| blob | 02b51576e2ddf8555b5bf452eebd530b837c5e72 |
1 2017-02-20 Marek Polacek <polacek@redhat.com>
3 PR middle-end/79537
4 * gimplify.c (gimplify_expr): Handle unused *&&L;.
6 PR sanitizer/79558
7 * ubsan.c (ubsan_type_descriptor): Check if TYPE_MAX_VALUE is null.
9 2017-02-20 Jakub Jelinek <jakub@redhat.com>
11 PR target/79568
12 * config/i386/i386.c (ix86_expand_builtin): Handle
13 OPTION_MASK_ISA_AVX512VL and OPTION_MASK_ISA_64BIT in
14 ix86_builtins_isa[fcode].isa as a requirement of those
15 flags and any other flag in the bitmask.
16 (ix86_init_mmx_sse_builtins): Use 0 instead of
17 ~OPTION_MASK_ISA_64BIT as mask.
18 * config/i386/i386-builtin.def (__builtin_ia32_rdtsc,
19 __builtin_ia32_rdtscp, __builtin_ia32_pause, __builtin_ia32_bsrsi,
20 __builtin_ia32_rdpmc, __builtin_ia32_rolqi, __builtin_ia32_rolhi,
21 __builtin_ia32_rorqi, __builtin_ia32_rorhi): Likewise.
23 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
25 PR target/78012
26 * lra-constraints.c (split_reg): Check requested split mode
27 is supported by the register.
29 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
31 * lra-constraints.c (simplify_operand_subreg): Remove early
32 return false.
34 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
36 PR target/78660
37 * lra-constraints.c (curr_insn_transform): Tighten condition
38 for converting SUBREG reloads from OP_OUT to OP_INOUT.
40 2017-02-20 Matthew Fortune <matthew.fortune@imgtec.com>
42 PR target/78660
43 * lra-constraints.c (curr_insn_transform): Handle
44 WORD_REGISTER_OPERATIONS requirements when reloading SUBREGs.
46 2017-02-19 Uros Bizjak <ubizjak@gmail.com>
48 Revert:
49 2016-05-30 Uros Bizjak <ubizjak@gmail.com>
51 * config/i386/sync.md (mfence_nosse): Use "lock orl $0, -4(%esp)".
53 2017-02-19 Jonathan Wakely <jwakely@redhat.com>
55 PR c++/69523
56 * doc/invoke.texi (C++ Dialect Options) [-Wliteral-suffix]: Update
57 description.
59 2017-02-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
61 * gimple-pretty-print.c (dump_ternary_rhs): Adjust gimple dump format
62 for FMA_EXPR.
64 2017-02-18 Jakub Jelinek <jakub@redhat.com>
66 * final.c (last_columnnum, override_columnnum): New variables.
67 (final_start_function): Set last_columnnum, pass it to begin_prologue
68 hook and pass 0 to dwarf2out_begin_prologue.
69 (final_scan_insn): Update override_columnnum. Pass last_columnnum
70 to source_line debug hook.
71 (notice_source_line): Compute last_columnnum and for debug_column_info
72 return true on column changes.
73 * debug.h (struct gcc_debug_hooks): Add column argument to
74 source_line and begin_prologue hooks.
75 (debug_nothing_int_charstar_int_bool): Remove prototype.
76 (debug_nothing_int_int_charstar,
77 debug_nothing_int_int_charstar_int_bool): New prototypes.
78 (dwarf2out_begin_prologue): Add column argument.
79 * debug.c (do_nothing_debug_hooks): Adjust source_line and
80 begin_prologue hooks.
81 (debug_nothing_int_charstar_int_bool): Remove.
82 (debug_nothing_int_int_charstar,
83 debug_nothing_int_int_charstar_int_bool): New functions.
84 * dwarf2out.c (dwarf2out_begin_prologue): Add column argument, pass it
85 through to dwarf2out_source_line.
86 (dwarf2_lineno_debug_hooks): Adjust begin_prologue hook.
87 (dwarf2out_source_line): Add column argument, emit it if requested.
88 * sdbout.c (sdbout_source_line, sdbout_begin_prologue): Add column
89 arguments.
90 * xcoffout.h (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
91 * xcoffout.c (xcoffout_begin_prologue, xcoffout_source_line): Likewise.
92 * vmsdbgout.c (vmsdbgout_begin_prologue): Add column argument, pass it
93 through to dwarf2out_begin_prologue.
94 (vmsdbgout_source_line): Add column argument, pass it through to
95 dwarf2out_source_line.
96 * dbxout.c (dbxout_begin_prologue): Add column argument, adjust
97 dbxout_source_line caller.
