1 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
4 * lra-constraints.c (curr_insn_transform): Handle
5 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
7 2017-02-19 Uros Bizjak <ubizjak@gmail.com>
10 2016-05-30 Uros Bizjak <ubizjak@gmail.com>
12 * config/i386/sync.md (mfence_nosse): Use "lock orl $0, -4(%esp)".
14 2017-02-19 Jonathan Wakely <jwakely@redhat.com>
17 * doc/invoke.texi (C++ Dialect Options) [-Wliteral-suffix]: Update
20 2017-02-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
22 * gimple-pretty-print.c (dump_ternary_rhs): Adjust gimple dump format
25 2017-02-18 Jakub Jelinek <jakub@redhat.com>
27 * final.c (last_columnnum, override_columnnum): New variables.
28 (final_start_function): Set last_columnnum, pass it to begin_prologue
29 hook and pass 0 to dwarf2out_begin_prologue.
30 (final_scan_insn): Update override_columnnum. Pass last_columnnum
31 to source_line debug hook.
32 (notice_source_line): Compute last_columnnum and for debug_column_info
33 return true on column changes.
34 * debug.h (struct gcc_debug_hooks): Add column argument to
35 source_line and begin_prologue hooks.
36 (debug_nothing_int_charstar_int_bool): Remove prototype.
37 (debug_nothing_int_int_charstar,
38 debug_nothing_int_int_charstar_int_bool): New prototypes.
39 (dwarf2out_begin_prologue): Add column argument.
40 * debug.c (do_nothing_debug_hooks): Adjust source_line and
42 (debug_nothing_int_charstar_int_bool): Remove.
43 (debug_nothing_int_int_charstar,
44 debug_nothing_int_int_charstar_int_bool): New functions.
45 * dwarf2out.c (dwarf2out_begin_prologue): Add column argument, pass it
46 through to dwarf2out_source_line.
47 (dwarf2_lineno_debug_hooks): Adjust begin_prologue hook.
48 (dwarf2out_source_line): Add column argument, emit it if requested.
49 * sdbout.c (sdbout_source_line, sdbout_begin_prologue): Add column
51 * xcoffout.h (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
52 * xcoffout.c (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
53 * vmsdbgout.c (vmsdbgout_begin_prologue): Add column argument, pass it
54 through to dwarf2out_begin_prologue.
55 (vmsdbgout_source_line): Add column argument, pass it through to
56 dwarf2out_source_line.
57 * dbxout.c (dbxout_begin_prologue): Add column argument, adjust
58 dbxout_source_line caller.
59 (dbxout_source_line): Add column argument.
61 * common.opt (gno-column-info, gcolumn-info): New options.
62 * dwarf2out.c (dwarf2_lineno_debug_hooks): Formatting fix.
63 (check_die): Also test for multiple DW_AT_decl_column attributes.
64 (add_src_coords_attributes, dwarf2out_imported_module_or_decl_1): Add
65 DW_AT_decl_column if requested.
66 (gen_subprogram_die): Compare and/or add also DW_AT_decl_column
68 (gen_variable_die): Likewise.
69 (add_call_src_coords_attributes): Add DW_AT_call_column if requested.
70 * doc/invoke.texi (-gcolumn-info, -gno-column-info): Document.
73 * config/i386/i386.opt (m3dnowa): Replace Undocumented with Report.
74 * common/config/i386/i386-common.c (OPTION_MASK_ISA_3DNOW_A_SET): Define.
75 (ix86_handle_option): Handle OPT_m3dnowa.
76 * doc/invoke.texi (-m3dnowa): Document.
77 * doc/extend.texi (__builtin_ia32_pmulhuw, __builtin_ia32_pf2iw): Use
78 -m3dnowa instead of -m3dnow -march=athlon.
81 * config/i386/i386.c (ix86_print_operand): Use output_operand_lossage
82 instead of gcc_assert for K, r and R code checks. Formatting fixes.
84 2017-02-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
87 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
88 support for CODE_FOR_vsx_xxpermdi_v2d[fi]_be.
89 * config/rs6000/rs6000.md (reload_gpr_from_vsx<mode>): Call
90 generator for vsx_xxpermdi_<mode>_be.
91 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Remove logic to
92 force big-endian semantics.
93 (vsx_xxpermdi_<mode>_be): New define_expand with same
94 implementation as previous version of vsx_xxpermdi_<mode>.
96 2017-02-17 Jakub Jelinek <jakub@redhat.com>
98 PR tree-optimization/79327
99 * gimple-ssa-sprintf.c (format_integer): Remove likely_adjust
100 variable, its initialization and use.
102 2017-02-17 Julia Koval <julia.koval@intel.com>
104 * common/config/i386/i386-common.c (OPTION_MASK_ISA_RDPID_SET): New.
105 (OPTION_MASK_ISA_PKU_UNSET): New.
106 (ix86_handle_option): Handle -mrdpid.
107 * config/i386/cpuid.h (bit_RDPID): New.
108 * config/i386/driver-i386.c (host_detect_local_cpu):
109 Detect RDPID feature.
110 * config/i386/i386-builtin.def (__builtin_ia32_rdpid): New.
111 * config/i386/i386-c.c (ix86_target_macros_internal):
113 * config/i386/i386.c (ix86_target_string): Add -mrdpid to isa2_opts.
114 (ix86_valid_target_attribute_inner_p): Add "rdpid".
115 (ix86_expand_builtin): Handle IX86_BUILTIN_RDPID.
116 * config/i386/i386.h (TARGET_RDPID, TARGET_RDPID_P): New.
117 * config/i386/i386.md (define_insn "rdpid"): New.
118 * config/i386/i386.opt Add -mrdpid.
119 * config/i386/immintrin.h (_rdpid_u32): New.
121 2017-02-17 Vladimir Makarov <vmakarov@redhat.com>
123 PR rtl-optimization/79541
124 * lra-constraints.c (curr_insn_transform): Remove wrong asm insn
125 instead of transforming it into USE.
127 2017-02-17 Segher Boessenkool <segher@kernel.crashing.org>
129 * config/rs6000/rs6000.md (extendsfdf2): Remove default arguments.
130 If HONOR_SNANS (SFmode) force the input to a register.
131 (*extendsfdf2_fpr): Add !HONOR_SNANS (SFmode) condition.
132 (*extendsfdf2_snan): New pattern, used when using SNaNs; it generates
133 an frsp or similar insn.
135 2017-02-17 Martin Liska <mliska@suse.cz>
137 PR rtl-optimization/79577
138 * params.def (selsched-max-sched-times): Increase minimum to 1.
140 2017-02-17 Martin Liska <mliska@suse.cz>
142 PR rtl-optimization/79574
143 * gcse.c (want_to_gcse_p): Prevent integer overflow.
145 2017-02-17 Martin Liska <mliska@suse.cz>
147 PR tree-optimization/79529
148 * tree-ssa-loop-unswitch.c (is_maybe_undefined): Use
149 ssa_defined_default_def_p to handle cases which are implicitly
151 * tree-ssa.c (ssa_defined_default_def_p): New function.
152 (ssa_undefined_value_p): Use ssa_defined_default_def_p to handle cases
153 which are implicitly defined.
154 * tree-ssa.h (ssa_defined_default_def_p): Declare.
156 2017-02-17 Richard Biener <rguenther@suse.de>
159 * params.def (max-ssa-name-query-depth): Limit to 10.
161 2017-02-17 Richard Biener <rguenther@suse.de>
163 PR tree-optimization/79552
164 * tree-ssa-structalias.c (visit_loadstore): Properly verify
167 2017-02-17 Richard Biener <rguenther@suse.de>
170 * genmatch.c (output_line_directive): Handle DIR_SEPARATOR_2.
172 2017-02-17 Marek Polacek <polacek@redhat.com>
175 * fold-const.c (fold_negate_expr_1): Renamed from fold_negate_expr.
176 (fold_negate_expr): New wrapper.
178 2017-02-16 Sandra Loosemore <sandra@codesourcery.com>
180 * doc/invoke.texi (C++ Dialect Options) [-Wno-non-template-friend]:
181 Correct terminology and de-emphasize pre-standard behavior.
183 2017-02-16 Alan Modra <amodra@gmail.com>
185 PR rtl-optimization/79286
186 * ira.c (def_dominates_uses): New function.
187 (update_equiv_regs): Don't create an equivalence for insns that
188 may trap where the register def does not dominate the use.
190 2017-02-16 Vladimir Makarov <vmakarov@redhat.com>
192 PR rtl-optimization/78127
193 * lra.c (lra): Call lra_eliminate before finish the loop after
196 2017-02-16 Richard Biener <rguenther@suse.de>
198 * graphite.h: Do not include isl/isl_val_gmp.h, instead include
200 * graphite-isl-ast-to-gimple.c (gmp_cst_to_tree): Remove.
201 (gcc_expression_from_isl_expr_int): Use generic isl_val interface.
202 * graphite-sese-to-poly.c: Do not include isl/isl_val_gmp.h.
203 (isl_val_int_from_wi): New function.
204 (extract_affine_gmp): Rename to ...
205 (extract_affine_wi): ... this, take a widest_int.
206 (extract_affine_int): Just wrap extract_affine_wi.
207 (add_param_constraints): Use isl_val_int_from_wi.
208 (add_loop_constraints): Likewise, and extract_affine_wi.
210 2017-02-15 Jeff Law <law@redhat.com>
213 * ira-costs.c (scan_one_insn): Check have_regs_of_mode before calling
214 ira_init_register_move_cost_if_necessary.
216 2017-02-15 Martin Sebor <msebor@redhat.com>
219 * doc/invoke.texi (-fdump-final-insns): Replace option accidentally
220 removed in a prior commit.
222 2017-02-15 Bin Cheng <bin.cheng@arm.com>
224 PR tree-optimization/79347
225 * tree-vect-loop-manip.c (vect_do_peeling): Maintain profile
226 counters during peeling.
228 2017-02-15 Thomas Schwinge <thomas@codesourcery.com>
230 * Makefile.in (site.exp): Remove "set ISLVER".
232 2017-02-15 Jakub Jelinek <jakub@redhat.com>
235 * real.c (real_from_integer): Call real_convert even for decimal.
237 2017-02-15 Dominik Vogt <vogt@linux.vnet.ibm.com>
240 * config/s390/s390.c: define TARGET_CUSTOM_FUNCTION_DESCRIPTORS.
242 2017-02-14 Andrew Pinski <apinski@cavium.com>
244 * config/aarch64/aarch64-cores.def (thunderx2t99): Move to under 'C"
245 cores and change the partno/implementer to be correct.
246 (thunderx2t99p1): New core which replaces thunderx2t99 and still has
247 the 'B" as the implementer.
248 * config/aarch64/aarch64-tune.md: Regenerate.
250 2017-02-14 Carl Love <cel@us.ibm.com>
252 * config/rs6000/rs6000.c: Add case statement entry to make the
253 xvcvuxdsp built-in argument unsigned.
254 * config/rs6000/vsx.md: Fix the source and return operand types so they
255 match the instruction definitions from the ISA document. Fix typo
256 in the instruction generation for the (define_insn "vsx_xvcvuxdsp"
259 2017-02-14 Vladimir Makarov <vmakarov@redhat.com>
262 * lra-int.h (struct lra_operand_data, struct lra_insn_reg): Add
263 member early_clobber_alts.
264 * lra-lives.c (reg_early_clobber_p): New.
265 (process_bb_lives): Use it.
266 * lra.c (new_insn_reg): New arg early_clobber_alts. Use it.
267 (debug_operand_data): Initialize early_clobber_alts.
268 (setup_operand_alternative): Set up early_clobber_alts.
269 (collect_non_operand_hard_regs): Ditto. Pass early clobber
270 alternatives to new_insn_reg.
271 (add_regs_to_insn_regno_info): Add arg early_clobber_alts. Use
273 (lra_update_insn_regno_info): Pass the new arg.
275 2017-02-14 Jakub Jelinek <jakub@redhat.com>
278 * omp-offload.c (free_oacc_loop): Release loop->ifns vector.
279 (new_oacc_loop_raw): Don't clear already cleared fields.
282 * config/i386/avx512pfintrin.h (_mm512_prefetch_i32gather_pd,
283 _mm512_prefetch_i32gather_ps, _mm512_prefetch_i64gather_pd,
284 _mm512_prefetch_i64gather_ps): New inline functions and macros.
286 2017-02-14 Uros Bizjak <ubizjak@gmail.com>
289 * config/i386/i386.md (*movxf_internal): Add (o,rC) alternative.
291 2017-02-14 H.J. Lu <hongjiu.lu@intel.com>
294 * config/i386/i386.c (timode_scalar_chain::convert_insn): Insert
295 the extra instruction to the right place to store 128-bit constant
298 2017-02-14 Martin Sebor <msebor@redhat.com>
301 * gimple-ssa-sprintf.c (format_directive): Avoid issuing INT_MAX
302 warning for strings of unknown length.
304 2017-02-13 Segher Boessenkool <segher@kernel.crashing.org>
306 * config.gcc (supported_defaults) [powerpc*-*-*]: Update.
308 2017-02-14 Jeff Law <law@redhat.com>
311 * ira-costs.c (scan_one_insn): Initialize register move costs
312 for pseudos seen in USE/CLOBBER insns.
314 PR tree-optimization/79095
315 * tree-vrp.c (extract_range_from_binary_expr_1): For EXACT_DIV_EXPR,
316 if the numerator has the range ~[0,0] make the resultant range ~[0,0].
317 (extract_range_from_binary_expr): For MINUS_EXPR with no derived range,
318 if the operands are known to be not equal, then the resulting range
320 (intersect_ranges): If the new range is ~[0,0] and the old range is
321 wide, then prefer ~[0,0].
322 * tree-vrp.c (overflow_comparison_p_1): New function.
323 (overflow_comparison_p): New function.
324 * tree-vrp.c (register_edge_assert_for_2): Register additional asserts
325 if NAME is used in an overflow test.
326 (vrp_evaluate_conditional_warnv_with_ops): If the ops represent an
327 overflow check that can be expressed as an equality test, then adjust
328 ops to be that equality test.
330 2017-02-14 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
332 * config/s390/s390-builtin-types.def: Remove flags argument.
333 * config/s390/s390.c (s390_init_builtins): Likewise.
335 2017-02-14 Martin Liska <mliska@suse.cz>
337 * tree-ssa-loop-unswitch.c (hoist_guard): Release get_loop_body
338 vector. Fix trailing white spaces.
340 2017-02-14 James Greenhalgh <james.greenhalgh@arm.com>
342 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Handle
345 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
347 PR rtl-optimization/68664
348 * config/arm/arm.c (arm_sched_can_speculate_insn):
349 New function. Declare prototype.
350 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
352 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
354 PR rtl-optimization/68664
355 * config/aarch64/aarch64.c (aarch64_sched_can_speculate_insn):
357 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
359 2017-02-14 Amit Pawar <amit.pawar@amd.com>
361 * config/i386/i386.c (znver1_cost): Fix the alignment for function and
362 max skip bytes for function, loop and jump.
364 2017-02-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
366 * gimple-pretty-print.c (dump_unary_rhs): Adjust dump format for
367 ABS_EXPR for gimple dump.
369 2017-02-14 Jakub Jelinek <jakub@redhat.com>
372 * config/sh/sh.c (expand_cbranchdi4): Don't clear operands[4].
374 PR tree-optimization/79408
375 * tree-vrp.c (simplify_div_or_mod_using_ranges): Handle also the
376 case when on TRUNC_MOD_EXPR op0 is INTEGER_CST.
377 (simplify_stmt_using_ranges): Call simplify_div_or_mod_using_ranges
378 also if rhs1 is INTEGER_CST.
380 2017-02-14 Richard Biener <rguenther@suse.de>
383 * tree-into-ssa.c (insert_phi_nodes): When the function can
384 have abnormal edges rewrite SSA names with broken use-def
385 dominance out of SSA and register them for PHI insertion.
387 2017-02-13 Martin Sebor <msebor@redhat.com>
390 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call): Avoid
391 clearing info.nowrite flag when snprintf size argument is a range.
393 2017-02-13 Jakub Jelinek <jakub@redhat.com>
395 * cprop.c (cprop_jump): Add missing space in string literal.
396 * tree-ssa-structalias.c (rewrite_constraints): Likewise.
397 (get_constraint_for_component_ref): Likewise.
398 * df-core.c (df_worklist_dataflow_doublequeue): Likewise.
