPERMD): Likewise.
1372 (VBPERM): New polymorphic function interface.
1373 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1374 Add entries for P9V_BUILTIN_VEC_VBPERM.
1375 * doc/extend.texi: Add interfaces for vec_bperm.
1377 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1379 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1380 first letter of error messages.
1381 (s390_resolve_overloaded_builtin): Likewise.
1382 * config/s390/s390.c (s390_expand_builtin): Likewise.
1383 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1384 (s390_valid_target_attribute_inner_p): Likewise.
1385 * config/s390/s390.md ("tabort"): Likewise.
1387 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1389 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1390 (ISA_AVOID_DIV_HILO): New macro.
1391 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1392 (ISA_HAS_DDIV): Likewise.
1394 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1396 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1398 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1400 * doc/invoke.texi (fabi-version): Spelling fix.
1402 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1405 * doc/invoke.texi (fabi-version): Mention mangling fix for
1408 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1411 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1413 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1416 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1417 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1419 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1421 (stack_protect_test): Ditto.
1422 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1423 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1425 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1426 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1427 -mstack-protector-guard-offset=.
1428 (RS/6000 and PowerPC Options): Ditto.
1430 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1432 * config/i386/i386.h (MASK_CLASS_P): New define.
1433 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1434 there are no registers from different register sets also when
1435 mask registers are used. Update function comment.
1436 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1437 to (*k/*r) and (*k/*km) alternatives.
1439 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1441 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1442 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1443 (EH_RETURN_HANDLER_RTX): New define.
1444 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1445 Force frame pointer in EH return functions.
1446 (aarch64_expand_epilogue): Add barrier for eh_return.
1447 (aarch64_final_eh_return_addr): Remove.
1448 (aarch64_eh_return_handler_rtx): New function.
1449 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1451 (aarch64_eh_return_handler_rtx): New prototype.
1453 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1455 * config/rs6000/altivec.h (vec_rlmi): New #define.
1456 (vec_vrlnm): Likewise.
1457 (vec_rlnm): Likewise.
1458 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1459 (UNSPEC_VRLNM): Likewise.
1460 (VIlong): New mode iterator.
1461 (altivec_vrl<VI_char>mi): New define_insn.
1462 (altivec_vrl<VI_char>nm): Likewise.
1463 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1466 (RLNM): New polymorphic function entry.
1467 (VRLWMI): New monomorphic function entry.
1469 (RLMI): New polymorphic function entry.
1470 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1471 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1472 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1475 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1478 * dwarf2out.c (field_byte_offset): Restore the
1479 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1480 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1481 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1484 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1487 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1489 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1492 * dwarf2out.c (add_data_member_location_attribute): For constant
1493 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1494 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1495 DW_AT_byte_size attributes.
1497 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1499 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1500 after forcing to constant memory when the code model is medium.
1502 2017-01-17 Julia Koval <julia.koval@intel.com>
1505 * config/i386/avx512fintrin.h
1506 (_mm512_i32gather_ps): Change __addr type to void const*.
1507 (_mm512_mask_i32gather_ps): Ditto.
1508 (_mm512_i32gather_pd): Ditto.
1509 (_mm512_mask_i32gather_pd): Ditto.
1510 (_mm512_i64gather_ps): Ditto.
1511 (_mm512_mask_i64gather_ps): Ditto.
1512 (_mm512_i64gather_pd): Ditto.
1513 (_mm512_mask_i64gather_pd): Ditto.
1514 (_mm512_i32gather_epi32): Ditto.
1515 (_mm512_mask_i32gather_epi32): Ditto.
1516 (_mm512_i32gather_epi64): Ditto.
1517 (_mm512_mask_i32gather_epi64): Ditto.
1518 (_mm512_i64gather_epi32): Ditto.
1519 (_mm512_mask_i64gather_epi32): Ditto.
1520 (_mm512_i64gather_epi64): Ditto.
1521 (_mm512_mask_i64gather_epi64): Ditto.
1522 (_mm512_i32scatter_ps): Change __addr type to void*.
1523 (_mm512_mask_i32scatter_ps): Ditto.
1524 (_mm512_i32scatter_pd): Ditto.
1525 (_mm512_mask_i32scatter_pd): Ditto.
1526 (_mm512_i64scatter_ps): Ditto.
1527 (_mm512_mask_i64scatter_ps): Ditto.
1528 (_mm512_i64scatter_pd): Ditto.
1529 (_mm512_mask_i64scatter_pd): Ditto.
1530 (_mm512_i32scatter_epi32): Ditto.
1531 (_mm512_mask_i32scatter_epi32): Ditto.
1532 (_mm512_i32scatter_epi64): Ditto.
1533 (_mm512_mask_i32scatter_epi64): Ditto.
1534 (_mm512_i64scatter_epi32): Ditto.
1535 (_mm512_mask_i64scatter_epi32): Ditto.
1536 (_mm512_i64scatter_epi64): Ditto.
1537 (_mm512_mask_i64scatter_epi64): Ditto.
1538 * config/i386/avx512pfintrin.h
1539 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1540 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1541 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1542 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1543 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1544 (_mm512_prefetch_i32scatter_ps): Ditto.
1545 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1546 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1547 (_mm512_prefetch_i64scatter_pd): Ditto.
1548 (_mm512_prefetch_i64scatter_ps): Ditto.
1549 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1550 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1551 * config/i386/avx512vlintrin.h
1552 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1553 (_mm_mmask_i32gather_ps): Ditto.
1554 (_mm256_mmask_i32gather_pd): Ditto.
1555 (_mm_mmask_i32gather_pd): Ditto.
1556 (_mm256_mmask_i64gather_ps): Ditto.
1557 (_mm_mmask_i64gather_ps): Ditto.
1558 (_mm256_mmask_i64gather_pd): Ditto.
1559 (_mm_mmask_i64gather_pd): Ditto.
1560 (_mm256_mmask_i32gather_epi32): Ditto.
1561 (_mm_mmask_i32gather_epi32): Ditto.
1562 (_mm256_mmask_i32gather_epi64): Ditto.
1563 (_mm_mmask_i32gather_epi64): Ditto.
1564 (_mm256_mmask_i64gather_epi32): Ditto.
1565 (_mm_mmask_i64gather_epi32): Ditto.
1566 (_mm256_mmask_i64gather_epi64): Ditto.
1567 (_mm_mmask_i64gather_epi64): Ditto.
1568 (_mm256_i32scatter_ps): Change __addr type to void*.