98 (dbxout_source_line): Add column argument.
100 * common.opt (gno-column-info, gcolumn-info): New options.
101 * dwarf2out.c (dwarf2_lineno_debug_hooks): Formatting fix.
102 (check_die): Also test for multiple DW_AT_decl_column attributes.
103 (add_src_coords_attributes, dwarf2out_imported_module_or_decl_1): Add
104 DW_AT_decl_column if requested.
105 (gen_subprogram_die): Compare and/or add also DW_AT_decl_column
106 if requested.
107 (gen_variable_die): Likewise.
108 (add_call_src_coords_attributes): Add DW_AT_call_column if requested.
109 * doc/invoke.texi (-gcolumn-info, -gno-column-info): Document.
111 PR target/79569
112 * config/i386/i386.opt (m3dnowa): Replace Undocumented with Report.
113 * common/config/i386/i386-common.c (OPTION_MASK_ISA_3DNOW_A_SET): Define.
114 (ix86_handle_option): Handle OPT_m3dnowa.
115 * doc/invoke.texi (-m3dnowa): Document.
116 * doc/extend.texi (__builtin_ia32_pmulhuw, __builtin_ia32_pf2iw): Use
117 -m3dnowa instead of -m3dnow -march=athlon.
119 PR target/79559
120 * config/i386/i386.c (ix86_print_operand): Use output_operand_lossage
121 instead of gcc_assert for K, r and R code checks. Formatting fixes.
123 2017-02-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
125 PR target/79261
126 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
127 support for CODE_FOR_vsx_xxpermdi_v2d[fi]_be.
128 * config/rs6000/rs6000.md (reload_gpr_from_vsx<mode>): Call
129 generator for vsx_xxpermdi_<mode>_be.
130 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Remove logic to
131 force big-endian semantics.
132 (vsx_xxpermdi_<mode>_be): New define_expand with same
133 implementation as previous version of vsx_xxpermdi_<mode>.
135 2017-02-17 Jakub Jelinek <jakub@redhat.com>
137 PR tree-optimization/79327
138 * gimple-ssa-sprintf.c (format_integer): Remove likely_adjust
139 variable, its initialization and use.
141 2017-02-17 Julia Koval <julia.koval@intel.com>
143 * common/config/i386/i386-common.c (OPTION_MASK_ISA_RDPID_SET): New.
144 (OPTION_MASK_ISA_PKU_UNSET): New.
145 (ix86_handle_option): Handle -mrdpid.
146 * config/i386/cpuid.h (bit_RDPID): New.
147 * config/i386/driver-i386.c (host_detect_local_cpu):
148 Detect RDPID feature.
149 * config/i386/i386-builtin.def (__builtin_ia32_rdpid): New.
150 * config/i386/i386-c.c (ix86_target_macros_internal):
151 Handle RDPID flag.
152 * config/i386/i386.c (ix86_target_string): Add -mrdpid to isa2_opts.
153 (ix86_valid_target_attribute_inner_p): Add "rdpid".
154 (ix86_expand_builtin): Handle IX86_BUILTIN_RDPID.
155 * config/i386/i386.h (TARGET_RDPID, TARGET_RDPID_P): New.
156 * config/i386/i386.md (define_insn "rdpid"): New.
157 * config/i386/i386.opt Add -mrdpid.
158 * config/i386/immintrin.h (_rdpid_u32): New.
160 2017-02-17 Vladimir Makarov <vmakarov@redhat.com>
162 PR rtl-optimization/79541
163 * lra-constraints.c (curr_insn_transform): Remove wrong asm insn
164 instead of transforming it into USE.
166 2017-02-17 Segher Boessenkool <segher@kernel.crashing.org>
168 * config/rs6000/rs6000.md (extendsfdf2): Remove default arguments.
169 If HONOR_SNANS (SFmode) force the input to a register.
170 (*extendsfdf2_fpr): Add !HONOR_SNANS (SFmode) condition.
171 (*extendsfdf2_snan): New pattern, used when using SNaNs; it generates
172 an frsp or similar insn.
174 2017-02-17 Martin Liska <mliska@suse.cz>
176 PR rtl-optimization/79577
177 * params.def (selsched-max-sched-times): Increase minimum to 1.
179 2017-02-17 Martin Liska <mliska@suse.cz>
181 PR rtl-optimization/79574
182 * gcse.c (want_to_gcse_p): Prevent integer overflow.
184 2017-02-17 Martin Liska <mliska@suse.cz>
186 PR tree-optimization/79529
187 * tree-ssa-loop-unswitch.c (is_maybe_undefined): Use
188 ssa_defined_default_def_p to handle cases which are implicitly
189 defined.