399 * tree-outof-ssa.c (insert_partition_copy_on_edge): Likewise.
400 * lra-constraints.c (process_alt_operands): Likewise.
401 * ipa-inline.c (inline_small_functions): Likewise.
402 * tree-ssa-sccvn.c (visit_reference_op_store): Likewise.
403 * cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
404 * trans-mem.c (diagnose_tm_1_op): Likewise.
405 * omp-grid.c (grid_find_single_omp_among_assignments): Likewise.
406 (grid_parallel_clauses_gridifiable): Likewise.
408 * config/nvptx/mkoffload.c (process): Add space in between
411 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
412 "MOD4_SSE_REGS" and "ALL_REGS".
414 * spellcheck.c (test_data): Add , in between "foo" and "food".
416 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
419 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
420 boundary crossing check and subsequent code generation agree.
422 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
424 * config/aarch64/aarch64.c (has_memory_op): Delete.
425 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
428 2017-02-13 Jakub Jelinek <jakub@redhat.com>
430 PR rtl-optimization/79388
431 PR rtl-optimization/79450
432 * combine.c (distribute_notes): When removing TEM_INSN for which
433 corresponding dest has last value recorded, invalidate that last
436 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
438 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
439 of explicit '@'. Add missing assembly comment marker on branch costs
442 2017-02-13 Nathan Sidwell <nathan@acm.org>
444 * gengtype-lex.l (<in_struct>): Add '/'.
446 2017-02-13 Martin Liska <mliska@suse.cz>
449 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
451 2017-02-13 Richard Biener <rguenther@suse.de>
453 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
455 * configure: Re-generate.
456 * config.in: Likewise.
457 * graphite-dependences.c: Simplify as if
458 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
459 * graphite-isl-ast-to-gimple.c: Likewise.
460 * graphite-optimize-isl.c: Likewise.
461 * graphite-poly.c: Likewise.
462 * graphite-sese-to-poly.c: Likewise.
463 * graphite.h: Likewise.
464 * toplev.c: Include isl/version.h and use isl_version () for
465 printing the ISL version.
466 * doc/install.texi: Update ISL requirement.
468 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
470 * doc/standards.texi (Standards): Update reference to
473 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
475 * doc/extend.texi (Named Address Spaces): sourceware.org now
477 * doc/install.texi (Binaries): Ditto.
480 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
482 * doc/cpp.texi: Replace "stringify"/"stringification" with C
483 standard terminology "stringize"/"stringizing" throughout.
484 * doc/cppinternals.texi: Likewise.
486 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
488 * doc/extend.texi: Fix some spelling mistakes and typos.
489 * doc/invoke.texi: Likewise.
491 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
494 * params.def (inline-min-speedup) Change from 10 to 8.
496 2017-02-11 Jakub Jelinek <jakub@redhat.com>
498 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
501 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
504 * ipa-inline-analysis.c (get_minimal_bb): New function.
505 (record_modified): Use it.
506 (remap_edge_change_prob): Handle also ancestor functions.
508 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
510 * doc/contrib.texi (Contributors): Remove broken link into
511 the Mauve CVS repository.
513 2017-02-11 Jakub Jelinek <jakub@redhat.com>
516 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
517 result computation whenever lhs doesn't have vector mode, not
518 just when it has BLKmode.
520 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
522 * doc/makefile.texi (profiledbootstrap): Refer to the
523 installation instructions only in textual form.
525 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
528 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
530 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
532 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
533 (Specific): Update mingw-w64 reference.
535 (Specific): Remove broken link to Renesas RX processor.
537 2017-02-10 Richard Biener <rguenther@suse.de>
539 * toplev.c (process_options): Do not mention obsolete graphite
540 options when printing sorry message about missing graphite support.
541 Mention -floop-nest-optimize.
543 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
545 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
546 (vtst_p16): Likewise.
547 (vtstq_p8): Likewise.
548 (vtstq_p16): Likewise.
550 (vtstq_p64): Likewise.
551 * config/arm/arm_neon.h (vgetq_lane_p64): New.
552 (vset_lane_p64): New.
553 (vsetq_lane_p64): New.
555 2017-02-10 Jakub Jelinek <jakub@redhat.com>
557 PR tree-optimization/79411
558 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
559 stmt operands are SSA_NAMEs used in abnormal phis.
560 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
563 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
566 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
569 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
571 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
573 2017-02-09 Jakub Jelinek <jakub@redhat.com>
575 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
579 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
580 not arbitrary TREE_CONSTANT.
583 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
584 "omp declare target link" attribute unless is_global_var.
585 * omp-offload.c (find_link_var_op): Likewise.
587 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
588 Chung-Lin Tang <cltang@codesourcery.com>
590 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
592 (gimplify_adjust_omp_clauses): Don't delete TILE.
593 (gimplify_omp_for): Deal with TILE.
594 * internal-fn.c (expand_GOACC_TILE): New function.
595 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
597 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
598 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
600 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
601 avoid DIV for outermost collapse var.
602 (expand_oacc_for): Insert tile element loop as needed. Adjust.
603 Remove out of date comments, fix whitespace.
604 * omp-general.c (omp_extract_for_data): Deal with tiling.
605 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
606 adjust OLF_DIM_BASE value.
607 (struct omp_for_data): Add tiling field.
608 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
609 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
610 for auto loops. Remove default auto determining, moved to
611 oacc_loop_fixed_partitions.
612 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
613 stmts, add e_mask field.
614 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
615 (oacc_thread_numbers): Use oacc_dim_call.
616 (oacc_xform_tile): New.
617 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
618 (finish_oacc_loop): Adjust for ifns vector.
619 (oacc_loop_discover_walk): Append loop abstraction sites to list,
620 add case for GOACC_TILE fns.
621 (oacc_loop_xform_loop): Delete.
622 (oacc_loop_process): Iterate over call list directly, and add
623 handling for GOACC_TILE fns.
624 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
626 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
627 vector partitioning to outer loops. Assign 2 partitions to loops
628 when available. Add TILE handling.
629 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
630 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
631 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
632 * tree.c (omp_clause_num_ops): Adjust TILE ops.
633 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
635 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
637 * configure.ac (ACX_BUGURL): Update.
638 * configure: Regenerate.
640 2017-02-09 Richard Biener <rguenther@suse.de>
642 PR tree-optimization/69823
643 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
644 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
646 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
648 * config/arc/arc-c.def: Add __NPS400__ definition.
649 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
650 (TARGET_NPS400): Define.
652 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
654 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
656 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
658 (arc_cpu_types): Fill the arch_info field with a pointer into the
659 arc_arch_types table.
660 (arc_selected_cpu): Declare.
661 * config/arc/arc.c (arc_selected_cpu): Make global.
662 (arc_selected_arch): Delete.
663 (arc_base_cpu): Delete.
664 (arc_override_options): Remove references to deleted variables,
665 update access to arch information.
666 (ARC_OPT): Update access to arch information.
667 (ARC_OPTX): Likewise.
668 * config/arc/arc.h (arc_base_cpu): Remove declaration.
669 (TARGET_ARC600): Update access to arch information.
670 (TARGET_ARC601): Likewise.
671 (TARGET_ARC700): Likewise.
672 (TARGET_EM): Likewise.
673 (TARGET_HS): Likewise.
674 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
677 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
680 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
681 condition/operands for integer GE/LE/GEU/LEU operations.
683 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
686 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
689 2017-02-08 Martin Jambor <mjambor@suse.cz>
692 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
693 whether allocation happened.
694 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
695 nothing was allocated.
697 2017-02-08 Jakub Jelinek <jakub@redhat.com>
699 PR tree-optimization/79408
700 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
701 constant, but SSA_NAME with a known integer range, use the minimum
702 of that range instead of op1 to determine if modulo can be replaced
703 with its first operand.
705 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
707 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
709 2017-02-08 Richard Biener <rguenther@suse.de>
711 PR tree-optimization/71824
712 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
713 Check all loops contained in the merged region.
715 2017-02-07 Andrew Pinski <apinski@cavium.com>
717 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
719 2017-02-07 Andrew Pinski <apinski@cavium.com>
721 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
722 (thunderxt88): Likewise.
723 (thunderxt81): Disable LSE and change v8.1 to v8.
724 (thunderxt83): Likewise.
726 2017-02-07 Jakub Jelinek <jakub@redhat.com>
727 Richard Biener <rguenther@suse.de>
730 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
731 type from int to size_t.
732 * ira-costs.c (struct_costs_size): Change type from int to size_t.
734 2017-02-07 Jakub Jelinek <jakub@redhat.com>
736 PR rtl-optimization/79386
737 * cprop.c (bypass_conditional_jumps): Initialize
738 bypass_last_basic_block already before splitting bbs after
739 unconditional traps...
740 (bypass_conditional_jumps): ... rather than here.
743 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
744 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
745 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
746 fix -masm=intel patterns.
748 2017-02-07 Richard Biener <rguenther@suse.de>
750 PR tree-optimization/79256
752 * builtins.c (get_object_alignment_2): Use min_align_of_type
753 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
754 and ADJUST_FIELD_ALIGN.
756 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
758 * doc/tm.texi: Regenerate.
759 * stor-layout.c (layout_decl): Adjust.
760 (update_alignment_for_field): Likewise.
761 (place_field): Likewise.
762 (min_align_of_type): Likewise.
763 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
764 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
765 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
766 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
767 * config/frv/frv.c (frv_adjust_field_align): Likewise.
768 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
769 * config/i386/i386.c (x86_field_alignment): Likewise.
770 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
771 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
772 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
773 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
774 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
775 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
779 2017-01-30 Richard Biener <rguenther@suse.de>
781 PR tree-optimization/79256
782 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
783 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
786 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
788 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
789 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
790 builtins to SImode and emit a zero-extend, if necessary.
792 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
794 * docs/invoke.texi (RISC-V Options): Alphabetize.
796 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
798 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
801 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
803 * config/riscv/riscv.c: New file.
804 * gcc/common/config/riscv/riscv-common.c: Likewise.
805 * config.gcc: Likewise.
806 * config/riscv/constraints.md: Likewise.
807 * config/riscv/elf.h: Likewise.
808 * config/riscv/generic.md: Likewise.
809 * config/riscv/linux.h: Likewise.
810 * config/riscv/multilib-generator: Likewise.
811 * config/riscv/peephole.md: Likewise.
812 * config/riscv/pic.md: Likewise.
813 * config/riscv/predicates.md: Likewise.
814 * config/riscv/riscv-builtins.c: Likewise.
815 * config/riscv/riscv-c.c: Likewise.
816 * config/riscv/riscv-ftypes.def: Likewise.
817 * config/riscv/riscv-modes.def: Likewise.
818 * config/riscv/riscv-opts.h: Likewise.
819 * config/riscv/riscv-protos.h: Likewise.
820 * config/riscv/riscv.h: Likewise.
821 * config/riscv/riscv.md: Likewise.
822 * config/riscv/riscv.opt: Likewise.
823 * config/riscv/sync.md: Likewise.
824 * config/riscv/t-elf-multilib: Likewise.
825 * config/riscv/t-linux: Likewise.
826 * config/riscv/t-linux-multilib: Likewise.
827 * config/riscv/t-riscv: Likewise.
828 * configure.ac: Likewise.
829 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
830 Waterman as RISC-V maintainers.
831 * doc/install.texi: Add RISC-V entries.
832 * doc/invoke.texi: Add RISC-V options section.
833 * doc/md.texi: Add RISC-V constraints section.
834 * configure: Regenerated.
836 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
839 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
840 false values to be constant vectors with all 0 or all 1 bits set.
841 (vcondu<mode><mode>): Likewise.
842 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
844 (fpmask_comparison_operator): Update comment.
845 (vecint_comparison_operator): New predicate.
846 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
847 vector conditionals when the true and false values are constant
848 vectors with all 0 bits or all 1 bits set.
850 2017-02-06 Martin Sebor <msebor@redhat.com>
852 PR tree-optimization/79376
853 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
855 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
857 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
858 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
859 to simplify split condition.
861 2017-02-06 Jakub Jelinek <jakub@redhat.com>
863 * omp-expand.c (oxpand_omp_atomic_fetch_op,
864 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
867 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
869 PR rtl-optimization/68664
870 * target.def (can_speculate_insn): New hook.
871 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
872 * doc/tm.texi: Regenerate.
873 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
874 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
875 (rs6000_sched_can_speculate_insn): New function.
877 2017-02-06 Jakub Jelinek <jakub@redhat.com>
879 PR tree-optimization/79284
880 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
881 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
882 vectorizable_mask_load_store, vectorizable_operation,
883 vect_is_simple_cond, get_same_sized_vectype): Use it instead
884 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
885 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
886 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
887 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
888 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
889 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
890 is_gimple_assign (stmt). Replace another such test with
891 is_gimple_assign (stmt).
893 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
896 * config/avr/avr.c (rtl-iter.h): Include it.
897 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
898 (avr_legitimate_combined_insn): ...and implementation.
900 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
902 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
903 * config/s390/s390.c (s390_const_operand_ok)
904 (s390_canonicalize_comparison, s390_extract_part)
905 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
906 (s390_contiguous_bitmask_p, s390_rtx_costs)
907 (legitimize_pic_address): Likewise.
908 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
909 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
910 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
911 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
912 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
914 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
916 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
917 REGNO($0) == REGNO($1).
919 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
921 * config/s390/linux.h(SIZE_TYPE): Add comment.
923 2017-02-06 Julian Brown <julian@codesourcery.com>
924 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
925 Virendra Pathak <virendra.pathak@broadcom.com>
927 * config/aarch64/aarch64-cores.def: Change the scheduler
929 * config/aarch64/aarch64.md: Include thunderx2t99.md.
930 * config/aarch64/thunderx2t99.md: New file.
932 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
934 * doc/standards.texi (Go Language): Update link to language
937 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
939 * tree-eh.c (lower_resx): Sanitize profile.
940 (cleanup_empty_eh_move_lp): Likewise.
942 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
945 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
947 * cfgloopmanip.h (loop_version): Update prototype.
948 * modulo-sched.c (sms_schedule): Update call of loop_version.
949 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
950 * tree-parloops.c (gen_parallel_loop): Likewise.
951 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
952 * tree-ssa-loop-split.c (split_loop): Likewise.
953 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
954 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
956 2017-02-05 Martin Liska <mliska@suse.cz>
959 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
961 (print_operand_address): Initialize a struct to zero.
963 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
965 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
966 garbage collector only in textual form.
968 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
970 * doc/extend.texi (x86 specific memory model extensions for
971 transactional memory): Simplify a phrase.
973 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
976 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
977 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
978 (atomic_storedi_1): Likewise.
980 2017-02-04 Jakub Jelinek <jakub@redhat.com>
982 PR tree-optimization/79338
983 * tree-parloops.c (gather_scalar_reductions): Don't call
984 vect_analyze_loop_form for loop->inner before destroying loop's
987 2017-02-03 Martin Sebor <msebor@redhat.com>
989 PR tree-optimization/79327
990 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
991 when precision has resulted in leading zeros.
992 (format_integer): Adjust the likely counter to assume an unknown
993 argument that may be zero is non-zero.
995 2017-02-03 Jason Merrill <jason@redhat.com>
998 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
999 avoid copying non-taken branch.
1001 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1003 PR tree-optimization/79340
1004 * tree-vect-loop.c (vectorizable_reduction): Release
1005 vec_defs elements after safe_splicing them into other vectors.
1008 PR tree-optimization/79327
1009 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
1010 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
1012 (format_integer): Use wide_int_to_tree instead of build_int_cst
1013 + to_?hwi. If argmin is NULL, just set argmin and argmax to
1014 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
1015 of shortest and longest sequence.
1017 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
1019 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
1020 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
1022 2017-02-03 Walter Lee <walt@tilera.com>
1025 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
1026 after initial stackframe link reg save.
1027 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
1029 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1032 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
1033 wu for stxssp alternative.
1035 2017-02-03 Martin Sebor <msebor@redhat.com>
1037 PR tree-optimization/79352
1038 * gimple-fold.c (get_range_strlen): Add argument.
1039 (get_range_strlen): Change return type to bool.
1040 (get_maxval_strlen): Pass in a dummy argument.
1041 * gimple-fold.h (get_range_strlen): Change return type to bool.
1042 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
1043 * tree.h (array_at_struct_end_p): Add argument.
1044 * tree.c (array_at_struct_end_p): Handle it.