1569 (_mm256_mask_i32scatter_ps): Ditto.
1570 (_mm_i32scatter_ps): Ditto.
1571 (_mm_mask_i32scatter_ps): Ditto.
1572 (_mm256_i32scatter_pd): Ditto.
1573 (_mm256_mask_i32scatter_pd): Ditto.
1574 (_mm_i32scatter_pd): Ditto.
1575 (_mm_mask_i32scatter_pd): Ditto.
1576 (_mm256_i64scatter_ps): Ditto.
1577 (_mm256_mask_i64scatter_ps): Ditto.
1578 (_mm_i64scatter_ps): Ditto.
1579 (_mm_mask_i64scatter_ps): Ditto.
1580 (_mm256_i64scatter_pd): Ditto.
1581 (_mm256_mask_i64scatter_pd): Ditto.
1582 (_mm_i64scatter_pd): Ditto.
1583 (_mm_mask_i64scatter_pd): Ditto.
1584 (_mm256_i32scatter_epi32): Ditto.
1585 (_mm256_mask_i32scatter_epi32): Ditto.
1586 (_mm_i32scatter_epi32): Ditto.
1587 (_mm_mask_i32scatter_epi32): Ditto.
1588 (_mm256_i32scatter_epi64): Ditto.
1589 (_mm256_mask_i32scatter_epi64): Ditto.
1590 (_mm_i32scatter_epi64): Ditto.
1591 (_mm_mask_i32scatter_epi64): Ditto.
1592 (_mm256_i64scatter_epi32): Ditto.
1593 (_mm256_mask_i64scatter_epi32): Ditto.
1594 (_mm_i64scatter_epi32): Ditto.
1595 (_mm_mask_i64scatter_epi32): Ditto.
1596 (_mm256_i64scatter_epi64): Ditto.
1597 (_mm256_mask_i64scatter_epi64): Ditto.
1598 (_mm_i64scatter_epi64): Ditto.
1599 (_mm_mask_i64scatter_epi64): Ditto.
1600 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1601 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1602 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1603 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1604 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1605 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1606 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1607 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1608 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1609 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1610 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1611 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1612 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1613 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1614 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1615 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1616 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1617 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1618 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1619 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1620 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1621 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1622 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1623 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1624 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1625 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1626 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1627 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1628 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1629 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1630 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1631 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1632 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1633 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1634 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1635 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1636 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1637 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1638 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1639 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1640 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1641 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1642 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1643 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1644 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1645 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1646 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1647 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1648 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1649 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1650 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1651 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1652 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1653 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1654 definitions accordingly.
1656 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1657 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1660 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1663 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1666 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1667 hard regs for inner regno.
1669 2017-01-17 Martin Liska <mliska@suse.cz>
1672 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1673 assumption and add comment.
1675 2017-01-17 Nathan Sidwell <nathan@acm.org>
1677 * ipa-visibility.c (localize_node): New function, broken out of ...
1678 (function_and_variable_visibility): ... here. Call it.
1680 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1683 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1684 correctly set frequency of oudgoing edge.
1685 (duplicate_thread_path): Fix profile updating.
1687 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1690 * configure.ac: Add GCC_BASE_VER.
1691 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1692 version from BASE-VER file.
1693 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1694 (gcc.o): Depend on $(BASEVER).
1695 * common.opt (dumpfullversion): New option.
1696 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1697 * doc/invoke.texi: Document -dumpfullversion.
1698 * doc/install.texi: Document --with-gcc-major-version-only.
1699 * configure: Regenerated.
1701 2017-01-17 Richard Biener <rguenther@suse.de>
1703 PR tree-optimization/71433
1704 * tree-vrp.c (register_new_assert_for): Merge same asserts
1705 on all incoming edges.
1706 (process_assert_insertions_for): Handle insertions at the
1709 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1711 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1712 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1714 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1717 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1720 2017-01-17 Alan Modra <amodra@gmail.com>
1723 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1724 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1725 symbolic stack limit when pic.
1727 2017-01-16 Martin Sebor <msebor@redhat.com>
1729 PR tree-optimization/78608
1730 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1732 2017-01-16 Jeff Law <law@redhat.com>
1735 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1736 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1737 for several include directories that may be relative to sysroot.
1738 * config/i386/x-mingw32 (gplus_includedir): Define.
1739 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1740 (native_system_includedir): Likewise.
1741 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1742 override if TARGET_SYSTEM_ROOT is defined.
1743 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1745 PR tree-optimization/79090
1746 PR tree-optimization/33562
1747 PR tree-optimization/61912
1748 PR tree-optimization/77485
1749 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1750 and computed trims into the dump file.
1752 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1754 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1756 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1759 * gimplify.c (gimplify_init_constructor): If want_value and
1760 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1764 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1765 sequence. Formatting fixes.
1766 (doloop_optimize): Formatting fixes.
1769 * gcc.c (debug_level_greater_than_spec_func): New function.
1770 (static_spec_functions): Add debug-level-gt spec function.
1771 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1773 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1774 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1775 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1776 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1777 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1778 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1780 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1782 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1783 QImode fixups to general and mask registers only.
1785 2017-01-16 Carl Love <cel@us.ibm.com>
1787 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1788 for built-in functions
1789 vector signed char vec_nabs (vector signed char)
1790 vector signed short vec_nabs (vector signed short)
1791 vector signed int vec_nabs (vector signed int)
1792 vector signed long long vec_nabs (vector signed long long)
1793 vector float vec_nabs (vector float)
1794 vector double vec_nabs (vector double)
1795 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1797 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1798 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1799 * doc/extend.texi: Update the documentation file for the new built-in
1802 2017-01-16 Martin Sebor <msebor@redhat.com>
1804 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1807 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1809 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1810 UNSPEC_VSX__XXSPLTD to require special splat handling.
1812 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1815 * system.h: Poison strndup.
1817 2017-01-16 Alan Modra <amodra@gmail.com>
1820 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1823 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1825 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1827 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1829 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1830 call recog here. Assert that INSN_CODE (insn) is non-negative.
1832 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1835 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1837 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1838 in the currently scheduled RTL fragment.
1840 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1842 PR rtl-optimization/78751
1843 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1846 2017-01-14 Jeff Law <law@redhat.com>
1848 PR tree-optimization/79090
1849 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1850 variable length stores.
1851 (compute_trims): Delete dead assignment to *trim_tail.
1852 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1855 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1857 PR rtl-optimization/78626
1858 PR rtl-optimization/78727
1859 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1860 of a block, and split such blocks after everything else is finished.