190 * tree-ssa.c (ssa_defined_default_def_p): New function.
191 (ssa_undefined_value_p): Use ssa_defined_default_def_p to handle cases
192 which are implicitly defined.
193 * tree-ssa.h (ssa_defined_default_def_p): Declare.
195 2017-02-17 Richard Biener <rguenther@suse.de>
197 PR middle-end/79576
198 * params.def (max-ssa-name-query-depth): Limit to 10.
200 2017-02-17 Richard Biener <rguenther@suse.de>
202 PR tree-optimization/79552
203 * tree-ssa-structalias.c (visit_loadstore): Properly verify
204 default defs.
206 2017-02-17 Richard Biener <rguenther@suse.de>
208 PR bootstrap/79567
209 * genmatch.c (output_line_directive): Handle DIR_SEPARATOR_2.
211 2017-02-17 Marek Polacek <polacek@redhat.com>
213 PR middle-end/79536
214 * fold-const.c (fold_negate_expr_1): Renamed from fold_negate_expr.
215 (fold_negate_expr): New wrapper.
217 2017-02-16 Sandra Loosemore <sandra@codesourcery.com>
219 * doc/invoke.texi (C++ Dialect Options) [-Wno-non-template-friend]:
220 Correct terminology and de-emphasize pre-standard behavior.
222 2017-02-16 Alan Modra <amodra@gmail.com>
224 PR rtl-optimization/79286
225 * ira.c (def_dominates_uses): New function.
226 (update_equiv_regs): Don't create an equivalence for insns that
227 may trap where the register def does not dominate the use.
229 2017-02-16 Vladimir Makarov <vmakarov@redhat.com>
231 PR rtl-optimization/78127
232 * lra.c (lra): Call lra_eliminate before finish the loop after
233 lra_constraint.
235 2017-02-16 Richard Biener <rguenther@suse.de>
237 * graphite.h: Do not include isl/isl_val_gmp.h, instead include
238 isl/isl_val.h.
239 * graphite-isl-ast-to-gimple.c (gmp_cst_to_tree): Remove.
240 (gcc_expression_from_isl_expr_int): Use generic isl_val interface.
241 * graphite-sese-to-poly.c: Do not include isl/isl_val_gmp.h.
242 (isl_val_int_from_wi): New function.
243 (extract_affine_gmp): Rename to ...
244 (extract_affine_wi): ... this, take a widest_int.
245 (extract_affine_int): Just wrap extract_affine_wi.
246 (add_param_constraints): Use isl_val_int_from_wi.
247 (add_loop_constraints): Likewise, and extract_affine_wi.
249 2017-02-15 Jeff Law <law@redhat.com>
251 PR middle-end/79521
252 * ira-costs.c (scan_one_insn): Check have_regs_of_mode before calling
253 ira_init_register_move_cost_if_necessary.
255 2017-02-15 Martin Sebor <msebor@redhat.com>
257 PR middle-end/32003
258 * doc/invoke.texi (-fdump-final-insns): Replace option accidentally
259 removed in a prior commit.
261 2017-02-15 Bin Cheng <bin.cheng@arm.com>
263 PR tree-optimization/79347
264 * tree-vect-loop-manip.c (vect_do_peeling): Maintain profile
265 counters during peeling.
267 2017-02-15 Thomas Schwinge <thomas@codesourcery.com>
269 * Makefile.in (site.exp): Remove "set ISLVER".
271 2017-02-15 Jakub Jelinek <jakub@redhat.com>
273 PR target/79487
274 * real.c (real_from_integer): Call real_convert even for decimal.
276 2017-02-15 Dominik Vogt <vogt@linux.vnet.ibm.com>
278 PR target/79421
279 * config/s390/s390.c: define TARGET_CUSTOM_FUNCTION_DESCRIPTORS.
281 2017-02-14 Andrew Pinski <apinski@cavium.com>
283 * config/aarch64/aarch64-cores.def (thunderx2t99): Move to under 'C"
284 cores and change the partno/implementer to be correct.
285 (thunderx2t99p1): New core which replaces thunderx2t99 and still has
286 the 'B" as the implementer.
287 * config/aarch64/aarch64-tune.md: Regenerate.
289 2017-02-14 Carl Love <cel@us.ibm.com>
291 * config/rs6000/rs6000.c: Add case statement entry to make the
292 xvcvuxdsp built-in argument unsigned.
293 * config/rs6000/vsx.md: Fix the source and return operand types so they
294 match the instruction definitions from the ISA document. Fix typo
295 in the instruction generation for the (define_insn "vsx_xvcvuxdsp"
296 statement.