1046 2017-02-03 Martin Liska <mliska@suse.cz>
1049 * multiple_target.c (create_dispatcher_calls): Redirect edge
1050 from a caller of a dispatcher.
1051 (expand_target_clones): Make the clones local.
1052 (ipa_target_clone): Do both target clones and resolvers.
1053 (ipa_dispatcher_calls): Remove the pass.
1054 (pass_dispatcher_calls::gate): Likewise.
1055 (make_pass_dispatcher_calls): Likewise.
1056 * passes.def (pass_target_clone): Put as very first IPA early
1059 2017-02-03 Martin Liska <mliska@suse.cz>
1061 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
1062 in case of a function with ifunc attribute.
1064 2017-02-03 Martin Liska <mliska@suse.cz>
1066 * cgraph.c (cgraph_node::dump): Dump function version info.
1067 * symtab.c (symtab_node::dump_base): Add missing new line.
1069 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
1071 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
1072 (ifcombine_ifandif): Use it.
1074 2017-02-03 Martin Liska <mliska@suse.cz>
1076 * doc/invoke.texi: Document default value for
1077 use-after-scope-direct-emission-threshold.
1079 2017-02-03 Martin Liska <mliska@suse.cz>
1081 PR tree-optimization/79339
1082 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
1083 (format_floating): Likewise.
1085 2017-02-03 Martin Liska <mliska@suse.cz>
1088 * ipa-prop.c (ipa_node_params_t::insert): Remove current
1090 (ipa_node_params_t::remove): Likewise.
1091 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
1092 initialization from removed ipa_node_params_t::insert.
1093 (ipa_node_params::~ipa_node_params): Move from removed
1094 ipa_node_params_t::release.
1095 * symbol-summary.h (symbol_summary::m_released): New member.
1096 Do not release a summary twice. Do not allow to call finalizer
1097 for types of a summary that live in GGC memory.
1099 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
1101 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
1104 2017-02-02 Martin Sebor <msebor@redhat.com>
1107 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
1108 (format_string): Tighten up the range of output for non-constant
1109 strings and correct the expected range for wide non-constant strings.
1111 2017-02-02 Martin Sebor <msebor@redhat.com>
1113 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
1116 * doc/invoke.texi (-fdump-tree-): Remove pass-specific options from
1118 (-fdump-tree-@var): Add to index and document how to come up
1119 with pass-specific option and dump file names.
1120 (-fdump-passes): Clarify where to look for output.
1122 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
1125 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
1126 statistics of the analyzed path; allow threading for speed when
1127 any of BBs along the path are optimized for speed.
1129 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
1132 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
1133 settings of the virtual registers.
1136 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
1138 * explow.c (get_dynamic_stack_size): Take known alignment of stack
1139 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
1142 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1144 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
1145 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
1147 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1149 * config/s390/s390.md: Add missing comments with the expanded
1151 * config/s390/vector.md: Likewise.
1152 * config/s390/vx-builtins.md: Likewise.
1154 2017-02-02 Jakub Jelinek <jakub@redhat.com>
1157 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
1158 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
1159 conditions on a single line.
1161 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1163 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
1164 __S390_VX__ to __VX__.
1166 2017-02-01 Andrew Pinski <apinski@cavium.com>
1168 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
1169 stmt_info to record_stmt_cost.
1170 (vect_get_known_peeling_cost): Pass stmt_info if known to
1172 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
1173 cpu_vector_cost field into
1174 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
1175 field into vec_int_stmt_cost and vec_fp_stmt_cost.
1176 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
1177 splitting of scalar_stmt_cost and vec_stmt_cost.
1178 (thunderx_vector_cost): Likewise.
1179 (cortexa57_vector_cost): LIkewise.
1180 (exynosm1_vector_cost): Likewise.
1181 (xgene1_vector_cost): Likewise.
1182 (thunderx2t99_vector_cost): Improve after the splitting of the two
1184 (aarch64_builtin_vectorization_cost): Update for the splitting of
1185 scalar_stmt_cost and vec_stmt_cost.
1187 2017-02-01 Torvald Riegel <triegel@redhat.com>
1188 Richard Henderson <rth@redhat.com>
1190 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
1191 conditional on existance of a fast atomic load.
1192 * optabs-query.c (can_atomic_load_p): New function.
1193 * optabs-query.h (can_atomic_load_p): Declare it.
1194 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
1195 no fast atomic load is available for the particular size of access.
1196 (expand_atomic_compare_and_swap): Likewise.
1197 (expand_atomic_load): Likewise.
1198 (expand_atomic_store): Likewise.
1199 (expand_atomic_fetch_op): Likewise.
1200 * testsuite/lib/target-supports.exp
1201 (check_effective_target_sync_int_128): Remove x86 because it provides
1202 no fast atomic load.
1203 (check_effective_target_sync_int_128_runtime): Likewise.
1205 2017-02-01 Richard Biener <rguenther@suse.de>
1207 * graphite.c: Include tree-vectorizer.h for find_loop_location.
1208 (graphite_transform_loops): Provide opt-info for optimized nests.
1209 * tree-parloop.c (parallelize_loops): Provide opt-info for
1212 2017-02-01 Richard Biener <rguenther@suse.de>
1215 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
1218 2017-02-01 Richard Biener <rguenther@suse.de>
1220 PR tree-optimization/71824
1221 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
1222 Verify the loops are valid in the merged SESE region.
1223 (scop_detection::can_represent_loop_1): Check analyzing the
1224 evolution of the number of iterations in the region succeeds.
1226 2017-01-31 Ian Lance Taylor <iant@golang.org>
1228 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
1229 REG_ARGS_SIZE note to 32-bit push insns and call insn.
1231 2017-01-31 David Malcolm <dmalcolm@redhat.com>
1233 PR preprocessor/79210
1234 * input.c (get_substring_ranges_for_loc): Replace line_width
1235 assertion with error-handling.
1237 2017-01-31 Richard Biener <rguenther@suse.de>
1239 PR tree-optimization/77318
1240 * graphite-sese-to-poly.c (extract_affine): Fix assert.
1241 (create_pw_aff_from_tree): Take loop parameter.
1242 (add_condition_to_pbb): Pass loop of the condition to
1243 create_pw_aff_from_tree.
1245 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1247 * config/s390/s390.c (s390_asan_shadow_offset): New function.
1248 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
1250 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1254 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
1256 (convert_int_to_float128): Likewise.
1257 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
1258 (convert_int_to_float128): Likewise.
1259 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
1260 (UNSPEC_IEEE128_CONVERT): Likewise.
1261 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
1262 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
1263 Use local variables for IBM extended format.
1264 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
1265 (fix_trunc<mode>si2_fprs): Likewise.
1266 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
1267 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
1268 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
1269 to know that we can now have integers of all sizes in vector
1271 (fix<uns>_<mode>di2_hw): Likewise.
1272 (float<uns>_<mode>si2_hw): Likewise.
1273 (fix_<mode>si2_hw): Likewise.
1274 (fixuns_<mode>si2_hw): Likewise.
1275 (float<uns>_<mode>di2_hw): Likewise.
1276 (float_<mode>di2_hw): Likewise.
1277 (float_<mode>si2_hw): Likewise.
1278 (floatuns_<mode>di2_hw): Likewise.
1279 (floatuns_<mode>si2_hw): Likewise.
1280 (xscvqp<su>wz_<mode>): Delete, no longer used.
1281 (xscvqp<su>dz_<mode>): Likewise.
1282 (xscv<su>dqp_<mode>): Likewise.
1283 (ieee128_mfvsrd_64bit): Likewise.
1284 (ieee128_mfvsrd_32bit): Likewise.
1285 (ieee128_mfvsrwz): Likewise.
1286 (ieee128_mtvsrw): Likewise.
1287 (ieee128_mtvsrd_64bit): Likewise.
1288 (ieee128_mtvsrd_32bit): Likewise.
1290 2017-01-31 Martin Liska <mliska@suse.cz>
1293 * ipa-prop.c (ipa_free_all_node_params): Call release method
1294 instead of ~sumbol_summary to not to trigger double times
1297 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
1299 PR tree-optimization/71691
1300 * bitmap.h (class auto_bitmap): New.
1301 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
1302 is_maybe_undefined instead of ssa_undefined_value_p.
1304 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1306 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
1307 __S390_ARCH_LEVEL__ to __ARCH__.
1309 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1311 PR tree-optimization/79267
1312 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
1313 if should_remove_lhs_p is true.
1315 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
1318 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
1319 (add_alignment_attribute): New.
1320 (base_type_die): Add alignment attribute.
1321 (subrange_type_die): Likewise.
1322 (modified_type_die): Likewise.
1323 (gen_array_type_die): Likewise.
1324 (gen_descr_array_type_die: Likewise.
1325 (gen_enumeration_type_die): Likewise.
1326 (gen_subprogram_die): Likewise.
1327 (gen_variable_die): Likewise.
1328 (gen_field_die): Likewise.
1329 (gen_ptr_to_mbr_type_die): Likewise.
1330 (gen_struct_or_union_type_die): Likewise.
1331 (gen_subroutine_type_die): Likewise.
1332 (gen_typedef_die): Likewise.
1333 (base_type_cmp): Compare alignment attribute.
1335 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1338 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
1339 (setb_unsigned) New pattern for setb with CCUNS.
1340 * config/rs6000/rs6000.c (expand_block_compare): Use a different
1341 subfc./subfe sequence to avoid overflow problems. Generate a
1342 shorter sequence with cmpld/setb for power9.
1343 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
1344 for generating subfc. instruction.
1345 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
1346 now uses this instruction.
1348 2017-01-30 Ian Lance Taylor <iant@google.com>
1351 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
1352 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
1354 2017-01-30 Martin Sebor <msebor@redhat.com>
1356 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
1357 Move constant to the right of a relational operator.
1358 (get_mpfr_format_length, format_character, format_string): Ditto.
1359 (should_warn_p, maybe_warn): Same.
1361 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
1363 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
1366 * asan.c (get_translation_unit_decl): Remove function.
1367 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
1369 2017-01-30 Martin Liska <mliska@suse.cz>
1371 PR gcov-profile/79259
1372 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
1375 2017-01-30 Martin Liska <mliska@suse.cz>
1378 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
1379 Initialize variables with NULL value.
1381 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
1384 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
1386 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
1388 2017-01-30 Richard Biener <rguenther@suse.de>
1390 PR tree-optimization/79276
1391 * tree-vrp.c (process_assert_insertions): Properly adjust common
1392 when removing a duplicate.
1394 * gcc.dg/torture/pr79276.c: New testcase.
1396 2017-01-30 Richard Biener <rguenther@suse.de>
1398 PR tree-optimization/79256
1399 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
1400 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
1402 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
1404 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
1407 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
1408 ("*r<noxa>sbg_<mode>_sll_bitmask")
1409 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
1410 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
1411 Use contiguous_bitmask_nowrap_operand.
1413 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1416 * config/rs6000/altivec.h (vec_xl): Revise #define.
1417 (vec_xst): Likewise.
1419 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
1421 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
1423 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
1425 PR rtl-optimization/79194
1426 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
1427 traps before call to bypass_conditional_jumps.
1429 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1431 PR tree-optimization/71374
1432 * lra-constraints.c (check_conflict_input_operands): New.
1433 (match_reload): Use it.
1435 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1438 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
1439 account to calculate conflict_set.
1441 2017-01-27 Bin Cheng <bin.cheng@arm.com>
1443 PR rtl-optimization/78559
1444 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
1445 other_insn in combine.
1447 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
1449 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
1450 uint16_type_node for BT_UINT16.
1452 2017-01-27 David Malcolm <dmalcolm@redhat.com>
1454 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
1455 "RTL Tests" to menu.
1456 (GIMPLE Tests): New node.
1457 (RTL Tests): New node.
1459 2017-01-27 Richard Biener <rguenther@suse.de>
1461 PR tree-optimization/79245
1462 * tree-loop-distribution.c (distribute_loop): Apply cost
1463 modeling also to detected patterns.
1465 2017-01-27 Richard Biener <rguenther@suse.de>
1467 PR tree-optimization/71433
1468 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
1469 (compare_assert_loc): New function.
1470 (process_assert_insertions): Sort and optimize assert locations
1471 to remove duplicates and push down identical assertions on
1472 edges to their destination block.
1474 2017-01-27 Richard Biener <rguenther@suse.de>
1476 PR tree-optimization/79244
1477 * tree-vrp.c (remove_range_assertions): Forcefully propagate
1478 out SSA names even if abnormal.
1480 2017-01-27 Jakub Jelinek <jakub@redhat.com>
1482 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
1483 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
1484 instead of MPFR_RNDN.
1486 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
1489 * arm.c (arm_option_override): Don't call build_target_option_node
1490 until after doing all option overrides.
1491 (arm_valid_target_attribute_tree): Likewise.
1493 2017-01-27 Martin Liska <mliska@suse.cz>
1495 * doc/invoke.texi (-fprofile-arcs): Document profiling support
1496 for {cd}tors and C++ {cd}tors.
1498 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
1500 * config/s390/s390.md ("*setmem_long_and")
1501 ("*setmem_long_and_31z"): Use zero_extend instead of and.
1503 2017-01-26 Martin Sebor <msebor@redhat.com>
1505 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
1508 2017-01-26 Martin Sebor <msebor@redhat.com>
1510 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
1511 HAVE_DFmode before using XFmode or DFmode.
1512 (parse_directive): Avoid using the z length modifier to avoid
1513 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
1516 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
1517 to accept adjustment as an array.
1518 (get_int_range): New function.
1519 (struct directive): Make width and prec arrays.
1520 (directive::set_width, directive::set_precision): Call get_int_range.
1521 (format_integer, format_floating): Handle width and precision ranges.
1522 (format_string, parse_directive): Same.
1524 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1527 * dwarf2out.c (generate_skeleton_bottom_up): For children with
1528 comdat_type_p set, just clone them, but keep the children in the
1532 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
1533 which have direct callers with -fvar-tracking-assignments enabled
1535 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
1536 inside of type units.
1538 2017-01-26 Martin Sebor <msebor@redhat.com>
1541 * gimple-ssa-sprintf.c (struct result_range): Add likely and
1543 (struct format_result): Replace number_chars, number_chars_min,
1544 and number_chars_max with a single member of struct result_range.
1546 (format_result::operator+=): Adjust.
1547 (struct fmtresult): Remove bounded. Handle likely and unlikely
1549 (fmtresult::adjust_for_width_or_precision): New function.
1550 (fmtresult:type_max_digits): New function.
1551 (bytes_remaining): Handle likely and unlikely counters.
1552 (min_bytes_remaining): Remove.
1553 (format_percent): Simplify.
1554 (format_integer, format_floating): Set likely and unlikely counters.
1555 (get_string_length, format_character, format_string): Same.
1556 (format_plain, should_warn_p): New function.
1557 (maybe_warn): Call should_warn_p. Update diagnostic messages
1558 and handle those for all directives, including plain strings.
1559 (format_directive): Handle likely and unlikely counters.
1560 Remove unnecessary quoting from diagnostics. Add an informational
1562 (add_bytes): Remove.
1563 (pass_sprintf_length::compute_format_length): Simplify.
1564 (try_substitute_return_value): Handle likely and unlikely counters.
1566 2017-01-26 Carl Love <cel@us.ibm.com>
1568 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
1569 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
1571 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
1574 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
1575 endianess for subregs into account.
1576 * lra-constraints.c (lra_constraints): Do risky transformations
1577 always on the first iteration.
1578 * lra-lives.c (check_pseudos_live_through_calls): Add arg
1579 last_call_used_reg_set.
1580 (process_bb_lives): Define and use last_call_used_reg_set.
1581 * lra.c (lra): Always continue after lra_constraints on the first
1584 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
1586 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
1588 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
1590 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1592 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
1593 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
1594 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
1595 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
1596 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
1597 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
1598 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
1599 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
1600 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
1602 2017-01-26 Marek Polacek <polacek@redhat.com>
1605 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
1606 for the third operand.
1608 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1611 * omp-low.c (struct omp_context): Add simt_stmt field.
1612 (scan_omp_for): Return omp_context *.
1613 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
1614 context to the _simt_ SIMD stmt.
1615 (lower_omp_for): For combined SIMD with sibling _simt_
1616 SIMD, make sure to use the same decls in _looptemp_
1617 clauses as in the sibling.
1619 2017-01-26 David Sherwood <david.sherwood@arm.com>
1622 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
1625 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1628 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
1629 emit fld b; fld a; if possible.
1631 * brig-builtins.def: Update copyright years.