1862 2017-01-14 Alan Modra <amodra@gmail.com>
1865 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1866 target legitimate_combined_insn.
1867 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1868 (rs6000_legitimate_combined_insn): New function.
1869 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1871 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1872 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1873 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1875 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1877 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1879 2017-01-13 Jeff Law <law@redhat.com>
1881 PR tree-optimization/33562
1882 PR tree-optimization/61912
1883 PR tree-optimization/77485
1884 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1886 (delete_dead_assignment): Likewise.
1887 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1888 statement to delete_dead_call and delete_dead_assignment.
1890 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1893 * substring-locations.c (format_warning_va): Strengthen case 1 so
1894 that both endpoints of the substring must be within the format
1895 range for just the substring to be printed.
1897 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1899 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1900 * config/i386/i386.c (ix86_target_string): Add missing options
1901 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1902 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1903 flags_other and ix86_target_other to flags2_other. Display unknown
1905 (ix86_valid_target_attribute_inner_p): Add missing options and
1906 reorder options by implied ISAs, as in ix86_target_string.
1908 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1910 * hash-table.h (hash_table::too_empty_p): New function.
1911 (hash_table::expand): Use it.
1912 (hash_table::traverse): Likewise.
1913 (hash_table::empty_slot): Use sizeof (value_type) instead of
1914 sizeof (PTR) to convert bytes to elements. Shrink the table
1915 if the current size is excessive for the current number of
1918 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1920 * ira-costs.c (record_reg_classes): Break from the inner loop
1921 early once alt_fail is known to be true. Update outer loop
1922 handling accordingly.
1924 2017-01-13 Jeff Law <law@redhat.com>
1926 * tree-ssa-dse.c (decrement_count): New function.
1927 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1928 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1929 when we know the partially dead statement is a mem* function.
1931 PR tree-optimization/61912
1932 PR tree-optimization/77485
1933 * tree-ssa-dse.c: Include expr.h.
1934 (maybe_trim_constructor_store): New function.
1935 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1937 PR tree-optimization/33562
1938 PR tree-optimization/61912
1939 PR tree-optimization/77485
1940 * doc/invoke.texi: Document new dse-max-object-size param.
1941 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1942 * tree-ssa-dse.c: Include params.h.
1943 (dse_store_status): New enum.
1944 (initialize_ao_ref_for_dse): New, partially extracted from
1946 (valid_ao_ref_for_dse, normalize_ref): New.
1947 (setup_live_bytes_from_ref, compute_trims): Likewise.
1948 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1949 (maybe_trim_partially_dead_store): Likewise.
1950 (maybe_trim_complex_store): Likewise.
1951 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1952 Track what bytes live from the original store. Return tri-state
1953 for dead, partially dead or live.
1954 (dse_dom_walker): Add constructor, destructor and new private members.
1955 (delete_dead_call, delete_dead_assignment): New extracted from
1957 (dse_optimize_stmt): Make a member of dse_dom_walker.
1958 Use initialize_ao_ref_for_dse.
1960 PR tree-optimization/33562
1961 PR tree-optimization/61912
1962 PR tree-optimization/77485
1963 * sbitmap.h (bitmap_count_bits): Prototype.
1964 (bitmap_clear_range, bitmap_set_range): Likewise.
1965 * sbitmap.c (bitmap_clear_range): New function.
1966 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1968 2017-01-13 Martin Liska <mliska@suse.cz>
1971 * function.c (set_cfun): Add new argument force.
1972 * function.h (set_cfun): Likewise.
1973 * ipa-inline-transform.c (inline_call): Use the function when
1974 strict alising from is dropped for function we inline to.
1976 2017-01-13 Richard Biener <rguenther@suse.de>
1978 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1979 for dumping GIMPLE INTEGER_CSTs.
1981 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1983 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1984 to 201112L since C++17.
1986 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1989 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1990 if -fsanitize=kernel-address is present.
1992 2017-01-13 Richard Biener <rguenther@suse.de>
1994 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1995 as _Literal ( type ) number in case usual suffixes do not
1996 preserve all information.
1998 2017-01-13 Richard Biener <rguenther@suse.de>
2000 PR tree-optimization/77283
2001 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2002 and ssa-iterators.h.
2003 (is_feasible_trace): Implement a cost model based on joiner
2006 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2009 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2010 char or short to __float128/_Float128 directly.
2012 2017-01-12 Martin Sebor <msebor@redhat.com>
2014 to -Wformat-overflow.
2015 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2016 (min_bytes_remaining): Same.
2017 (get_string_length): Same.
2018 (format_string): Same.
2019 (format_directive): Same.
2021 (pass_sprintf_length::handle_gimple_call): Same.
2023 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2025 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2026 info.nowrite calls with no lhs that can't throw. Return bool
2027 whether gsi_remove has been called or not.
2028 (pass_sprintf_length::handle_gimple_call): Return bool whether
2029 try_substitute_return_value called gsi_remove. Formatting fix.
2030 (pass_sprintf_length::execute): Don't use gsi_remove if
2031 handle_gimple_call returned true.
2034 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2035 be removed due to side-effects, don't remove following barrier nor
2036 turn the successor edge into fallthru edge.
2038 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2041 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2042 element-reversing loads and stores as not swappable.
2044 2017-01-12 Nathan Sidwell <nathan@acm.org>
2045 Nicolai Stange <nicstange@gmail.com>
2047 * combine.c (try_combine): Don't ignore result of overlap checking
2048 loop. Combine overlap & asm check into single loop.
2050 2017-01-12 Richard Biener <rguenther@suse.de>
2052 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2053 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2055 2017-01-12 Richard Biener <rguenther@suse.de>
2057 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2058 and TS_TARGET_OPTION directly derive from TS_BASE.
2059 * tree-core.h (tree_optimization_option): Derive from tree_base.
2060 (tree_target_option): Likewise.
2062 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2064 * config/i386/i386.c (memory_address_length): Increase len
2065 only when rip_relative_addr_p returns false.
2067 2017-01-11 Julia Koval <julia.koval@intel.com>
2069 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2070 (OPTION_MASK_ISA_SGX_SET): New.
2071 (ix86_handle_option): Handle OPT_msgx.
2072 * config.gcc: Added sgxintrin.h.
2073 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2074 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2075 * config/i386/i386.c (ix86_target_string): Add -msgx.
2077 (ix86_option_override_internal): Handle new options.
2078 (ix86_valid_target_attribute_inner_p): Add sgx.