1632 * config/arm/arm_acle_builtins.def: Update copyright years.
1634 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
1637 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
1638 constraint instead of o for the stxsd instruction.
1640 2017-01-25 Carl Love <cel@us.ibm.com>
1642 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
1643 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
1645 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
1647 * doc/invoke.texi (C++ Dialect Options): Fix typo.
1649 2017-01-25 Richard Biener <rguenther@suse.de>
1651 PR tree-optimization/69264
1652 * target.def (vector_alignment_reachable): Improve documentation.
1653 * doc/tm.texi: Regenerate.
1654 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
1656 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
1657 earlier changes with respect to TYPE_USER_ALIGN.
1658 (vector_alignment_reachable_p): Likewise. Improve dumping.
1660 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1663 * config/arm/arm.md (xordi3): Force constant operand into a register
1666 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1668 * doc/invoke.texi (-fstore-merging): Correct default optimization
1669 levels at which it is enabled.
1670 (-O): Move -fstore-merging from list to...
1673 2017-01-25 Richard Biener <rguenther@suse.de>
1676 * omp-expand.c: Include debug.h.
1677 (expand_omp_taskreg): Make sure to generate early debug before
1678 outlining anything from a function.
1679 (expand_omp_target): Likewise.
1680 (grid_expand_target_grid_body): Likewise.
1682 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
1685 * asan.c (get_translation_unit_decl): New function.
1686 (asan_add_global): Extract modules file name from globals
1687 TRANSLATION_UNIT_DECL name.
1689 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
1692 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
1693 for long calls with APCS frame and VFP.
1695 2017-01-24 David Malcolm <dmalcolm@redhat.com>
1697 * cfg.c (original_copy_tables_initialized_p): New function.
1698 * cfg.h (original_copy_tables_initialized_p): New decl.
1699 * cfgrtl.c (relink_block_chain): Guard the call to
1700 free_original_copy_tables with a call to
1701 original_copy_tables_initialized_p.
1702 * cgraph.h (symtab_node::native_rtl_p): New decl.
1703 * cgraphunit.c (symtab_node::native_rtl_p): New function.
1704 (symtab_node::needed_p): Don't assert for early assembly output
1705 for __RTL functions.
1706 (cgraph_node::finalize_function): Set "force_output" for __RTL
1708 (cgraph_node::analyze): Bail out early for __RTL functions.
1709 (analyze_functions): Update assertion to support __RTL functions.
1710 (cgraph_node::expand): Bail out early for __RTL functions.
1711 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
1713 * function.h (struct function): Update comment for field
1715 * gimple-expr.c: Include "tree-pass.h".
1716 (gimple_has_body_p): Return false for __RTL functions.
1717 * Makefile.in (OBJS): Add run-rtl-passes.o.
1718 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
1720 (gcc::pass_manager::get_clean_slate): New accessor.
1721 * passes.c: Include "insn-addr.h".
1722 (should_skip_pass_p): Add logging. Update logic for running
1723 "expand" to be compatible with both __GIMPLE and __RTL. Guard
1724 property-provider override so it is only done for gimple passes.
1726 (skip_pass): New function.
1727 (execute_one_pass): Call skip_pass when skipping passes.
1728 * read-md.c (md_reader::read_char): Support filtering
1729 the input to a subset of line numbers.
1730 (md_reader::md_reader): Initialize fields
1731 m_first_line and m_last_line.
1732 (md_reader::read_file_fragment): New function.
1733 * read-md.h (md_reader::read_file_fragment): New decl.
1734 (md_reader::m_first_line): New field.
1735 (md_reader::m_last_line): New field.
1736 * read-rtl-function.c (function_reader::create_function): Only
1737 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
1738 curr_properties. Set DECL_INITIAL to a dummy block.
1739 (read_rtl_function_body_from_file_range): New function.
1740 * read-rtl-function.h (read_rtl_function_body_from_file_range):
1742 * run-rtl-passes.c: New file.
1743 * run-rtl-passes.h: New file.
1745 2017-01-24 Jeff Law <law@redhat.com>
1747 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
1750 2017-01-24 Bin Cheng <bin.cheng@arm.com>
1752 PR tree-optimization/79159
1753 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
1754 (record_nonwrapping_iv): Improve boundary using above function if no
1755 value range information.
1757 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
1758 Martin Jambor <mjambor@suse.cz>
1760 * brig-builtins.def: New file.
1761 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
1762 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
1763 (DEF_HSAIL_SAT_BUILTIN): Likewise.
1764 (DEF_HSAIL_INTR_BUILTIN): Likewise.
1765 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
1766 * builtin-types.def (BT_INT8): New.
1767 (BT_INT16): Likewise.
1768 (BT_UINT8): Likewise.
1769 (BT_UINT16): Likewise.
1770 (BT_FN_ULONG): Likewise.
1771 (BT_FN_UINT_INT): Likewise.
1772 (BT_FN_UINT_ULONG): Likewise.
1773 (BT_FN_UINT_LONG): Likewise.
1774 (BT_FN_UINT_PTR): Likewise.
1775 (BT_FN_ULONG_PTR): Likewise.
1776 (BT_FN_INT8_FLOAT): Likewise.
1777 (BT_FN_INT16_FLOAT): Likewise.
1778 (BT_FN_UINT32_FLOAT): Likewise.
1779 (BT_FN_UINT16_FLOAT): Likewise.
1780 (BT_FN_UINT8_FLOAT): Likewise.
1781 (BT_FN_UINT64_FLOAT): Likewise.
1782 (BT_FN_UINT16_UINT32): Likewise.
1783 (BT_FN_UINT32_UINT16): Likewise.
1784 (BT_FN_UINT16_UINT16_UINT16): Likewise.
1785 (BT_FN_INT_PTR_INT): Likewise.
1786 (BT_FN_UINT_PTR_UINT): Likewise.
1787 (BT_FN_LONG_PTR_LONG): Likewise.
1788 (BT_FN_ULONG_PTR_ULONG): Likewise.
1789 (BT_FN_VOID_UINT64_UINT64): Likewise.
1790 (BT_FN_UINT8_UINT8_UINT8): Likewise.
1791 (BT_FN_INT8_INT8_INT8): Likewise.
1792 (BT_FN_INT16_INT16_INT16): Likewise.
1793 (BT_FN_INT_INT_INT): Likewise.
1794 (BT_FN_UINT_FLOAT_UINT): Likewise.
1795 (BT_FN_FLOAT_UINT_UINT): Likewise.
1796 (BT_FN_ULONG_UINT_UINT): Likewise.
1797 (BT_FN_ULONG_UINT_PTR): Likewise.
1798 (BT_FN_ULONG_ULONG_ULONG): Likewise.
1799 (BT_FN_UINT_UINT_UINT): Likewise.
1800 (BT_FN_VOID_UINT_PTR): Likewise.
1801 (BT_FN_UINT_UINT_PTR: Likewise.
1802 (BT_FN_UINT32_UINT64_PTR): Likewise.
1803 (BT_FN_INT_INT_UINT_UINT): Likewise.
1804 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
1805 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
1806 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
1807 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
1808 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
1809 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
1810 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
1811 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
1812 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
1813 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
1814 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
1815 * doc/frontends.texi: List BRIG FE.
1816 * doc/install.texi (Testing): Add BRIG tesring requirements.
1817 * doc/invoke.texi (Overall Options): Mention BRIG.
1818 * doc/standards.texi (Standards): Doucment BRIG HSA version.
1820 2017-01-24 Richard Biener <rguenther@suse.de>
1822 PR translation/79208
1823 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
1825 2017-01-24 Martin Jambor <mjambor@suse.cz>
1828 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
1829 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
1832 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
1835 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
1836 casts from signed to unsigned really don't have a range.
1838 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1840 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
1841 GMP_RNDx for compatiblity.
1843 2017-01-24 Martin Liska <mliska@suse.cz>
1846 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
1847 that would prevent us to call alloca with -1 as argument.
1849 2017-01-24 Jakub Jelinek <jakub@redhat.com>
1851 * dwarf2out.c (output_compilation_unit_header, output_file_names):
1852 Avoid -Wformat-security warning.
1854 2017-01-23 Andrew Pinski <apinski@cavium.com>
1856 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
1859 2017-01-23 Martin Sebor <msebor@redhat.com>
1862 * gimple-ssa-sprintf.c (warn_level): New global.
1863 (format_integer): Use it here and throughout the rest of the file.
1864 Use the same switch to compute sign as base.
1865 (maybe_warn): New function.
1866 (format_directive): Factor out warnings into maybe_warn.
1867 Add debugging output. Use warn_level.
1868 (add_bytes): Use warn_level.
1869 (pass_sprintf_length::compute_format_length): Add debugging output.
1870 (try_substitute_return_value): Same.
1871 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
1874 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
1875 (struct fmtresult, format_integer, format_floating): Adjust.
1876 (fmtresult::fmtresult): Set max correctly in two argument ctor.
1877 (get_string_length, format_string,format_directive): Same.
1878 (pass_sprintf_length::compute_format_length): Same.
1879 (try_substitute_return_value): Simplify slightly.
1882 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
1883 (fmtresult::operator+=): Outlined.
1884 (struct fmtresult): Add ctors.
1885 (struct conversion_spec): Rename...
1886 (struct directive): ...to this. Add and remove data members.
1887 (directive::set_width, directive::set_precision): New functions.
1888 (format_percent): Use fmtresult ctor.
1889 (get_width_and_precision): Remove.
1890 (format_integer): Make naming changes. Avoid computing width and
1892 (format_floating): Same. Adjust indentation.
1893 (format_character, format_none): New functions.
1894 (format_string): Moved character handling to format_character.
1895 (format_directive): Remove arguments, change return type.
1896 (parse_directive): New function.
1897 (pass_sprintf_length::compute_format_length): Move directive
1898 parsing to parse_directive.
1900 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1902 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
1903 (assign_assembler_name_if_needed): ... this.
1904 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
1905 (assign_assembler_name_if_needed): ... this.
1906 (free_lang_data_in_cgraph): Adjust callers.
1907 * cgraphunit.c (cgraph_node::analyze): Likewise.
1908 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
1911 2017-01-23 Richard Biener <rguenther@suse.de>
1913 PR tree-optimization/79088
1914 PR tree-optimization/79188
1915 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
1916 resetting loop bounds after last path deletion. Reset loop
1917 bounds of the target loop, make code match the comments.
1918 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
1919 Make sure loops need no fixups.
1921 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1923 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1924 exponent support with double type for first argument.
1925 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1926 type returned by __builtin_vec_extract_sig,
1927 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1928 functions from "vector int" to "vector unsigned int" or from
1929 "vector long long int" to "vector unsigned long long int".
1930 Changed type returned by __builtin_vec_extract_exp,
1931 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1932 functions from "vector int" to "vector unsigned int" or from
1933 "vector long long int" to "vector unsigned long long int".
1934 Changed return type of __builtin_vec_test_data_class,
1935 __builtin_vec_test_data_class_sp, and
1936 __builtin_vec_test_data_class_dp from "vector int" to
1937 "vector bool int" or from "vector long long int" to "vector bool
1938 long long int" and changed second argument type from "unsigned
1939 int" to "int". Added new overloaded function forms "vector float
1940 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1941 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1942 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1943 double, vector unsigned long long int)" and "vector double
1944 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1945 long int)". Changed return type of
1946 __builtin_scalar_test_data_class and
1947 __builtin_scalar_test_data_class_sp and
1948 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1949 int" and changed second argument from "unsigned int" to "int".
1950 Changed type returned by __builtin_scalar_test_neg,
1951 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1952 from "int" to "bool int". Added new overloaded function form
1953 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1954 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1955 exponent double-precision with floating point first argument.
1956 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1957 documentation of scalar_test_data_class, scalar_test_neg,
1958 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1959 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1960 vec_test_data_class built-in functions to reflect refinements in
1961 their type signatures.
1963 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1965 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1967 (aarch64_elf_asm_destructor): Likewise.
1969 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1971 PR rtl-optimization/78634
1972 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1973 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1974 * ifcvt.c (noce_try_cmove): Add missing cost check.
1976 PR rtl-optimization/71724
1977 * combine.c (if_then_else_cond): Look for situations where it is
1978 beneficial to undo the work of one of the recursive calls.
1980 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1982 PR tree-optimization/70754
1983 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1984 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1985 combined stmt before it if not NULL.
1986 (combine_chains): Process refs reversely and compute dominance point
1989 2017-01-23 Martin Liska <mliska@suse.cz>
1991 PR tree-optimization/79196
1992 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1993 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1995 (strlen_optimize_stmt): Call the renamed function.
1997 2017-01-23 Michael Matz <matz@suse.de>
1999 PR tree-optimization/78384
2000 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
2002 2017-01-23 Richard Biener <rguenther@suse.de>
2004 PR tree-optimization/79186
2005 * tree-vrp.c (register_new_assert_for): Make sure we've seen
2006 both incoming edges before moving an assert.
2008 2017-01-23 Martin Jambor <mjambor@suse.cz>
2010 * ipa-prop.c (load_from_param_1): Removed.
2011 (load_from_unmodified_param): Bits from load_from_param_1 put back
2013 (load_from_param): Removed.
2014 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
2015 with stmt. Reverted back to use of load_from_unmodified_param.
2017 2017-01-23 Martin Jambor <mjambor@suse.cz>
2020 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
2021 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
2022 field a pointer to garbage collected vector, mark lattices and
2023 ipcp_orig_node with GTY((skip)).
2024 (ipa_get_param_count): Adjust to descriptors being a pointer.
2025 (ipa_get_param): Likewise.
2026 (ipa_get_type): Likewise.
2027 (ipa_get_param_move_cost): Likewise.
2028 (ipa_set_param_used): Likewise.
2029 (ipa_get_controlled_uses): Likewise.
2030 (ipa_set_controlled_uses): Likewise.
2031 (ipa_is_param_used): Likewise.
2032 (ipa_node_params_t): Move into garbage collector. New methods insert
2034 (ipa_node_params_sum): Annotate wth GTY(()).
2035 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
2037 (ipa_load_from_parm_agg): Adjust declaration.
2038 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
2039 * ipa-profile.c (ipa_profile): Likewise.
2040 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
2041 (ipa_populate_param_decls): Make descriptors parameter garbage
2043 (ipa_dump_param): Adjust to descriptors being a pointer.
2044 (ipa_alloc_node_params): Likewise.
2045 (ipa_initialize_node_params): Likewise.
2046 (load_from_param_1): Make descriptors parameter garbage collected.
2047 (load_from_unmodified_param): Likewise.
2048 (load_from_param): Likewise.
2049 (ipa_load_from_parm_agg): Likewise.
2050 (ipa_node_params::~ipa_node_params): Removed.
2051 (ipa_free_all_node_params): Remove call to delete operator.
2052 (ipa_node_params_t::insert): New.
2053 (ipa_node_params_t::remove): Likewise.
2054 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
2055 copy known_csts and known_contexts vectors.
2056 (ipa_read_node_info): Adjust to descriptors being a pointer.
2057 (ipcp_modif_dom_walker): Make m_descriptors field garbage
2059 (ipcp_transform_function): Make descriptors variable garbage
2062 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
2064 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
2065 * config/i386/avx512dqintrin.h: Ditto.
2066 * config/i386/avx512fintrin.h: Ditto.
2067 * gcc/config/i386/i386.c: Handle new builtins.
2068 * config/i386/i386-builtin.def: Add new builtins.
2069 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
2070 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
2072 2017-01-23 Jakub Jelinek <jakub@redhat.com>
2073 Martin Liska <mliska@suse.cz>
2075 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
2076 * asan.c (asan_expand_poison_ifn): Support stores and use
2077 appropriate ASAN report function.
2078 * internal-fn.c (expand_ASAN_POISON_USE): New function.
2079 * internal-fn.def (ASAN_POISON_USE): Declare.
2080 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
2081 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
2082 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
2083 ASAN_POISON calls w/o LHS.
2084 * tree-ssa.c (execute_update_addresses_taken): Create clobber
2085 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
2086 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
2087 * gimplify.c (asan_poison_variables): Add attribute
2088 use_after_scope_memory to variables that really needs to live
2090 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
2091 having the attribute.
2093 2017-01-23 Martin Liska <mliska@suse.cz>
2095 * asan.c (create_asan_shadow_var): New function.
2096 (asan_expand_poison_ifn): Likewise.
2097 * asan.h (asan_expand_poison_ifn): New declaration.
2098 * internal-fn.c (expand_ASAN_POISON): Likewise.
2099 * internal-fn.def (ASAN_POISON): New builtin.
2100 * sanopt.c (pass_sanopt::execute): Expand
2101 asan_expand_poison_ifn.
2102 * tree-inline.c (copy_decl_for_dup_finish): Make function
2104 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
2105 * tree-ssa.c (is_asan_mark_p): New function.
2106 (execute_update_addresses_taken): Rewrite local variables
2107 (identified just by use-after-scope as addressable) into SSA.
2109 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
2111 * doc/install.texi (Specific): opensource.apple.com uses https
2112 now. Remove trailing slash.
2114 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
2116 * README.Portability: Remove note on an Irix compatibility issue.
2118 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
2120 * gcov.c (INCLUDE_ALGORITHM): Define.
2121 (INCLUDE_VECTOR): Define.
2122 No longer include <vector> and <algorithm> directly.
2124 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
2126 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
2128 * doc/invoke.texi (Code Gen Options): Ditto.
2130 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
2133 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
2135 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
2137 rtl-optimization/79125
2138 * cprop.c (local_cprop_pass): Handle cases where we make an
2141 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
2145 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
2146 read from, for big endian.
2148 2017-01-20 Jiong Wang <jiong.wang@arm.com>
2150 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
2151 register pauth builtins for LP64 only.
2153 2017-01-20 Marek Polacek <polacek@redhat.com>
2156 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
2159 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
2161 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
2163 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
2164 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
2165 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
2167 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2170 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
2171 in big-endian mode when they are not a single duplicated value.
2173 2017-01-20 Richard Biener <rguenther@suse.de>
2175 * BASE-VER: Bump to 7.0.1.
2177 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
2179 * omp-low.c (omplow_simd_context): New struct. Use it...
2180 (lower_rec_simd_input_clauses): ...here and...
2181 (lower_rec_input_clauses): ...here to hold common data. Adjust all
2182 references to idx, lane, max_vf, is_simt.
2184 2017-01-20 Graham Markall <graham.markall@embecosm.com>
2186 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
2189 2017-01-20 Martin Jambor <mjambor@suse.cz>
2191 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
2192 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
2194 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
2195 (GTFILES): Rename hsa.c to hsa-common.c.
2196 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
2197 * hsa-dump.c: Likewise.
2198 * hsa-gen.c: Likewise.
2199 * hsa-regalloc.c: Likewise.
2200 * ipa-hsa.c: Likewise.
2201 * omp-expand.c: Likewise.
2202 * omp-low.c: Likewise.
2203 * toplev.c: Likewise.
2205 2017-01-20 Marek Polacek <polacek@redhat.com>
2208 * doc/invoke.texi: Document -Wduplicated-branches.
2209 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
2210 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
2211 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
2212 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
2213 return 0 only when not OEP_LEXICOGRAPHIC.
2214 (fold_build_cleanup_point_expr): Use the expression
2215 location when building CLEANUP_POINT_EXPR.
2216 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
2217 * tree.c (add_expr): Handle error_mark_node.
2219 2017-01-20 Martin Liska <mliska@suse.cz>
2222 * tree-profile.c (init_ic_make_global_vars): Do not call
2224 (gimple_init_gcov_profiler): Likewise.
2226 2017-01-20 Martin Liska <mliska@suse.cz>
2229 * cgraph.h (maybe_create_reference): Remove argument and
2231 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
2233 * ipa-cp.c (create_specialized_node): Likewise.
2234 * symtab.c (symtab_node::maybe_create_reference): Handle
2235 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
2237 2017-01-20 Martin Liska <mliska@suse.cz>
2239 * read-rtl-function.c (function_reader::create_function): Use
2240 build_decl instread of build_decl_stat.
2242 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
2244 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
2245 * config/i386/avx512dqintrin.h: Ditto.
2246 * config/i386/avx512fintrin.h: Ditto.
2247 * config/i386/i386-builtin-types.def: Add new types.
2248 * gcc/config/i386/i386.c: Handle new types.
2249 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
2250 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
2251 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
2252 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
2253 (__builtin_ia32_kshiftridi): New.
2254 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
2256 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
2260 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
2261 define to rs6000_init_stack_protect_guard.
2262 (rs6000_init_stack_protect_guard): New function.
2264 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2265 Yunqiang Su <yunqiang.su@imgtec.com>
2267 * config.gcc (supported_defaults): Add madd4.
2268 (with_madd4): Add validation.
2269 (all_defaults): Add madd4.
2270 * config/mips/mips.opt (mmadd4): New option.
2271 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2273 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2275 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
2276 (ISA_HAS_FUSED_MADD4): Likewise.
2277 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
2278 * gcc/doc/install.texi (--with-madd4): Document the new option.
2280 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2282 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
2283 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
2284 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
2285 (aarch64_init_pauth_hint_builtins): New.
2286 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
2287 (aarch64_expand_builtin): Expand new builtins.
2289 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2291 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
2292 * combine-stack-adj.c (no_unhandled_cfa): Handle
2293 REG_CFA_TOGGLE_RA_MANGLE.
2294 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
2295 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
2296 info for return address signing.
2297 (aarch64_expand_epilogue): Likewise.
2299 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2301 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
2302 * config/aarch64/aarch64-protos.h
2303 (aarch64_return_address_signing_enabled): New declaration.
2304 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
2306 (aarch64_expand_prologue): Sign return address before it's pushed onto
2308 (aarch64_expand_epilogue): Authenticate return address fetched from
2310 (aarch64_override_options): Sanity check for ILP32 and ISA level.
2311 (aarch64_attributes): New function attributes for "sign-return-address".
2312 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
2313 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
2314 ("*do_return"): Generate combined instructions according to key index.
2315 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
2316 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
2318 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
2319 * config/aarch64/aarch64.opt (msign-return-address=): New.
2320 * doc/extend.texi (AArch64 Function Attributes): Documents
2321 "sign-return-address=".
2322 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
2324 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2326 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
2327 overall option summary.
2329 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2331 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
2332 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
2333 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
2334 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
2336 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
2338 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
2339 -mpower9-minmax by default for -mcpu=power9.
2340 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
2341 128-bit floating point.
2343 2017-01-20 Alan Modra <amodra@gmail.com>
2345 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
2346 optimizing for size.
2348 2017-01-20 Alan Modra <amodra@gmail.com>
2351 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
2352 for strcmp and strncmp from corresponding builtin decl.
2354 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2356 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
2357 instead of i386/rtems-64.h.
2358 * config/i386/rtems-64.h: Remove.
2360 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2364 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2366 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
2368 2017-01-19 Tamar Christina <tamar.christina@arm.com>
2370 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
2371 Change int to HOST_WIDE_INT.
2372 * config/aarch64/aarch64-protos.h
2373 (aarch64_simd_gen_const_vector_dup): Likewise.
2374 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
2376 2017-01-19 David Malcolm <dmalcolm@redhat.com>
2378 * langhooks-def.h (lhd_type_for_size): New decl.
2379 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
2380 * langhooks.c (lhd_type_for_size): New function, taken from
2383 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
2385 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
2386 define_bypass for CR latency.
2387 (power9-cracked-alu): Update bypass latency and remove power9-branch.
2388 (power9-alu2): Add define_bypass for CR latency.
2390 (power9-mul): Update insn latency.
2391 (power9-mul-compare): Update insn latency, bypass latency and remove
2394 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2396 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
2398 * config/aarch64/aarch64.md
2399 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
2400 aarch64_nopcrelative_literal_loads.
2401 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
2403 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
2405 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
2407 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
2409 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
2412 * config.gcc (supported_defaults): Add lxc1-sxc1.
2413 (with_lxc1_sxc1): Add validation.
2414 (all_defaults): Add lxc1-sxc1.
2415 * config/mips/mips.opt (mlxc1-sxc1): New option.
2416 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2418 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2419 __mips_no_lxc1_sxc1.
2420 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
2421 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
2422 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
2424 2017-01-19 Richard Biener <rguenther@suse.de>
2426 PR tree-optimization/72488
2427 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
2428 sure to restore SSA info.
2429 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
2431 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
2433 PR rtl-optimization/79121
2434 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
2435 of the inner type when shifting an extended value.
2437 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2440 * symtab.c (symtab_node::equal_address_to): Fix comparing of
2441 interposable aliases.
2443 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
2446 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
2447 Use the evmergelohi instruction.
2448 (mov_si<mode>_e500_subreg4_2_le): Likewise.
2449 (mov_sitf_e500_subreg8_2_be): Likewise.
2450 (mov_sitf_e500_subreg12_2_le): Likewise.
2451 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
2452 (mov_si<mode>_e500_subreg4_2_be): Likewise.
2453 (mov_sitf_e500_subreg8_2_le): Likewise.
2454 (mov_sitf_e500_subreg12_2_be): Likewise.
2456 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2458 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
2459 attribute from vecsimple to vecperm.
2460 (altivec_vbpermq2): Likewise.
2462 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2465 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
2467 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
2468 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
2469 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
2470 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
2471 case where N arg is SIZE_MAX.
2472 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
2473 (cmpstrsi): Add pattern.
2475 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2477 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2478 __builtin_vec_revb builtins.
2479 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
2480 built-in functions to support generation of the ISA 3.0 XXBR<x>
2481 vector byte reverse instructions.
2482 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
2483 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
2484 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
2485 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
2486 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
2487 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
2488 (P9V_BUILTIN_VEC_REVB): Likewise.
2489 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
2490 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
2491 (p9_xxbrq_v16qi): Likewise.
2492 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
2493 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
2494 (p9_xxbrh_v8hi): Likewise.
2495 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
2496 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
2497 vec_revb built-in functions.
2499 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
2501 PR rtl-optimization/78952
2502 * config/i386/i386.md (any_extract): New code iterator.
2503 (*insvqi_2): Use any_extract for source operand.
2504 (*insvqi_3): Use any_shiftrt for source operand.
2506 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
2508 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
2510 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
2512 2017-01-18 Matthias Klose <doko@ubuntu.com>
2514 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
2516 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2518 * config/rs6000/altivec.h (vec_bperm): Change #define.
2519 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
2520 (altivec_vbpermq2): New define_insn.
2521 (altivec_vbpermd): Likewise.
2522 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
2524 (VBPERMD): Likewise.
2525 (VBPERM): New polymorphic function interface.
2526 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
2527 Add entries for P9V_BUILTIN_VEC_VBPERM.
2528 * doc/extend.texi: Add interfaces for vec_bperm.
2530 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2532 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
2533 first letter of error messages.
2534 (s390_resolve_overloaded_builtin): Likewise.
2535 * config/s390/s390.c (s390_expand_builtin): Likewise.
2536 (s390_invalid_arg_for_unprototyped_fn): Likewise.
2537 (s390_valid_target_attribute_inner_p): Likewise.
2538 * config/s390/s390.md ("tabort"): Likewise.
2540 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
2542 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
2543 (ISA_AVOID_DIV_HILO): New macro.
2544 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
2545 (ISA_HAS_DDIV): Likewise.
2547 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2549 * doc/invoke.texi (fabi-version): Correct number of occurrences.
2551 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2553 * doc/invoke.texi (fabi-version): Spelling fix.
2555 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2558 * doc/invoke.texi (fabi-version): Mention mangling fix for
2561 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2564 * doc/invoke.texi (fabi-version): Document discriminator mangling.
2566 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
2569 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
2570 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
2572 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
2574 (stack_protect_test): Ditto.
2575 * config/rs6000/rs6000.opt (mstack-protector-guard=,
2576 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
2578 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
2579 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
2580 -mstack-protector-guard-offset=.
2581 (RS/6000 and PowerPC Options): Ditto.
2583 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2585 * config/i386/i386.h (MASK_CLASS_P): New define.
2586 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
2587 there are no registers from different register sets also when
2588 mask registers are used. Update function comment.
2589 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
2590 to (*k/*r) and (*k/*km) alternatives.
2592 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
2594 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
2595 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
2596 (EH_RETURN_HANDLER_RTX): New define.
2597 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
2598 Force frame pointer in EH return functions.
2599 (aarch64_expand_epilogue): Add barrier for eh_return.
2600 (aarch64_final_eh_return_addr): Remove.
2601 (aarch64_eh_return_handler_rtx): New function.
2602 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
2604 (aarch64_eh_return_handler_rtx): New prototype.
2606 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2608 * config/rs6000/altivec.h (vec_rlmi): New #define.
2609 (vec_vrlnm): Likewise.
2610 (vec_rlnm): Likewise.
2611 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
2612 (UNSPEC_VRLNM): Likewise.
2613 (VIlong): New mode iterator.
2614 (altivec_vrl<VI_char>mi): New define_insn.
2615 (altivec_vrl<VI_char>nm): Likewise.
2616 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
2619 (RLNM): New polymorphic function entry.
2620 (VRLWMI): New monomorphic function entry.
2622 (RLMI): New polymorphic function entry.
2623 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
2624 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
2625 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
2628 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2631 * dwarf2out.c (field_byte_offset): Restore the
2632 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
2633 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
2634 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
2637 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2640 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
2642 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2645 * dwarf2out.c (add_data_member_location_attribute): For constant
2646 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
2647 instead of DW_AT_data_member_location, DW_AT_bit_offset and
2648 DW_AT_byte_size attributes.
2650 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2652 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
2653 after forcing to constant memory when the code model is medium.
2655 2017-01-17 Julia Koval <julia.koval@intel.com>
2658 * config/i386/avx512fintrin.h
2659 (_mm512_i32gather_ps): Change __addr type to void const*.
2660 (_mm512_mask_i32gather_ps): Ditto.
2661 (_mm512_i32gather_pd): Ditto.
2662 (_mm512_mask_i32gather_pd): Ditto.
2663 (_mm512_i64gather_ps): Ditto.
2664 (_mm512_mask_i64gather_ps): Ditto.
2665 (_mm512_i64gather_pd): Ditto.
2666 (_mm512_mask_i64gather_pd): Ditto.
2667 (_mm512_i32gather_epi32): Ditto.
2668 (_mm512_mask_i32gather_epi32): Ditto.
2669 (_mm512_i32gather_epi64): Ditto.
2670 (_mm512_mask_i32gather_epi64): Ditto.
2671 (_mm512_i64gather_epi32): Ditto.
2672 (_mm512_mask_i64gather_epi32): Ditto.
2673 (_mm512_i64gather_epi64): Ditto.
2674 (_mm512_mask_i64gather_epi64): Ditto.
2675 (_mm512_i32scatter_ps): Change __addr type to void*.
2676 (_mm512_mask_i32scatter_ps): Ditto.
2677 (_mm512_i32scatter_pd): Ditto.
2678 (_mm512_mask_i32scatter_pd): Ditto.
2679 (_mm512_i64scatter_ps): Ditto.
2680 (_mm512_mask_i64scatter_ps): Ditto.
2681 (_mm512_i64scatter_pd): Ditto.
2682 (_mm512_mask_i64scatter_pd): Ditto.
2683 (_mm512_i32scatter_epi32): Ditto.
2684 (_mm512_mask_i32scatter_epi32): Ditto.
2685 (_mm512_i32scatter_epi64): Ditto.
2686 (_mm512_mask_i32scatter_epi64): Ditto.
2687 (_mm512_i64scatter_epi32): Ditto.
2688 (_mm512_mask_i64scatter_epi32): Ditto.
2689 (_mm512_i64scatter_epi64): Ditto.
2690 (_mm512_mask_i64scatter_epi64): Ditto.
2691 * config/i386/avx512pfintrin.h
2692 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
2693 (_mm512_mask_prefetch_i32gather_ps): Ditto.
2694 (_mm512_mask_prefetch_i64gather_pd): Ditto.
2695 (_mm512_mask_prefetch_i64gather_ps): Ditto.
2696 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
2697 (_mm512_prefetch_i32scatter_ps): Ditto.
2698 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2699 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2700 (_mm512_prefetch_i64scatter_pd): Ditto.
2701 (_mm512_prefetch_i64scatter_ps): Ditto.
2702 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2703 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2704 * config/i386/avx512vlintrin.h
2705 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
2706 (_mm_mmask_i32gather_ps): Ditto.
2707 (_mm256_mmask_i32gather_pd): Ditto.
2708 (_mm_mmask_i32gather_pd): Ditto.
2709 (_mm256_mmask_i64gather_ps): Ditto.
2710 (_mm_mmask_i64gather_ps): Ditto.
2711 (_mm256_mmask_i64gather_pd): Ditto.