2079 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2080 * config/i386/i386.opt: Add msgx.
2081 * config/i386/sgxintrin.h: New file.
2082 * config/i386/x86intrin.h: Add sgxintrin.h.
2084 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2087 * fold-const.c (maybe_nonzero_address): Return 1 for function
2089 (tree_single_nonzero_warnv_p): Don't handle function local objects
2093 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2096 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2099 * opts.c: Include "spellcheck.h"
2100 (struct string_fragment): New struct.
2101 (struct edit_distance_traits<const string_fragment &>): New
2103 (get_closest_sanitizer_option): New function.
2104 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2106 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2108 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2110 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2111 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2112 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2113 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2114 for initial die_offset if dwarf_split_debug_info.
2115 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2116 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2118 (output_skeleton_debug_sections): Formatting fix. Use
2119 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2120 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2122 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2124 * config/arm/cortex-a53.md: Add bypasses for
2126 (cortex_a53_r2f): Only use for transfers.
2127 (cortex_a53_f2r): Likewise.
2128 (cortex_a53_r2f_cvt): Add reservation for conversions.
2129 (cortex_a53_f2r_cvt): Likewise.
2131 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2133 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2134 to all inlined functions, change static to extern.
2136 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2139 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2141 (arm_assemble_integer): Likewise.
2143 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2145 * config.gcc: Use new awk script to check CPU, FPU and architecture
2146 parameters for --with-... options.
2147 * config/arm/parsecpu.awk: New file
2148 * config/arm/arm-cpus.in: New file.
2149 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2151 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2153 * config/arm/t-arm: Update dependency rules.
2154 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2155 of processing .def files.
2156 * config/arm/genopt.sh: Deleted.
2157 * config/arm/gentune.sh: Deleted.
2158 * config/arm/arm-cores.def: Deleted.
2159 * config/arm/arm-arches.def: Deleted.
2160 * config/arm/arm-fpus.def: Deleted.
2161 * config/arm/arm-tune.md: Regenerated.
2162 * config/arm/arm-tables.opt: Regenerated.
2163 * config/arm/arm-cpu.h: New generated file.
2164 * config/arm/arm-cpu-data.h: New generated file.
2165 * config/arm/arm-cpu-cdata.h: New generated file.
2167 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2170 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2172 (input_varpool_node): Unpack dynamically_initialized bit.
2174 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2176 PR rtl-optimization/79032
2177 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2178 the alignment of the adjusted memory reference against that of MODE,
2179 instead of the alignment of the original memory reference.
2181 2017-01-11 Martin Jambor <mjambor@suse.cz>
2183 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2185 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2186 decorated functions.
2188 2017-01-11 Richard Biener <rguenther@suse.de>
2190 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2191 set range/nonnull info for PHI results. Do not set it on
2192 stmts marked for removal.
2194 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2196 * expr.c (store_field): In the bitfield case, fetch the return value
2197 from the registers before applying a single big-endian adjustment.
2198 Always do a final load for a BLKmode value not larger than a word.
2200 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2203 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2204 that we correctly handle column numbers greater than
2205 LINE_MAP_MAX_COLUMN_NUMBER.
2207 2017-01-10 Martin Sebor <msebor@redhat.com>
2210 * gimple-ssa-sprintf.c (get_destination_size): Call
2211 {init,fini}object_sizes.
2212 * tree-object-size.c (addr_object_size): Adjust.
2213 (pass_through_call): Adjust.
2214 (pass_object_sizes::execute): Adjust.
2215 * tree-object-size.h (fini_object_sizes): Declare.
2217 2017-01-10 Martin Sebor <msebor@redhat.com>
2219 PR tree-optimization/78775
2220 * builtins.c (get_size_range): Move...
2221 * calls.c: ...to here.
2222 (alloc_max_size): Accept zero argument.
2223 (operand_signed_p): Remove.
2224 (maybe_warn_alloc_args_overflow): Call get_size_range.
2225 * calls.h (get_size_range): Declare.
2227 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2229 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2230 from TI's devices.csv file as of September 2016.
2231 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2233 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2235 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2236 * doc/invoke.texi: Likewise.
2237 * doc/md.texi: Likewise.
2238 * doc/objc.texi: Likewise.
2240 2017-01-10 Joshua Conner <joshconner@google.com>
2242 * config/arm/fuchsia-elf.h: New file.
2243 * config/fuchsia.h: New file.
2244 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2245 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2247 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2249 2016-01-10 Richard Biener <rguenther@suse.de>
2251 PR tree-optimization/79034
2252 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2253 Propagate out degenerate PHIs in the joiner.
2255 2017-01-10 Martin Liska <mliska@suse.cz>
2257 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2258 (sort_congruence_classes_by_decl_uid): Likewise.
2259 (sort_congruence_class_groups_by_decl_uid): Likewise.
2260 (sem_item_optimizer::merge_classes): Sort class, groups in these
2261 classes and members in the groups by DECL_UID of declarations.
2262 This would make merge operations stable.
2264 2017-01-10 Martin Liska <mliska@suse.cz>
2266 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2267 usage of m_classes_vec.
2268 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2269 (sem_item_optimizer::get_group_by_hash): Likewise.
2270 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2271 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2272 (sem_item_optimizer::verify_classes): Likewise.
2273 (sem_item_optimizer::process_cong_reduction): Likewise.
2274 (sem_item_optimizer::dump_cong_classes): Likewise.
2275 (sem_item_optimizer::merge_classes): Likewise.
2276 * ipa-icf.h (congruence_class_hash): Rename from
2277 congruence_class_group_hash. Remove declaration of m_classes_vec.
2279 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2281 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2282 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2283 * config.gcc: Add avx512vpopcntdqintrin.h.
2284 * config/i386/avx512vpopcntdqintrin.h: New.
2285 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2286 * config/i386/i386-builtin-types.def: Add new types.
2287 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2288 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2289 __builtin_ia32_vpopcountq_v8di_mask): New.
2290 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2291 __AVX512VPOPCNTDQ__.
2292 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2293 (PTA_AVX512VPOPCNTDQ): Define.
2294 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2295 TARGET_AVX512VPOPCNTDQ_P): Define.
2296 * config/i386/i386.opt: Add mavx512vpopcntdq.
2297 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2298 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2300 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2303 * predict.def (PRED_CALL): Set to 67.
2305 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2307 * expr.c (store_field): In the bitfield case, if the value comes from
2308 a function call and is of an aggregate type returned in registers, do
2309 not modify the field mode; extract the value in all cases if the mode
2310 is BLKmode and the size is not larger than a word.
2312 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2315 * config/i386/cpuid.h: Fix undefined behavior.
2317 2017-01-04 Jeff Law <law@redhat.com>
2319 PR tree-optimization/79007
2320 PR tree-optimization/67955
2321 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2322 conservative for pt.null when flag_non_call_exceptions is on.
2324 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2326 PR translation/79019
2327 PR translation/79020
2328 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2329 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2330 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2332 * config/avr/avr.opt (maccumulate-args): Likewise.
2333 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2334 * common.opt (freport-bug): Likewise.
2335 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2336 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2337 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2338 translatable string.
2339 * config/i386/i386.c (function_value_32): Likewise.
2340 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2341 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2343 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2344 * common/config/msp430/msp430-common.c (msp430_handle_option):
2346 * symtab.c (symtab_node::verify_base): Likewise.
2347 * opts.c (set_debug_level): Likewise.
2348 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2349 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2350 missing whitespace to translatable strings.
2351 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2352 * config/sh/superh.h: Likewise.
2353 * config/i386/xopintrin.h: Likewise.
2354 * config/i386/znver1.md: Likewise.
2355 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2356 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2357 * double-int.h (struct double_int): Likewise.
2358 * double-int.c (div_and_round_double): Likewise.
2359 * wide-int.cc: Likewise.
2360 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2361 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2362 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2363 (crossjumps_occurred): ... this.
2364 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2367 PR tree-optimization/78899
2368 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2369 returning bool return struct loop *, NULL for failure and the new
2371 (versionable_outer_loop_p): Don't version outer loop if it has
2372 dont_vectorized bit set.
2373 (tree_if_conversion): When versioning outer loop, ensure
2374 tree_if_conversion is performed also on the inner loop of the
2375 non-vectorizable outer loop copy.
2376 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2377 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2378 prevent vectorization of it.
2379 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2380 the outer loop vectorization of the non-scalar version is attempted
2381 before vectorization of the inner loop in scalar version. If
2382 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2383 vectorization of its inner loop.
2384 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2385 has 2 inner loops, rename also on edges from bb whose single pred
2386 is outer_loop->header. Fix typo in function comment.
2388 2017-01-09 Martin Sebor <msebor@redhat.com>
2391 * asan.c (asan_emit_stack_protection): Increase local buffer size
2392 to avoid snprintf truncation warning.
2394 2017-01-09 Andrew Pinski <apinski@cavium.com>
2396 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2397 to reference thunderx2t99 for the tuning structure
2398 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2400 (thunderx2t99_extra_costs): This.
2401 * config/aarch64/aarch64-tune.md: Regenerate.
2402 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2403 (vulcan_addrcost_table): This.
2404 (vulcan_regmove_cost): Rename to ...
2405 (thunderx2t99_regmove_cost): This.
2406 (vulcan_vector_cost): Rename to ...
2407 (thunderx2t99_vector_cost): this.
2408 (vulcan_branch_cost): Rename to ...
2409 (thunderx2t99_branch_cost): This.
2410 (vulcan_tunings): Rename to ...
2411 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2412 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2414 2017-01-09 Martin Jambor <mjambor@suse.cz>
2418 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2419 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2420 (propagate_vr_accross_jump_function): Use the above function for all
2421 value range computations for pass-through jump functions and type
2422 converasion from explicit value range values.
2423 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2424 parameters from TYPE_ARG_TYPES.
2425 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2426 (ipa_write_node_info): Stream type of the actual argument.
2427 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2429 2017-01-09 Martin Liska <mliska@suse.cz>
2432 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2433 (lookup_compiler): Do not show error message with have_E.
2435 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2437 PR tree-optimization/78938
2438 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2439 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2440 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2441 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2444 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2446 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2449 2017-01-09 Richard Biener <rguenther@suse.de>
2451 PR tree-optimization/78997
2452 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2453 name condition properly.
2455 2017-01-09 Richard Biener <rguenther@suse.de>
2458 * dwarf2out.c (is_cxx): New overload with context.
2459 (is_naming_typedef_decl): Use it.
2461 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2463 * invoke.texi (Option Summary): Correct spacing in option lists
2464 and add line breaks to fix over-long lines.
2466 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2470 * extend.texi (Common Variable Attributes): Add xref to GCC
2471 Internals manual to explain mode attribute keywords.
2473 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2476 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2477 and Preprocessor Options.
2478 (Options for Linking): Document -pthread here....
2479 (RS/6000 and PowerPC Options): ...not here.
2480 (Solaris 2 Options): ...or here.
2481 * doc/cppopts.texi: Document -pthread.
2483 2017-01-08 Martin Sebor <msebor@redhat.com>
2486 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2487 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2488 New member functions.
2489 (format_directive): Used them.
2491 (pass_sprintf_length::handle_gimple_call): Same.
2492 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2493 to avoid truncation for any argument.
2494 (extract_affine_mul): Same.
2495 * tree.c (get_file_function_name): Same.
2497 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2500 * predict.def (PRED_INDIR_CALL): Set to 86.
2502 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2504 PR preprocessor/54124
2505 * doc/cppopts.texi: Reformat -d subtable to list the full name
2506 of the options. Add cross-reference to the docs for the general
2507 compiler -d options.
2508 * doc/invoke.texi (Developer Options): Add cross-reference to the
2509 preprocessor-specific -d option documentation.
2511 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2513 PR preprocessor/13498
2514 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2515 redudant material, and reflect new command-line options.
2516 (System Headers): Likewise.
2518 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2520 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2521 -isystem, and -idirafter. Copy-edit.
2522 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2523 default for -ftrack-macro-expansion. Delete obsolete and
2524 badly-formatted implementation details about -fdebug-cpp output.
2525 * doc/cppwarnopts.texi: Copy-edit.
2527 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2530 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2531 that the transition from a max line width >= 1<<10 to narrower
2532 lines works correctly.
2534 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2536 * doc/options.texi (PerFunction): New.
2537 * opt-functions.awk (switch_flags): Map both Optimization and
2538 PerFunction to CL_OPTIMIZATION.
2539 * opth-gen.awk: Test for PerFunction flag along with
2541 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2542 it only when the latter is present. Skip those that don't in
2543 the hash function generator.
2544 * common.opt (fvar-tracking): Mark as PerFunction instead of
2546 (fvar-tracking-assignments): Likewise.
2547 (fvar-tracking-assignments-toggle): Likewise.
2548 (fvar-tracking-uninit): Likewise.