2712 (_mm_mmask_i64gather_pd): Ditto.
2713 (_mm256_mmask_i32gather_epi32): Ditto.
2714 (_mm_mmask_i32gather_epi32): Ditto.
2715 (_mm256_mmask_i32gather_epi64): Ditto.
2716 (_mm_mmask_i32gather_epi64): Ditto.
2717 (_mm256_mmask_i64gather_epi32): Ditto.
2718 (_mm_mmask_i64gather_epi32): Ditto.
2719 (_mm256_mmask_i64gather_epi64): Ditto.
2720 (_mm_mmask_i64gather_epi64): Ditto.
2721 (_mm256_i32scatter_ps): Change __addr type to void*.
2722 (_mm256_mask_i32scatter_ps): Ditto.
2723 (_mm_i32scatter_ps): Ditto.
2724 (_mm_mask_i32scatter_ps): Ditto.
2725 (_mm256_i32scatter_pd): Ditto.
2726 (_mm256_mask_i32scatter_pd): Ditto.
2727 (_mm_i32scatter_pd): Ditto.
2728 (_mm_mask_i32scatter_pd): Ditto.
2729 (_mm256_i64scatter_ps): Ditto.
2730 (_mm256_mask_i64scatter_ps): Ditto.
2731 (_mm_i64scatter_ps): Ditto.
2732 (_mm_mask_i64scatter_ps): Ditto.
2733 (_mm256_i64scatter_pd): Ditto.
2734 (_mm256_mask_i64scatter_pd): Ditto.
2735 (_mm_i64scatter_pd): Ditto.
2736 (_mm_mask_i64scatter_pd): Ditto.
2737 (_mm256_i32scatter_epi32): Ditto.
2738 (_mm256_mask_i32scatter_epi32): Ditto.
2739 (_mm_i32scatter_epi32): Ditto.
2740 (_mm_mask_i32scatter_epi32): Ditto.
2741 (_mm256_i32scatter_epi64): Ditto.
2742 (_mm256_mask_i32scatter_epi64): Ditto.
2743 (_mm_i32scatter_epi64): Ditto.
2744 (_mm_mask_i32scatter_epi64): Ditto.
2745 (_mm256_i64scatter_epi32): Ditto.
2746 (_mm256_mask_i64scatter_epi32): Ditto.
2747 (_mm_i64scatter_epi32): Ditto.
2748 (_mm_mask_i64scatter_epi32): Ditto.
2749 (_mm256_i64scatter_epi64): Ditto.
2750 (_mm256_mask_i64scatter_epi64): Ditto.
2751 (_mm_i64scatter_epi64): Ditto.
2752 (_mm_mask_i64scatter_epi64): Ditto.
2753 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
2754 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
2755 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
2756 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
2757 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
2758 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
2759 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
2760 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
2761 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
2762 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
2763 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
2764 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
2765 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
2766 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
2767 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
2768 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
2769 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
2770 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
2771 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
2772 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
2773 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
2774 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
2775 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
2776 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
2777 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
2778 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
2779 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
2780 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
2781 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
2782 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
2783 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
2784 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
2785 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
2786 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
2787 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
2788 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
2789 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
2790 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
2791 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
2792 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
2793 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
2794 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
2795 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
2796 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
2797 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
2798 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
2799 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
2800 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
2801 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
2802 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
2803 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
2804 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
2805 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
2806 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
2807 definitions accordingly.
2809 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
2810 Kuan-Lin Chen <kuanlinchentw@gmail.com>
2813 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
2816 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
2819 * ira-conflicts.c (ira_build_conflicts): Update total conflict
2820 hard regs for inner regno.
2822 2017-01-17 Martin Liska <mliska@suse.cz>
2825 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
2826 assumption and add comment.
2828 2017-01-17 Nathan Sidwell <nathan@acm.org>
2830 * ipa-visibility.c (localize_node): New function, broken out of ...
2831 (function_and_variable_visibility): ... here. Call it.
2833 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2836 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
2837 correctly set frequency of oudgoing edge.
2838 (duplicate_thread_path): Fix profile updating.
2840 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2843 * configure.ac: Add GCC_BASE_VER.
2844 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
2845 version from BASE-VER file.
2846 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
2847 (gcc.o): Depend on $(BASEVER).
2848 * common.opt (dumpfullversion): New option.
2849 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
2850 * doc/invoke.texi: Document -dumpfullversion.
2851 * doc/install.texi: Document --with-gcc-major-version-only.
2852 * configure: Regenerated.
2854 2017-01-17 Richard Biener <rguenther@suse.de>
2856 PR tree-optimization/71433
2857 * tree-vrp.c (register_new_assert_for): Merge same asserts
2858 on all incoming edges.
2859 (process_assert_insertions_for): Handle insertions at the
2862 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
2864 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
2865 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
2867 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
2870 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
2873 2017-01-17 Alan Modra <amodra@gmail.com>
2876 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
2877 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
2878 symbolic stack limit when pic.
2880 2017-01-16 Martin Sebor <msebor@redhat.com>
2882 PR tree-optimization/78608
2883 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
2885 2017-01-16 Jeff Law <law@redhat.com>
2888 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
2889 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
2890 for several include directories that may be relative to sysroot.
2891 * config/i386/x-mingw32 (gplus_includedir): Define.
2892 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
2893 (native_system_includedir): Likewise.
2894 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
2895 override if TARGET_SYSTEM_ROOT is defined.
2896 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
2898 PR tree-optimization/79090
2899 PR tree-optimization/33562
2900 PR tree-optimization/61912
2901 PR tree-optimization/77485
2902 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
2903 and computed trims into the dump file.
2905 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2907 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
2909 2017-01-16 Jakub Jelinek <jakub@redhat.com>
2912 * gimplify.c (gimplify_init_constructor): If want_value and
2913 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
2917 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
2918 sequence. Formatting fixes.
2919 (doloop_optimize): Formatting fixes.
2922 * gcc.c (debug_level_greater_than_spec_func): New function.
2923 (static_spec_functions): Add debug-level-gt spec function.
2924 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2926 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2927 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2928 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2929 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2930 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2931 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2933 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2935 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2936 QImode fixups to general and mask registers only.
2938 2017-01-16 Carl Love <cel@us.ibm.com>
2940 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2941 for built-in functions
2942 vector signed char vec_nabs (vector signed char)
2943 vector signed short vec_nabs (vector signed short)
2944 vector signed int vec_nabs (vector signed int)
2945 vector signed long long vec_nabs (vector signed long long)
2946 vector float vec_nabs (vector float)
2947 vector double vec_nabs (vector double)
2948 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2950 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2951 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2952 * doc/extend.texi: Update the documentation file for the new built-in
2955 2017-01-16 Martin Sebor <msebor@redhat.com>
2957 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2960 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2962 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2963 UNSPEC_VSX__XXSPLTD to require special splat handling.
2965 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2968 * system.h: Poison strndup.
2970 2017-01-16 Alan Modra <amodra@gmail.com>
2973 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2976 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2978 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2980 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2982 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2983 call recog here. Assert that INSN_CODE (insn) is non-negative.
2985 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2988 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2990 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2991 in the currently scheduled RTL fragment.
2993 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2995 PR rtl-optimization/78751
2996 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2999 2017-01-14 Jeff Law <law@redhat.com>
3001 PR tree-optimization/79090
3002 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
3003 variable length stores.
3004 (compute_trims): Delete dead assignment to *trim_tail.
3005 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
3008 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
3010 PR rtl-optimization/78626
3011 PR rtl-optimization/78727
3012 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
3013 of a block, and split such blocks after everything else is finished.
3015 2017-01-14 Alan Modra <amodra@gmail.com>
3018 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
3019 target legitimate_combined_insn.
3020 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
3021 (rs6000_legitimate_combined_insn): New function.
3022 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
3024 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
3025 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
3026 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
3028 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
3030 * doc/frontends.texi (G++ and GCC): Remove references to Java.
3032 2017-01-13 Jeff Law <law@redhat.com>
3034 PR tree-optimization/33562
3035 PR tree-optimization/61912
3036 PR tree-optimization/77485
3037 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
3039 (delete_dead_assignment): Likewise.
3040 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
3041 statement to delete_dead_call and delete_dead_assignment.
3043 2017-01-13 David Malcolm <dmalcolm@redhat.com>
3046 * substring-locations.c (format_warning_va): Strengthen case 1 so
3047 that both endpoints of the substring must be within the format
3048 range for just the substring to be printed.
3050 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
3052 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
3053 * config/i386/i386.c (ix86_target_string): Add missing options
3054 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
3055 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
3056 flags_other and ix86_target_other to flags2_other. Display unknown
3058 (ix86_valid_target_attribute_inner_p): Add missing options and
3059 reorder options by implied ISAs, as in ix86_target_string.
3061 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3063 * hash-table.h (hash_table::too_empty_p): New function.
3064 (hash_table::expand): Use it.
3065 (hash_table::traverse): Likewise.
3066 (hash_table::empty_slot): Use sizeof (value_type) instead of
3067 sizeof (PTR) to convert bytes to elements. Shrink the table
3068 if the current size is excessive for the current number of
3071 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3073 * ira-costs.c (record_reg_classes): Break from the inner loop
3074 early once alt_fail is known to be true. Update outer loop
3075 handling accordingly.
3077 2017-01-13 Jeff Law <law@redhat.com>
3079 * tree-ssa-dse.c (decrement_count): New function.
3080 (increment_start_addr, maybe_trim_memstar_call): Likewise.
3081 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
3082 when we know the partially dead statement is a mem* function.
3084 PR tree-optimization/61912
3085 PR tree-optimization/77485
3086 * tree-ssa-dse.c: Include expr.h.
3087 (maybe_trim_constructor_store): New function.
3088 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
3090 PR tree-optimization/33562
3091 PR tree-optimization/61912
3092 PR tree-optimization/77485
3093 * doc/invoke.texi: Document new dse-max-object-size param.
3094 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
3095 * tree-ssa-dse.c: Include params.h.
3096 (dse_store_status): New enum.
3097 (initialize_ao_ref_for_dse): New, partially extracted from
3099 (valid_ao_ref_for_dse, normalize_ref): New.
3100 (setup_live_bytes_from_ref, compute_trims): Likewise.
3101 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
3102 (maybe_trim_partially_dead_store): Likewise.
3103 (maybe_trim_complex_store): Likewise.
3104 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
3105 Track what bytes live from the original store. Return tri-state
3106 for dead, partially dead or live.
3107 (dse_dom_walker): Add constructor, destructor and new private members.
3108 (delete_dead_call, delete_dead_assignment): New extracted from
3110 (dse_optimize_stmt): Make a member of dse_dom_walker.
3111 Use initialize_ao_ref_for_dse.
3113 PR tree-optimization/33562
3114 PR tree-optimization/61912
3115 PR tree-optimization/77485
3116 * sbitmap.h (bitmap_count_bits): Prototype.
3117 (bitmap_clear_range, bitmap_set_range): Likewise.
3118 * sbitmap.c (bitmap_clear_range): New function.
3119 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
3121 2017-01-13 Martin Liska <mliska@suse.cz>
3124 * function.c (set_cfun): Add new argument force.
3125 * function.h (set_cfun): Likewise.
3126 * ipa-inline-transform.c (inline_call): Use the function when
3127 strict alising from is dropped for function we inline to.
3129 2017-01-13 Richard Biener <rguenther@suse.de>
3131 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
3132 for dumping GIMPLE INTEGER_CSTs.
3134 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3136 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
3137 to 201112L since C++17.
3139 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
3142 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
3143 if -fsanitize=kernel-address is present.
3145 2017-01-13 Richard Biener <rguenther@suse.de>
3147 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
3148 as _Literal ( type ) number in case usual suffixes do not
3149 preserve all information.
3151 2017-01-13 Richard Biener <rguenther@suse.de>
3153 PR tree-optimization/77283
3154 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
3155 and ssa-iterators.h.
3156 (is_feasible_trace): Implement a cost model based on joiner
3159 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3162 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
3163 char or short to __float128/_Float128 directly.
3165 2017-01-12 Martin Sebor <msebor@redhat.com>
3167 to -Wformat-overflow.
3168 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
3169 (min_bytes_remaining): Same.
3170 (get_string_length): Same.
3171 (format_string): Same.
3172 (format_directive): Same.
3174 (pass_sprintf_length::handle_gimple_call): Same.
3176 2017-01-12 Jakub Jelinek <jakub@redhat.com>
3178 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
3179 info.nowrite calls with no lhs that can't throw. Return bool
3180 whether gsi_remove has been called or not.
3181 (pass_sprintf_length::handle_gimple_call): Return bool whether
3182 try_substitute_return_value called gsi_remove. Formatting fix.
3183 (pass_sprintf_length::execute): Don't use gsi_remove if
3184 handle_gimple_call returned true.
3187 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
3188 be removed due to side-effects, don't remove following barrier nor
3189 turn the successor edge into fallthru edge.
3191 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3194 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
3195 element-reversing loads and stores as not swappable.
3197 2017-01-12 Nathan Sidwell <nathan@acm.org>
3198 Nicolai Stange <nicstange@gmail.com>
3200 * combine.c (try_combine): Don't ignore result of overlap checking
3201 loop. Combine overlap & asm check into single loop.
3203 2017-01-12 Richard Biener <rguenther@suse.de>
3205 * tree-pretty-print.c (dump_generic_node): Provide -gimple
3206 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
3208 2017-01-12 Richard Biener <rguenther@suse.de>
3210 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
3211 and TS_TARGET_OPTION directly derive from TS_BASE.
3212 * tree-core.h (tree_optimization_option): Derive from tree_base.
3213 (tree_target_option): Likewise.
3215 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
3217 * config/i386/i386.c (memory_address_length): Increase len
3218 only when rip_relative_addr_p returns false.
3220 2017-01-11 Julia Koval <julia.koval@intel.com>
3222 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
3223 (OPTION_MASK_ISA_SGX_SET): New.
3224 (ix86_handle_option): Handle OPT_msgx.
3225 * config.gcc: Added sgxintrin.h.
3226 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
3227 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
3228 * config/i386/i386.c (ix86_target_string): Add -msgx.
3230 (ix86_option_override_internal): Handle new options.
3231 (ix86_valid_target_attribute_inner_p): Add sgx.
3232 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
3233 * config/i386/i386.opt: Add msgx.
3234 * config/i386/sgxintrin.h: New file.
3235 * config/i386/x86intrin.h: Add sgxintrin.h.
3237 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3240 * fold-const.c (maybe_nonzero_address): Return 1 for function
3242 (tree_single_nonzero_warnv_p): Don't handle function local objects
3246 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
3249 2017-01-11 David Malcolm <dmalcolm@redhat.com>
3252 * opts.c: Include "spellcheck.h"
3253 (struct string_fragment): New struct.
3254 (struct edit_distance_traits<const string_fragment &>): New
3256 (get_closest_sanitizer_option): New function.
3257 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
3259 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3261 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
3263 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
3264 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
3265 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
3266 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
3267 for initial die_offset if dwarf_split_debug_info.
3268 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
3269 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
3271 (output_skeleton_debug_sections): Formatting fix. Use
3272 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
3273 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
3275 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
3277 * config/arm/cortex-a53.md: Add bypasses for
3279 (cortex_a53_r2f): Only use for transfers.
3280 (cortex_a53_f2r): Likewise.
3281 (cortex_a53_r2f_cvt): Add reservation for conversions.
3282 (cortex_a53_f2r_cvt): Likewise.
3284 2017-01-11 Tamar Christina <tamar.christina@arm.com>
3286 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
3287 to all inlined functions, change static to extern.
3289 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
3292 * config/arm/arm.c (legitimize_pic_address): Handle reference to
3294 (arm_assemble_integer): Likewise.
3296 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
3298 * config.gcc: Use new awk script to check CPU, FPU and architecture
3299 parameters for --with-... options.
3300 * config/arm/parsecpu.awk: New file
3301 * config/arm/arm-cpus.in: New file.
3302 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
3304 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
3306 * config/arm/t-arm: Update dependency rules.
3307 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
3308 of processing .def files.
3309 * config/arm/genopt.sh: Deleted.
3310 * config/arm/gentune.sh: Deleted.
3311 * config/arm/arm-cores.def: Deleted.
3312 * config/arm/arm-arches.def: Deleted.
3313 * config/arm/arm-fpus.def: Deleted.
3314 * config/arm/arm-tune.md: Regenerated.