2550 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2552 PR translation/79018
2553 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2556 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2559 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2560 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2561 TARGET_LONG_JUMP_TABLE_OFFSETS.
2562 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2563 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2564 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2565 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2566 * config/m68k/m68k.md (tablejump expander): Likewise.
2567 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2568 TARGET_LONG_JUMP_TABLE_OFFSETS.
2569 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2570 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2572 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2573 David Holsgrove <david.holsgrove@xilinx.com>
2575 * common/config/microblaze/microblaze-common.c
2576 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2577 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2579 * config/microblaze/microblaze.c (microblaze_must_save_register)
2580 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2582 (microblaze_eh_return): New function.
2583 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2584 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2585 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2586 * config/microblaze/microblaze.md (eh_return): New pattern.
2588 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2590 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2591 GCC_DIAGNOSTIC_STRINGIFY): Define.
2593 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2595 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2597 * config/arm/arm.md (<mcrr>): New.
2599 * config/arm/arm.c (arm_arch5te): New.
2600 (arm_option_override): Set arm_arch5te.
2601 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2603 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2604 (arm_mcrr_qualifiers): ... this. New.
2605 (MRRC_QUALIFIERS): Define to...
2606 (arm_mrrc_qualifiers): ... this. New.
2607 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2609 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2610 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2611 (MRRCI, mrrc, MRRC): New.
2612 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2613 VUNSPEC_MRRC2): New.
2615 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2617 * config/arm/arm.md (<mcr>): New.
2619 * config/arm/arm.c (arm_coproc_builtin_available): Add
2620 support for mcr, mrc, mcr2 and mrc2.
2621 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2622 (arm_mcr_qualifiers): ... this. New.
2623 (MRC_QUALIFIERS): Define to ...
2624 (arm_mrc_qualifiers): ... this. New.
2625 (MCR_QUALIFIERS): Define to ...
2626 (arm_mcr_qualifiers): ... this. New.
2627 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2629 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2630 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2631 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2634 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2636 * config/arm/arm.md (*ldc): New.
2640 * config/arm/arm.c (arm_coproc_builtin_available): Add
2641 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2642 (arm_coproc_ldc_stc_legitimate_address): New.
2643 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2644 'qualifier_const_pointer'.
2645 (LDC_QUALIFIERS): Define to...
2646 (arm_ldc_qualifiers): ... this. New.
2647 (STC_QUALIFIERS): Define to...
2648 (arm_stc_qualifiers): ... this. New.
2649 * config/arm/arm-protos.h
2650 (arm_coproc_ldc_stc_legitimate_address): New.
2651 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2652 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2653 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2654 stc2, stcl, stc2l): New.
2655 * config/arm/constraints.md (Uz): New.
2656 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2657 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2658 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2659 VUNSPEC_STC2L): New.
2661 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2663 * config/arm/arm.md (<cdp>): New.
2664 * config/arm/arm.c (neon_const_bounds): Rename this ...
2665 (arm_const_bounds): ... this.
2666 (arm_coproc_builtin_available): New.
2667 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2668 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2669 (CDP_QUALIFIERS): Define to...
2670 (arm_cdp_qualifiers): ... this. New.
2672 (arm_expand_builtin_args): Add case for 6 arguments.
2673 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2674 (arm_const_bounds): ... this.
2675 (arm_coproc_builtin_available): New.
2676 * config/arm/arm_acle.h (__arm_cdp): New.
2678 * config/arm/arm_acle_builtins.def (cdp): New.
2680 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2681 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2683 * config/arm/types.md (coproc): New.
2684 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2685 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2686 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2687 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2689 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2691 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2692 (UBINOP_QUALIFIERS): New.
2694 (acle_builtin_data): New. Change comment.
2695 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2696 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2697 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2698 arm_acle_builtins.def.
2699 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2700 (arm_init_acle_builtins): New.
2701 (CRC32_BUILTIN): Remove.
2702 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2703 crc32cb, crc32ch and crc32cw.
2704 (arm_init_crc32_builtins): Remove.
2705 (arm_init_builtins): Use arm_init_acle_builtins rather
2706 than arm_init_crc32_builtins.
2707 (arm_expand_acle_builtin): New.
2708 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2709 * config/arm/arm_acle_builtins.def: New.
2711 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2713 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2714 (arm_builtin_datum): ... this.
2715 (arm_init_neon_builtin): Rename to ...
2716 (arm_init_builtin): ... this. Add a new parameters PREFIX
2717 and USE_SIG_IN_NAME.
2718 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2719 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2720 'arm_builtin_datum'.
2721 (arm_init_vfp_builtins): Likewise.
2722 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2723 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2724 (arm_expand_neon_args): Rename to ...
2725 (arm_expand_builtin_args): ... this. Rename builtin_arg
2726 enum values and differentiate between ARG_BUILTIN_MEMORY
2727 and ARG_BUILTIN_NEON_MEMORY.
2728 (arm_expand_neon_builtin_1): Rename to ...
2729 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2730 values, arm_expand_builtin_args and add bool parameter NEON.
2731 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2732 (arm_expand_vfp_builtin): Likewise.
2733 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2735 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2738 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2739 * predict.c (tree_estimate_probability_bb): Reverse direction of
2740 polymorphic call predictor.
2742 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2744 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2745 (determine_pass_name_match): ...this new function and...
2746 (should_skip_pass_p): ...this new function.
2748 2017-01-06 Nathan Sidwell <nathan@acm.org>
2750 * ipa-visibility.c (function_and_variable_visibility): Reformat
2751 comments and long lines. Remove extrneous if.
2752 * symtab.c (symtab_node::make_decl_local): Fix code format.
2753 (symtab_node::set_section_for_node): Fix comment typo.
2755 2017-01-06 Martin Liska <mliska@suse.cz>
2758 * lra-constraints.c: Rename invariant to lra_invariant.
2759 * predict.c (set_even_probabilities): Initialize e to NULL.
2761 2017-01-05 Martin Sebor <msebor@redhat.com>
2763 PR tree-optimization/78910
2764 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2765 (format_integer): Correct off-by-one error in the handling
2766 of precision with negative numbers in signed conversions..
2768 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2770 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2772 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2774 PR tree-optimization/71016
2775 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2776 factor_out_conditional_conversion. Formatting fix.
2777 (factor_out_conditional_conversion): Add cond_stmt argument.
2778 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2779 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2782 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2784 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2785 read-rtl-function.o, and selftest-rtl.o.
2786 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2787 (selftest::aarch64_test_loading_full_dump): New function.
2788 (selftest::aarch64_run_selftests): New function.
2789 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2790 selftest::aarch64_run_selftests.
2791 * config/i386/i386.c
2792 (selftest::ix86_test_loading_dump_fragment_1): New function.
2793 (selftest::ix86_test_loading_call_insn): New function.
2794 (selftest::ix86_test_loading_full_dump): New function.
2795 (selftest::ix86_test_loading_unspec): New function.
2796 (selftest::ix86_run_selftests): Call the new functions.
2797 * emit-rtl.c (maybe_set_max_label_num): New function.
2798 * emit-rtl.h (maybe_set_max_label_num): New decl.
2799 * function.c (instantiate_decls): Guard call to
2800 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2801 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2803 * gensupport.c (gen_reader::gen_reader): Pass "false"
2804 for new "compact" param of rtx_reader.
2805 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2806 rather than an empty string for NULL strings.
2807 * read-md.c: Potentially include config.h rather than bconfig.h.
2808 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2809 (have_error): New global, copied from errors.c.
2810 (md_reader::read_name): Rename to...
2811 (md_reader::read_name_1): ...this, adding "out_loc" param,
2812 and converting "missing name or number" to returning false, rather
2814 (md_reader::read_name): Reimplement in terms of read_name_1.
2815 (md_reader::read_name_or_nil): New function.
2816 (md_reader::read_string): Handle "(nil)" by returning NULL.
2817 (md_reader::md_reader): Add new param "compact".
2818 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2819 (md_reader::read_file): New method.
2820 * read-md.h (md_reader::md_reader): Add new param "compact".
2821 (md_reader::read_file): New method.
2822 (md_reader::is_compact): New accessor.
2823 (md_reader::read_name): Convert return type from void to file_location.
2824 (md_reader::read_name_or_nil): New decl.
2825 (md_reader::read_name_1): New decl.
2826 (md_reader::m_compact): New field.
2827 (noop_reader::noop_reader): Pass "false" for new "compact" param
2829 (rtx_reader::rtx_reader): Add new "compact" param.
2830 (rtx_reader::read_rtx_operand): Make virtual and convert return
2831 type from void to rtx.
2832 (rtx_reader::read_until): New decl.
2833 (rtx_reader::handle_any_trailing_information): New virtual function.
2834 (rtx_reader::postprocess): New virtual function.
2835 (rtx_reader::finalize_string): New virtual function.
2836 (rtx_reader::m_in_call_function_usage): New field.
2837 (rtx_reader::m_reuse_rtx_by_id): New field.
2838 * read-rtl-function.c: New file.
2839 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2840 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2841 (selftest::verify_three_block_rtl_cfg): New decl.
2842 * read-rtl-function.h: New file.
2843 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2844 For host, include function.h, memmodel.h, and emit-rtl.h.
2845 (one_time_initialization): New function.
2846 (struct compact_insn_name): New struct.
2847 (compact_insn_names): New array.
2848 (find_code): Handle insn codes in compact dumps.
2849 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2850 (bind_subst_iter_and_attr): Likewise.
2851 (add_condition_to_string): Likewise.
2852 (add_condition_to_rtx): Likewise.
2853 (apply_attribute_uses): Likewise.
2854 (add_current_iterators): Likewise.
2855 (apply_iterators): Likewise.
2856 (initialize_iterators): Guard usage of apply_subst_iterator with
2857 #ifdef GENERATOR_FILE.
2858 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2859 (md_reader::read_mapping): Likewise.
2860 (add_define_attr_for_define_subst): Likewise.
2861 (add_define_subst_attr): Likewise.
2862 (read_subst_mapping): Likewise.
2863 (check_code_iterator): Likewise.
2864 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2866 (one_time_initialization): New function.
2867 (rtx_reader::read_until): New method.
2868 (read_flags): New function.
2869 (parse_reg_note_name): New function.
2870 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2871 Handle reuse_rtx ids.
2872 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2873 Add parsing support for RTL dumps, mirroring the special-cases in
2874 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2875 values, and calling handle_any_trailing_information.
2876 (rtx_reader::read_rtx_operand): Convert return type from void
2877 to rtx, returning return_rtx. Handle case 'e'. Call
2878 finalize_string on XSTR and XTMPL fields.
2879 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2880 "(nil)" values were omitted. Call the postprocess vfunc on the
2882 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2883 class ctor. Initialize m_in_call_function_usage. Call
2884 one_time_initialization.
2885 * rtl-tests.c (selftest::test_uncond_jump): Call
2886 set_new_first_and_last_insn.
2887 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2888 * selftest-rtl.c: New file.
2889 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2890 (selftest::get_insn_by_uid): New decl.
2891 * selftest-run-tests.c (selftest::run_tests): Call
2892 read_rtl_function_c_tests.
2893 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2894 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2897 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2899 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2900 operands in a special way. Assert that pos+len <= mode precision.
2902 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2904 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2905 3 argument Alias with unlimited for the negative form.
2906 (fno-vect-cost-model): Removed.
2908 2017-01-05 Martin Liska <mliska@suse.cz>
2910 * hsa-gen.c (gen_hsa_divmod): New function.
2911 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2913 2017-01-05 Martin Liska <mliska@suse.cz>
2916 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2919 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2921 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2922 small constant length operands.
2924 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2926 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2927 between loop iterations.
2929 2017-01-05 Martin Liska <mliska@suse.cz>
2932 * gimplify.c (gimplify_decl_expr): Compare to
2933 asan_poisoned_variables instread of checking flags.
2934 (gimplify_target_expr): Likewise.
2935 (gimplify_expr): Likewise.
2936 (gimplify_function_tree): Conditionally initialize
2937 asan_poisoned_variables.
2939 2017-01-04 Jeff Law <law@redhat.com>
2941 PR tree-optimizatin/78812
2942 * rtl.h (contains_mem_rtx_p): Prototype.
2943 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2944 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2945 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2946 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2947 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2949 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2951 * input.c (assert_char_at_range): Default-initialize actual_range.
2953 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2955 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2956 to match the caller.
2958 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2960 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2961 insns after final jump in test to emit dummy move.
2963 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2965 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2966 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2968 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2970 * multiple_target.c (create_dispatcher_calls): Init e_next.
2971 * tree-ssa-loop-split.c (split_loop): Init border.
2972 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2975 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2980 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2981 (altivec_register_operand): Do not return true if the operand
2982 contains a SUBREG mixing SImode and SFmode.
2983 (vsx_register_operand): Likewise.