3315 * config/arm/arm-tables.opt: Regenerated.
3316 * config/arm/arm-cpu.h: New generated file.
3317 * config/arm/arm-cpu-data.h: New generated file.
3318 * config/arm/arm-cpu-cdata.h: New generated file.
3320 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
3323 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
3325 (input_varpool_node): Unpack dynamically_initialized bit.
3327 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
3329 PR rtl-optimization/79032
3330 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
3331 the alignment of the adjusted memory reference against that of MODE,
3332 instead of the alignment of the original memory reference.
3334 2017-01-11 Martin Jambor <mjambor@suse.cz>
3336 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
3338 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
3339 decorated functions.
3341 2017-01-11 Richard Biener <rguenther@suse.de>
3343 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
3344 set range/nonnull info for PHI results. Do not set it on
3345 stmts marked for removal.
3347 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
3349 * expr.c (store_field): In the bitfield case, fetch the return value
3350 from the registers before applying a single big-endian adjustment.
3351 Always do a final load for a BLKmode value not larger than a word.
3353 2017-01-10 David Malcolm <dmalcolm@redhat.com>
3356 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3357 that we correctly handle column numbers greater than
3358 LINE_MAP_MAX_COLUMN_NUMBER.
3360 2017-01-10 Martin Sebor <msebor@redhat.com>
3363 * gimple-ssa-sprintf.c (get_destination_size): Call
3364 {init,fini}object_sizes.
3365 * tree-object-size.c (addr_object_size): Adjust.
3366 (pass_through_call): Adjust.
3367 (pass_object_sizes::execute): Adjust.
3368 * tree-object-size.h (fini_object_sizes): Declare.
3370 2017-01-10 Martin Sebor <msebor@redhat.com>
3372 PR tree-optimization/78775
3373 * builtins.c (get_size_range): Move...
3374 * calls.c: ...to here.
3375 (alloc_max_size): Accept zero argument.
3376 (operand_signed_p): Remove.
3377 (maybe_warn_alloc_args_overflow): Call get_size_range.
3378 * calls.h (get_size_range): Declare.
3380 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
3382 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
3383 from TI's devices.csv file as of September 2016.
3384 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
3386 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
3388 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
3389 * doc/invoke.texi: Likewise.
3390 * doc/md.texi: Likewise.
3391 * doc/objc.texi: Likewise.
3393 2017-01-10 Joshua Conner <joshconner@google.com>
3395 * config/arm/fuchsia-elf.h: New file.
3396 * config/fuchsia.h: New file.
3397 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
3398 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
3400 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
3402 2016-01-10 Richard Biener <rguenther@suse.de>
3404 PR tree-optimization/79034
3405 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
3406 Propagate out degenerate PHIs in the joiner.
3408 2017-01-10 Martin Liska <mliska@suse.cz>
3410 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
3411 (sort_congruence_classes_by_decl_uid): Likewise.
3412 (sort_congruence_class_groups_by_decl_uid): Likewise.
3413 (sem_item_optimizer::merge_classes): Sort class, groups in these
3414 classes and members in the groups by DECL_UID of declarations.
3415 This would make merge operations stable.
3417 2017-01-10 Martin Liska <mliska@suse.cz>
3419 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
3420 usage of m_classes_vec.
3421 (sem_item_optimizer::~sem_item_optimizer): Likewise.
3422 (sem_item_optimizer::get_group_by_hash): Likewise.
3423 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
3424 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
3425 (sem_item_optimizer::verify_classes): Likewise.
3426 (sem_item_optimizer::process_cong_reduction): Likewise.
3427 (sem_item_optimizer::dump_cong_classes): Likewise.
3428 (sem_item_optimizer::merge_classes): Likewise.
3429 * ipa-icf.h (congruence_class_hash): Rename from
3430 congruence_class_group_hash. Remove declaration of m_classes_vec.
3432 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
3434 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
3435 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
3436 * config.gcc: Add avx512vpopcntdqintrin.h.
3437 * config/i386/avx512vpopcntdqintrin.h: New.
3438 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
3439 * config/i386/i386-builtin-types.def: Add new types.
3440 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
3441 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
3442 __builtin_ia32_vpopcountq_v8di_mask): New.
3443 * config/i386/i386-c.c (ix86_target_macros_internal): Define
3444 __AVX512VPOPCNTDQ__.
3445 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
3446 (PTA_AVX512VPOPCNTDQ): Define.
3447 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
3448 TARGET_AVX512VPOPCNTDQ_P): Define.
3449 * config/i386/i386.opt: Add mavx512vpopcntdq.
3450 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
3451 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
3453 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3456 * predict.def (PRED_CALL): Set to 67.
3458 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
3460 * expr.c (store_field): In the bitfield case, if the value comes from
3461 a function call and is of an aggregate type returned in registers, do
3462 not modify the field mode; extract the value in all cases if the mode
3463 is BLKmode and the size is not larger than a word.
3465 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
3468 * config/i386/cpuid.h: Fix undefined behavior.
3470 2017-01-04 Jeff Law <law@redhat.com>
3472 PR tree-optimization/79007
3473 PR tree-optimization/67955
3474 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
3475 conservative for pt.null when flag_non_call_exceptions is on.
3477 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3479 PR translation/79019
3480 PR translation/79020
3481 * params.def (PARAM_INLINE_MIN_SPEEDUP,
3482 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
3483 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
3485 * config/avr/avr.opt (maccumulate-args): Likewise.
3486 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
3487 * common.opt (freport-bug): Likewise.
3488 * cif-code.def (CIF_FINAL_ERROR): Likewise.
3489 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
3490 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
3491 translatable string.
3492 * config/i386/i386.c (function_value_32): Likewise.
3493 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
3494 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
3496 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
3497 * common/config/msp430/msp430-common.c (msp430_handle_option):
3499 * symtab.c (symtab_node::verify_base): Likewise.
3500 * opts.c (set_debug_level): Likewise.
3501 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
3502 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
3503 missing whitespace to translatable strings.
3504 * config/avr/avr.md (bswapsi2): Fix typo in comment.
3505 * config/sh/superh.h: Likewise.
3506 * config/i386/xopintrin.h: Likewise.
3507 * config/i386/znver1.md: Likewise.
3508 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
3509 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
3510 * double-int.h (struct double_int): Likewise.
3511 * double-int.c (div_and_round_double): Likewise.
3512 * wide-int.cc: Likewise.
3513 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3514 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
3515 * cfgcleanup.c (crossjumps_occured): Renamed to ...
3516 (crossjumps_occurred): ... this.
3517 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
3520 PR tree-optimization/78899
3521 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
3522 returning bool return struct loop *, NULL for failure and the new
3524 (versionable_outer_loop_p): Don't version outer loop if it has
3525 dont_vectorized bit set.
3526 (tree_if_conversion): When versioning outer loop, ensure
3527 tree_if_conversion is performed also on the inner loop of the
3528 non-vectorizable outer loop copy.
3529 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
3530 LOOP_VECTORIZED in inner loop of the scalar outer loop and
3531 prevent vectorization of it.
3532 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
3533 the outer loop vectorization of the non-scalar version is attempted
3534 before vectorization of the inner loop in scalar version. If
3535 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
3536 vectorization of its inner loop.
3537 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
3538 has 2 inner loops, rename also on edges from bb whose single pred
3539 is outer_loop->header. Fix typo in function comment.
3541 2017-01-09 Martin Sebor <msebor@redhat.com>
3544 * asan.c (asan_emit_stack_protection): Increase local buffer size
3545 to avoid snprintf truncation warning.
3547 2017-01-09 Andrew Pinski <apinski@cavium.com>
3549 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
3550 to reference thunderx2t99 for the tuning structure
3551 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
3553 (thunderx2t99_extra_costs): This.
3554 * config/aarch64/aarch64-tune.md: Regenerate.
3555 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
3556 (vulcan_addrcost_table): This.
3557 (vulcan_regmove_cost): Rename to ...
3558 (thunderx2t99_regmove_cost): This.
3559 (vulcan_vector_cost): Rename to ...
3560 (thunderx2t99_vector_cost): this.
3561 (vulcan_branch_cost): Rename to ...
3562 (thunderx2t99_branch_cost): This.
3563 (vulcan_tunings): Rename to ...
3564 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
3565 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
3567 2017-01-09 Martin Jambor <mjambor@suse.cz>
3571 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
3572 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
3573 (propagate_vr_accross_jump_function): Use the above function for all
3574 value range computations for pass-through jump functions and type
3575 converasion from explicit value range values.
3576 (ipcp_propagate_stage): Do not attempt to deduce types of formal
3577 parameters from TYPE_ARG_TYPES.
3578 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
3579 (ipa_write_node_info): Stream type of the actual argument.
3580 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
3582 2017-01-09 Martin Liska <mliska@suse.cz>
3585 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
3586 (lookup_compiler): Do not show error message with have_E.
3588 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3590 PR tree-optimization/78938
3591 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
3592 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
3593 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
3594 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
3597 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3599 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
3602 2017-01-09 Richard Biener <rguenther@suse.de>
3604 PR tree-optimization/78997
3605 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
3606 name condition properly.
3608 2017-01-09 Richard Biener <rguenther@suse.de>
3611 * dwarf2out.c (is_cxx): New overload with context.
3612 (is_naming_typedef_decl): Use it.
3614 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3616 * invoke.texi (Option Summary): Correct spacing in option lists
3617 and add line breaks to fix over-long lines.
3619 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3623 * extend.texi (Common Variable Attributes): Add xref to GCC
3624 Internals manual to explain mode attribute keywords.
3626 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3629 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
3630 and Preprocessor Options.
3631 (Options for Linking): Document -pthread here....
3632 (RS/6000 and PowerPC Options): ...not here.
3633 (Solaris 2 Options): ...or here.
3634 * doc/cppopts.texi: Document -pthread.
3636 2017-01-08 Martin Sebor <msebor@redhat.com>
3639 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
3640 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
3641 New member functions.
3642 (format_directive): Used them.
3644 (pass_sprintf_length::handle_gimple_call): Same.
3645 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
3646 to avoid truncation for any argument.
3647 (extract_affine_mul): Same.
3648 * tree.c (get_file_function_name): Same.
3650 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3653 * predict.def (PRED_INDIR_CALL): Set to 86.
3655 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3657 PR preprocessor/54124
3658 * doc/cppopts.texi: Reformat -d subtable to list the full name
3659 of the options. Add cross-reference to the docs for the general
3660 compiler -d options.
3661 * doc/invoke.texi (Developer Options): Add cross-reference to the
3662 preprocessor-specific -d option documentation.
3664 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3666 PR preprocessor/13498
3667 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
3668 redudant material, and reflect new command-line options.
3669 (System Headers): Likewise.
3671 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3673 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
3674 -isystem, and -idirafter. Copy-edit.
3675 * doc/cppopts.texi: Copy-edit. Remove contradiction about
3676 default for -ftrack-macro-expansion. Delete obsolete and
3677 badly-formatted implementation details about -fdebug-cpp output.
3678 * doc/cppwarnopts.texi: Copy-edit.
3680 2017-01-07 David Malcolm <dmalcolm@redhat.com>
3683 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3684 that the transition from a max line width >= 1<<10 to narrower
3685 lines works correctly.
3687 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
3689 * doc/options.texi (PerFunction): New.
3690 * opt-functions.awk (switch_flags): Map both Optimization and
3691 PerFunction to CL_OPTIMIZATION.
3692 * opth-gen.awk: Test for PerFunction flag along with
3694 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
3695 it only when the latter is present. Skip those that don't in
3696 the hash function generator.
3697 * common.opt (fvar-tracking): Mark as PerFunction instead of
3699 (fvar-tracking-assignments): Likewise.
3700 (fvar-tracking-assignments-toggle): Likewise.
3701 (fvar-tracking-uninit): Likewise.
3703 2017-01-07 Jakub Jelinek <jakub@redhat.com>
3705 PR translation/79018
3706 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
3709 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
3712 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
3713 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
3714 TARGET_LONG_JUMP_TABLE_OFFSETS.
3715 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
3716 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
3717 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
3718 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
3719 * config/m68k/m68k.md (tablejump expander): Likewise.
3720 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
3721 TARGET_LONG_JUMP_TABLE_OFFSETS.
3722 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
3723 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
3725 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
3726 David Holsgrove <david.holsgrove@xilinx.com>
3728 * common/config/microblaze/microblaze-common.c
3729 (TARGET_EXCEPT_UNWIND_INFO): Remove.
3730 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
3732 * config/microblaze/microblaze.c (microblaze_must_save_register)
3733 (microblaze_expand_epilogue, microblaze_return_addr): Handle
3735 (microblaze_eh_return): New function.
3736 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
3737 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
3738 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
3739 * config/microblaze/microblaze.md (eh_return): New pattern.
3741 2017-01-06 Jakub Jelinek <jakub@redhat.com>
3743 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
3744 GCC_DIAGNOSTIC_STRINGIFY): Define.
3746 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
3748 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3750 * config/arm/arm.md (<mcrr>): New.
3752 * config/arm/arm.c (arm_arch5te): New.
3753 (arm_option_override): Set arm_arch5te.
3754 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
3756 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
3757 (arm_mcrr_qualifiers): ... this. New.
3758 (MRRC_QUALIFIERS): Define to...
3759 (arm_mrrc_qualifiers): ... this. New.
3760 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
3762 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
3763 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
3764 (MRRCI, mrrc, MRRC): New.
3765 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
3766 VUNSPEC_MRRC2): New.
3768 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3770 * config/arm/arm.md (<mcr>): New.
3772 * config/arm/arm.c (arm_coproc_builtin_available): Add
3773 support for mcr, mrc, mcr2 and mrc2.
3774 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
3775 (arm_mcr_qualifiers): ... this. New.
3776 (MRC_QUALIFIERS): Define to ...
3777 (arm_mrc_qualifiers): ... this. New.
3778 (MCR_QUALIFIERS): Define to ...
3779 (arm_mcr_qualifiers): ... this. New.
3780 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
3782 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
3783 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
3784 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
3787 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3789 * config/arm/arm.md (*ldc): New.
3793 * config/arm/arm.c (arm_coproc_builtin_available): Add
3794 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
3795 (arm_coproc_ldc_stc_legitimate_address): New.
3796 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
3797 'qualifier_const_pointer'.
3798 (LDC_QUALIFIERS): Define to...
3799 (arm_ldc_qualifiers): ... this. New.
3800 (STC_QUALIFIERS): Define to...
3801 (arm_stc_qualifiers): ... this. New.
3802 * config/arm/arm-protos.h
3803 (arm_coproc_ldc_stc_legitimate_address): New.
3804 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
3805 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
3806 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
3807 stc2, stcl, stc2l): New.
3808 * config/arm/constraints.md (Uz): New.
3809 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
3810 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
3811 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
3812 VUNSPEC_STC2L): New.
3814 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3816 * config/arm/arm.md (<cdp>): New.
3817 * config/arm/arm.c (neon_const_bounds): Rename this ...
3818 (arm_const_bounds): ... this.
3819 (arm_coproc_builtin_available): New.
3820 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
3821 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
3822 (CDP_QUALIFIERS): Define to...
3823 (arm_cdp_qualifiers): ... this. New.
3825 (arm_expand_builtin_args): Add case for 6 arguments.
3826 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
3827 (arm_const_bounds): ... this.
3828 (arm_coproc_builtin_available): New.
3829 * config/arm/arm_acle.h (__arm_cdp): New.
3831 * config/arm/arm_acle_builtins.def (cdp): New.
3833 * config/arm/iterators.md (CDPI,CDP,cdp): New.
3834 * config/arm/neon.md: Rename all 'neon_const_bounds' to
3836 * config/arm/types.md (coproc): New.
3837 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
3838 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
3839 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
3840 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
3842 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3844 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
3845 (UBINOP_QUALIFIERS): New.
3847 (acle_builtin_data): New. Change comment.
3848 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
3849 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
3850 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
3851 arm_acle_builtins.def.
3852 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
3853 (arm_init_acle_builtins): New.
3854 (CRC32_BUILTIN): Remove.
3855 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
3856 crc32cb, crc32ch and crc32cw.
3857 (arm_init_crc32_builtins): Remove.
3858 (arm_init_builtins): Use arm_init_acle_builtins rather
3859 than arm_init_crc32_builtins.
3860 (arm_expand_acle_builtin): New.
3861 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
3862 * config/arm/arm_acle_builtins.def: New.
3864 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3866 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
3867 (arm_builtin_datum): ... this.
3868 (arm_init_neon_builtin): Rename to ...
3869 (arm_init_builtin): ... this. Add a new parameters PREFIX
3870 and USE_SIG_IN_NAME.
3871 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
3872 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
3873 'arm_builtin_datum'.
3874 (arm_init_vfp_builtins): Likewise.
3875 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
3876 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
3877 (arm_expand_neon_args): Rename to ...
3878 (arm_expand_builtin_args): ... this. Rename builtin_arg
3879 enum values and differentiate between ARG_BUILTIN_MEMORY
3880 and ARG_BUILTIN_NEON_MEMORY.
3881 (arm_expand_neon_builtin_1): Rename to ...
3882 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
3883 values, arm_expand_builtin_args and add bool parameter NEON.
3884 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
3885 (arm_expand_vfp_builtin): Likewise.
3886 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
3888 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3891 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
3892 * predict.c (tree_estimate_probability_bb): Reverse direction of
3893 polymorphic call predictor.
3895 2017-01-06 David Malcolm <dmalcolm@redhat.com>
3897 * passes.c (execute_one_pass): Split out pass-skipping logic into...
3898 (determine_pass_name_match): ...this new function and...
3899 (should_skip_pass_p): ...this new function.
3901 2017-01-06 Nathan Sidwell <nathan@acm.org>
3903 * ipa-visibility.c (function_and_variable_visibility): Reformat
3904 comments and long lines. Remove extrneous if.
3905 * symtab.c (symtab_node::make_decl_local): Fix code format.
3906 (symtab_node::set_section_for_node): Fix comment typo.
3908 2017-01-06 Martin Liska <mliska@suse.cz>
3911 * lra-constraints.c: Rename invariant to lra_invariant.
3912 * predict.c (set_even_probabilities): Initialize e to NULL.
3914 2017-01-05 Martin Sebor <msebor@redhat.com>
3916 PR tree-optimization/78910
3917 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
3918 (format_integer): Correct off-by-one error in the handling
3919 of precision with negative numbers in signed conversions..
3921 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3923 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3925 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3927 PR tree-optimization/71016
3928 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3929 factor_out_conditional_conversion. Formatting fix.
3930 (factor_out_conditional_conversion): Add cond_stmt argument.
3931 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3932 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3935 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3937 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3938 read-rtl-function.o, and selftest-rtl.o.
3939 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3940 (selftest::aarch64_test_loading_full_dump): New function.
3941 (selftest::aarch64_run_selftests): New function.
3942 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3943 selftest::aarch64_run_selftests.
3944 * config/i386/i386.c
3945 (selftest::ix86_test_loading_dump_fragment_1): New function.
3946 (selftest::ix86_test_loading_call_insn): New function.
3947 (selftest::ix86_test_loading_full_dump): New function.
3948 (selftest::ix86_test_loading_unspec): New function.
3949 (selftest::ix86_run_selftests): Call the new functions.
3950 * emit-rtl.c (maybe_set_max_label_num): New function.
3951 * emit-rtl.h (maybe_set_max_label_num): New decl.
3952 * function.c (instantiate_decls): Guard call to
3953 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3954 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3956 * gensupport.c (gen_reader::gen_reader): Pass "false"
3957 for new "compact" param of rtx_reader.
3958 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3959 rather than an empty string for NULL strings.
3960 * read-md.c: Potentially include config.h rather than bconfig.h.
3961 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3962 (have_error): New global, copied from errors.c.
3963 (md_reader::read_name): Rename to...
3964 (md_reader::read_name_1): ...this, adding "out_loc" param,
3965 and converting "missing name or number" to returning false, rather
3967 (md_reader::read_name): Reimplement in terms of read_name_1.
3968 (md_reader::read_name_or_nil): New function.
3969 (md_reader::read_string): Handle "(nil)" by returning NULL.
3970 (md_reader::md_reader): Add new param "compact".
3971 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3972 (md_reader::read_file): New method.
3973 * read-md.h (md_reader::md_reader): Add new param "compact".
3974 (md_reader::read_file): New method.
3975 (md_reader::is_compact): New accessor.
3976 (md_reader::read_name): Convert return type from void to file_location.
3977 (md_reader::read_name_or_nil): New decl.
3978 (md_reader::read_name_1): New decl.
3979 (md_reader::m_compact): New field.
3980 (noop_reader::noop_reader): Pass "false" for new "compact" param
3982 (rtx_reader::rtx_reader): Add new "compact" param.
3983 (rtx_reader::read_rtx_operand): Make virtual and convert return
3984 type from void to rtx.
3985 (rtx_reader::read_until): New decl.
3986 (rtx_reader::handle_any_trailing_information): New virtual function.
3987 (rtx_reader::postprocess): New virtual function.
3988 (rtx_reader::finalize_string): New virtual function.
3989 (rtx_reader::m_in_call_function_usage): New field.
3990 (rtx_reader::m_reuse_rtx_by_id): New field.
3991 * read-rtl-function.c: New file.
3992 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3993 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3994 (selftest::verify_three_block_rtl_cfg): New decl.
3995 * read-rtl-function.h: New file.
3996 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3997 For host, include function.h, memmodel.h, and emit-rtl.h.
3998 (one_time_initialization): New function.
3999 (struct compact_insn_name): New struct.
4000 (compact_insn_names): New array.
4001 (find_code): Handle insn codes in compact dumps.
4002 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
4003 (bind_subst_iter_and_attr): Likewise.
4004 (add_condition_to_string): Likewise.
4005 (add_condition_to_rtx): Likewise.
4006 (apply_attribute_uses): Likewise.
4007 (add_current_iterators): Likewise.
4008 (apply_iterators): Likewise.
4009 (initialize_iterators): Guard usage of apply_subst_iterator with
4010 #ifdef GENERATOR_FILE.
4011 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
4012 (md_reader::read_mapping): Likewise.
4013 (add_define_attr_for_define_subst): Likewise.
4014 (add_define_subst_attr): Likewise.
4015 (read_subst_mapping): Likewise.
4016 (check_code_iterator): Likewise.
4017 (rtx_reader::read_rtx): Likewise. Move one-time initialization
4019 (one_time_initialization): New function.
4020 (rtx_reader::read_until): New method.
4021 (read_flags): New function.
4022 (parse_reg_note_name): New function.
4023 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
4024 Handle reuse_rtx ids.
4025 Wrap iterator lookup within #ifdef GENERATOR_FILE.
4026 Add parsing support for RTL dumps, mirroring the special-cases in
4027 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
4028 values, and calling handle_any_trailing_information.
4029 (rtx_reader::read_rtx_operand): Convert return type from void
4030 to rtx, returning return_rtx. Handle case 'e'. Call
4031 finalize_string on XSTR and XTMPL fields.
4032 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
4033 "(nil)" values were omitted. Call the postprocess vfunc on the
4035 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
4036 class ctor. Initialize m_in_call_function_usage. Call
4037 one_time_initialization.
4038 * rtl-tests.c (selftest::test_uncond_jump): Call
4039 set_new_first_and_last_insn.
4040 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
4041 * selftest-rtl.c: New file.
4042 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
4043 (selftest::get_insn_by_uid): New decl.
4044 * selftest-run-tests.c (selftest::run_tests): Call
4045 read_rtl_function_c_tests.
4046 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
4047 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
4050 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
4052 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
4053 operands in a special way. Assert that pos+len <= mode precision.
4055 2017-01-05 Jakub Jelinek <jakub@redhat.com>
4057 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
4058 3 argument Alias with unlimited for the negative form.
4059 (fno-vect-cost-model): Removed.
4061 2017-01-05 Martin Liska <mliska@suse.cz>
4063 * hsa-gen.c (gen_hsa_divmod): New function.
4064 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
4066 2017-01-05 Martin Liska <mliska@suse.cz>
4069 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
4072 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4074 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
4075 small constant length operands.
4077 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4079 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
4080 between loop iterations.
4082 2017-01-05 Martin Liska <mliska@suse.cz>
4085 * gimplify.c (gimplify_decl_expr): Compare to
4086 asan_poisoned_variables instread of checking flags.
4087 (gimplify_target_expr): Likewise.
4088 (gimplify_expr): Likewise.
4089 (gimplify_function_tree): Conditionally initialize
4090 asan_poisoned_variables.
4092 2017-01-04 Jeff Law <law@redhat.com>
4094 PR tree-optimizatin/78812
4095 * rtl.h (contains_mem_rtx_p): Prototype.
4096 * ifcvt.c (containts_mem_rtx_p): Move from here to...
4097 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
4098 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
4099 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
4100 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
4102 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4104 * input.c (assert_char_at_range): Default-initialize actual_range.
4106 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4108 * df-scan.c (df_ref_create_structure): Make regno unsigned,
4109 to match the caller.
4111 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4113 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
4114 insns after final jump in test to emit dummy move.
4116 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4118 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
4119 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
4121 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4123 * multiple_target.c (create_dispatcher_calls): Init e_next.
4124 * tree-ssa-loop-split.c (split_loop): Init border.
4125 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
4128 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
4133 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
4134 (altivec_register_operand): Do not return true if the operand
4135 contains a SUBREG mixing SImode and SFmode.
4136 (vsx_register_operand): Likewise.
4137 (vsx_reg_sfsubreg_ok): New predicate.
4138 (vfloat_operand): Do not return true if the operand contains a
4139 SUBREG mixing SImode and SFmode.
4140 (vint_operand): Likewise.
4141 (vlogical_operand): Likewise.
4142 (gpc_reg_operand): Likewise.
4143 (int_reg_operand): Likewise.
4144 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
4145 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
4146 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
4148 (rs6000_emit_move_si_sf_subreg): New helper function.
4149 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
4150 fixup SUBREGs involving SImode and SFmode.
4151 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
4152 numbers for the new peephole2 optimization.
4153 (peephole2 for SFmode unions): New peephole2 to optimize cases in
4154 the GLIBC math library that do AND/IOR/XOR operations on single
4155 precision floating point.
4156 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
4157 target macros to say whether we need to avoid SUBREGs mixing
4159 (TARGET_ALLOW_SF_SUBREG): Likewise.
4160 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
4161 (UNSPEC_SI_FROM_SF): Likewise.
4162 (iorxor): Change spacing.
4163 (and_ior_xor): New iterator for AND, IOR, and XOR.
4164 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
4165 (movdi_from_sf_zero_ext): Likewise.
4166 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
4167 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
4168 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
4169 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
4170 (fms<mode>4): Likewise.
4171 (fnma<mode>4): Likewise.
4172 (fnms<mode>4): Likewise.
4173 (nfma<mode>4): Likewise.
4174 (nfms<mode>4): Likewise.
4176 2017-01-04 Marek Polacek <polacek@redhat.com>
4179 * doc/invoke.texi: Document -Wpointer-compare.
4181 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4183 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
4186 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
4187 descriptions for -gdwarf-5 and emit them as uleb128 instead of
4190 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
4193 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
4194 documentation of the powerpc_popcntb_ok attribute.
4195 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
4196 code to issue warning messages if a requested CPU configuration is
4197 not supported by the binary (assembler and loader) toolchain.
4198 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
4199 made to define a built-in function that has been disabled.
4200 (paired_init_builtins): Add assertion to prevent ICE if attempt is
4201 made to define a built-in function that has been disabled.
4202 (altivec_init_builtins): Add comment explaining why definition
4203 of the DST built-in functions is not preceded by an assertion
4204 check. Add assertions to prevent ICE if attempts are made to
4205 define an altivec predicate or an abs* built-in function that has
4207 (htm_init_builtins): Add comment explaining why definition of the
4208 htm built-in functions is not preceded by an assertion check.
4210 2017-01-04 Jeff Law <law@redhat.com>
4212 PR tree-optimizatin/67955
4213 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
4214 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
4215 the points-to solution does not include pt_null. Use DECL_PT_UID
4218 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
4220 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
4221 Use gen_int_mode instead of gen_lopwart for const_int operands.
4223 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4225 PR tree-optimization/71563
4226 * match.pd: Simplify X << Y into X if Y is known to be 0 or
4227 out of range value - has low bits known to be zero.
4229 2017-01-04 Alan Modra <amodra@gmail.com>
4231 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
4232 * configure: Regenerate.
4233 * config.in: Regenerate.
4235 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4238 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
4239 a substring of the message, but strcmp with the whole message. Ifdef
4240 ENABLE_NLS, translate the message first using dgettext.
4242 2017-01-03 Jeff Law <law@redhat.com>
4244 PR tree-optimizatin/78856
4245 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
4246 (mark_threaded_blocks): Remove code to truncate thread paths that
4247 cross multiple loop headers. Instead invalidate the cached loop
4248 iteration information and handle case of a thread path walking
4249 into an irreducible region.
4251 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
4254 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
4255 assertions. Add support for doing the signbit if the IEEE 128-bit
4256 floating point value is in a GPR.
4257 * config/rs6000/rs6000.md (Fsignbit): Delete.
4258 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
4259 Update the length attribute if the value is in a GPR.
4260 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
4261 the sign or zero extension instruction, since the value is always 0/1.
4262 (signbit<mode>2_dm2): Delete using <Fsignbit>.
4265 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
4266 extracting SImode to a GPR register so that we can generate a
4267 store, limit the vector to be in a traditional Altivec register
4268 for the vextuwrx instruction.
4270 2017-01-03 Ian Lance Taylor <iant@google.com>
4272 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
4274 2017-01-03 Martin Sebor <msebor@redhat.com>
4276 PR tree-optimization/78696
4277 * gimple-ssa-sprintf.c (format_floating): Correct handling of
4278 precision. Use MPFR for %f for greater fidelity. Correct handling
4280 (pass_sprintf_length::compute_format_length): Set width and precision
4281 specified by asrerisk to void_node for vararg functions.
4282 (try_substitute_return_value): Adjust dump output.
4284 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
4286 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
4288 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
4290 * doc/invoke.texi (SPARC options): Document -mlra as the default.
4291 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
4292 -mlra/-mno-lra was passed to the compiler.
4294 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
4296 PR rtl-optimization/65618
4297 * emit-rtl.c (try_split): Move initialization of "before" and
4298 "after" to just before the call to emit_insn_after_setloc.
4300 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
4302 * doc/md.texi (Standard Names): Remove reference to Java frontend.
4304 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
4306 * dwarf2out.c (gen_enumeration_type_die): When
4307 -gno-strict-dwarf, add a DW_AT_encoding attribute.
4309 2017-01-03 Jakub Jelinek <jakub@redhat.com>
4311 PR tree-optimization/78965
4312 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
4313 Change first argument from const call_info & to call_info &. For %n
4314 set info.nowrite to false.
4317 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
4318 possibly throwing calls.
4320 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
4321 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
4322 and fns handling, rather than in a separate case SSA_NAME.
4324 2017-01-02 Jeff Law <law@redhat.com>
4326 * config/darwin-driver.c (darwin_driver_init): Const-correctness
4327 fixes for first_period and second_period variables.
4329 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
4332 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
4333 (*insvqi_1): New insn pattern.
4334 (*insvqi_1_mem_rex64): Ditto.
4336 (*insvqi_3): Rename from *insvqi.
4338 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
4340 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
4342 * doc/cfg.texi (Edges): Remove reference to Java.
4343 (Maintaining the CFG): Ditto.
4345 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4348 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
4349 transparent aliases.
4351 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4354 * predict.def (PRED_CALL): Update hitrate.
4355 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
4356 * predict.c (tree_estimate_probability_bb): Split CALL predictor
4357 into direct/indirect/polymorphic variants.
4359 2017-01-01 Jakub Jelinek <jakub@redhat.com>
4361 Update copyright years.
4363 * gcc.c (process_command): Update copyright notice dates.
4364 * gcov-dump.c (print_version): Ditto.
4365 * gcov.c (print_version): Ditto.
4366 * gcov-tool.c (print_version): Ditto.
4367 * gengtype.c (create_file): Ditto.
4368 * doc/cpp.texi: Bump @copying's copyright year.
4369 * doc/cppinternals.texi: Ditto.
4370 * doc/gcc.texi: Ditto.
4371 * doc/gccint.texi: Ditto.
4372 * doc/gcov.texi: Ditto.
4373 * doc/install.texi: Ditto.
4374 * doc/invoke.texi: Ditto.
4376 Copyright (C) 2017 Free Software Foundation, Inc.
4378 Copying and distribution of this file, with or without modification,
4379 are permitted in any medium without royalty provided the copyright
4380 notice and this notice are preserved.