2984 (vsx_reg_sfsubreg_ok): New predicate.
2985 (vfloat_operand): Do not return true if the operand contains a
2986 SUBREG mixing SImode and SFmode.
2987 (vint_operand): Likewise.
2988 (vlogical_operand): Likewise.
2989 (gpc_reg_operand): Likewise.
2990 (int_reg_operand): Likewise.
2991 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2992 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2993 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2995 (rs6000_emit_move_si_sf_subreg): New helper function.
2996 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2997 fixup SUBREGs involving SImode and SFmode.
2998 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2999 numbers for the new peephole2 optimization.
3000 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3001 the GLIBC math library that do AND/IOR/XOR operations on single
3002 precision floating point.
3003 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3004 target macros to say whether we need to avoid SUBREGs mixing
3006 (TARGET_ALLOW_SF_SUBREG): Likewise.
3007 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3008 (UNSPEC_SI_FROM_SF): Likewise.
3009 (iorxor): Change spacing.
3010 (and_ior_xor): New iterator for AND, IOR, and XOR.
3011 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3012 (movdi_from_sf_zero_ext): Likewise.
3013 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3014 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3015 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3016 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3017 (fms<mode>4): Likewise.
3018 (fnma<mode>4): Likewise.
3019 (fnms<mode>4): Likewise.
3020 (nfma<mode>4): Likewise.
3021 (nfms<mode>4): Likewise.
3023 2017-01-04 Marek Polacek <polacek@redhat.com>
3026 * doc/invoke.texi: Document -Wpointer-compare.
3028 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3030 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3033 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3034 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3037 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3040 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3041 documentation of the powerpc_popcntb_ok attribute.
3042 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3043 code to issue warning messages if a requested CPU configuration is
3044 not supported by the binary (assembler and loader) toolchain.
3045 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3046 made to define a built-in function that has been disabled.
3047 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3048 made to define a built-in function that has been disabled.
3049 (altivec_init_builtins): Add comment explaining why definition
3050 of the DST built-in functions is not preceded by an assertion
3051 check. Add assertions to prevent ICE if attempts are made to
3052 define an altivec predicate or an abs* built-in function that has
3054 (htm_init_builtins): Add comment explaining why definition of the
3055 htm built-in functions is not preceded by an assertion check.
3057 2017-01-04 Jeff Law <law@redhat.com>
3059 PR tree-optimizatin/67955
3060 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3061 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3062 the points-to solution does not include pt_null. Use DECL_PT_UID
3065 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3067 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3068 Use gen_int_mode instead of gen_lopwart for const_int operands.
3070 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3072 PR tree-optimization/71563
3073 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3074 out of range value - has low bits known to be zero.
3076 2017-01-04 Alan Modra <amodra@gmail.com>
3078 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3079 * configure: Regenerate.
3080 * config.in: Regenerate.
3082 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3085 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3086 a substring of the message, but strcmp with the whole message. Ifdef
3087 ENABLE_NLS, translate the message first using dgettext.
3089 2017-01-03 Jeff Law <law@redhat.com>
3091 PR tree-optimizatin/78856
3092 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3093 (mark_threaded_blocks): Remove code to truncate thread paths that
3094 cross multiple loop headers. Instead invalidate the cached loop
3095 iteration information and handle case of a thread path walking
3096 into an irreducible region.
3098 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3101 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3102 assertions. Add support for doing the signbit if the IEEE 128-bit
3103 floating point value is in a GPR.
3104 * config/rs6000/rs6000.md (Fsignbit): Delete.
3105 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3106 Update the length attribute if the value is in a GPR.
3107 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3108 the sign or zero extension instruction, since the value is always 0/1.
3109 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3112 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3113 extracting SImode to a GPR register so that we can generate a
3114 store, limit the vector to be in a traditional Altivec register
3115 for the vextuwrx instruction.
3117 2017-01-03 Ian Lance Taylor <iant@google.com>
3119 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3121 2017-01-03 Martin Sebor <msebor@redhat.com>
3123 PR tree-optimization/78696
3124 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3125 precision. Use MPFR for %f for greater fidelity. Correct handling
3127 (pass_sprintf_length::compute_format_length): Set width and precision
3128 specified by asrerisk to void_node for vararg functions.
3129 (try_substitute_return_value): Adjust dump output.
3131 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3133 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3135 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3137 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3138 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3139 -mlra/-mno-lra was passed to the compiler.
3141 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3143 PR rtl-optimization/65618
3144 * emit-rtl.c (try_split): Move initialization of "before" and
3145 "after" to just before the call to emit_insn_after_setloc.
3147 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3149 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3151 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3153 * dwarf2out.c (gen_enumeration_type_die): When
3154 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3156 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3158 PR tree-optimization/78965
3159 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3160 Change first argument from const call_info & to call_info &. For %n
3161 set info.nowrite to false.
3164 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3165 possibly throwing calls.
3167 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3168 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3169 and fns handling, rather than in a separate case SSA_NAME.
3171 2017-01-02 Jeff Law <law@redhat.com>
3173 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3174 fixes for first_period and second_period variables.
3176 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3179 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3180 (*insvqi_1): New insn pattern.
3181 (*insvqi_1_mem_rex64): Ditto.
3183 (*insvqi_3): Rename from *insvqi.
3185 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3187 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3189 * doc/cfg.texi (Edges): Remove reference to Java.
3190 (Maintaining the CFG): Ditto.
3192 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3195 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3196 transparent aliases.
3198 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3201 * predict.def (PRED_CALL): Update hitrate.
3202 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3203 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3204 into direct/indirect/polymorphic variants.
3206 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3208 Update copyright years.
3210 * gcc.c (process_command): Update copyright notice dates.
3211 * gcov-dump.c (print_version): Ditto.
3212 * gcov.c (print_version): Ditto.
3213 * gcov-tool.c (print_version): Ditto.
3214 * gengtype.c (create_file): Ditto.
3215 * doc/cpp.texi: Bump @copying's copyright year.
3216 * doc/cppinternals.texi: Ditto.
3217 * doc/gcc.texi: Ditto.
3218 * doc/gccint.texi: Ditto.
3219 * doc/gcov.texi: Ditto.
3220 * doc/install.texi: Ditto.
3221 * doc/invoke.texi: Ditto.
3223 Copyright (C) 2017 Free Software Foundation, Inc.
3225 Copying and distribution of this file, with or without modification,
3226 are permitted in any medium without royalty provided the copyright
3227 notice and this notice are preserved.