298 2017-02-14 Vladimir Makarov <vmakarov@redhat.com>
300 PR target/79282
301 * lra-int.h (struct lra_operand_data, struct lra_insn_reg): Add
302 member early_clobber_alts.
303 * lra-lives.c (reg_early_clobber_p): New.
304 (process_bb_lives): Use it.
305 * lra.c (new_insn_reg): New arg early_clobber_alts. Use it.
306 (debug_operand_data): Initialize early_clobber_alts.
307 (setup_operand_alternative): Set up early_clobber_alts.
308 (collect_non_operand_hard_regs): Ditto. Pass early clobber
309 alternatives to new_insn_reg.
310 (add_regs_to_insn_regno_info): Add arg early_clobber_alts. Use
311 it.
312 (lra_update_insn_regno_info): Pass the new arg.
314 2017-02-14 Jakub Jelinek <jakub@redhat.com>
316 PR middle-end/79505
317 * omp-offload.c (free_oacc_loop): Release loop->ifns vector.
318 (new_oacc_loop_raw): Don't clear already cleared fields.
320 PR target/79481
321 * config/i386/avx512pfintrin.h (_mm512_prefetch_i32gather_pd,
322 _mm512_prefetch_i32gather_ps, _mm512_prefetch_i64gather_pd,
323 _mm512_prefetch_i64gather_ps): New inline functions and macros.
325 2017-02-14 Uros Bizjak <ubizjak@gmail.com>
327 PR target/79495
328 * config/i386/i386.md (*movxf_internal): Add (o,rC) alternative.
330 2017-02-14 H.J. Lu <hongjiu.lu@intel.com>
332 PR target/79498
333 * config/i386/i386.c (timode_scalar_chain::convert_insn): Insert
334 the extra instruction to the right place to store 128-bit constant
335 when needed.
337 2017-02-14 Martin Sebor <msebor@redhat.com>
339 PR middle-end/79448
340 * gimple-ssa-sprintf.c (format_directive): Avoid issuing INT_MAX
341 warning for strings of unknown length.
343 2017-02-13 Segher Boessenkool <segher@kernel.crashing.org>
345 * config.gcc (supported_defaults) [powerpc*-*-*]: Update.
347 2017-02-14 Jeff Law <law@redhat.com>
349 PR target/79404
350 * ira-costs.c (scan_one_insn): Initialize register move costs
351 for pseudos seen in USE/CLOBBER insns.
353 PR tree-optimization/79095
354 * tree-vrp.c (extract_range_from_binary_expr_1): For EXACT_DIV_EXPR,
355 if the numerator has the range ~[0,0] make the resultant range ~[0,0].
356 (extract_range_from_binary_expr): For MINUS_EXPR with no derived range,
357 if the operands are known to be not equal, then the resulting range
358 is ~[0,0].
359 (intersect_ranges): If the new range is ~[0,0] and the old range is
360 wide, then prefer ~[0,0].
361 * tree-vrp.c (overflow_comparison_p_1): New function.
362 (overflow_comparison_p): New function.
363 * tree-vrp.c (register_edge_assert_for_2): Register additional asserts
364 if NAME is used in an overflow test.
365 (vrp_evaluate_conditional_warnv_with_ops): If the ops represent an
366 overflow check that can be expressed as an equality test, then adjust
367 ops to be that equality test.
369 2017-02-14 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
371 * config/s390/s390-builtin-types.def: Remove flags argument.
372 * config/s390/s390.c (s390_init_builtins): Likewise.
374 2017-02-14 Martin Liska <mliska@suse.cz>
376 * tree-ssa-loop-unswitch.c (hoist_guard): Release get_loop_body
377 vector. Fix trailing white spaces.
379 2017-02-14 James Greenhalgh <james.greenhalgh@arm.com>
381 * config/aarch64/aarch64.c (aarch64_simd_container_mode): Handle
382 HFmode.
384 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
386 PR rtl-optimization/68664
387 * config/arm/arm.c (arm_sched_can_speculate_insn):
388 New function. Declare prototype.
389 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
391 2017-02-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
393 PR rtl-optimization/68664
394 * config/aarch64/aarch64.c (aarch64_sched_can_speculate_insn):
395 New function.
396 (TARGET_SCHED_CAN_SPECULATE_INSN): Define.
398 2017-02-14 Amit Pawar <amit.pawar@amd.com>
400 * config/i386/i386.c (znver1_cost): Fix the alignment for function and
401 max skip bytes for function, loop and jump.
403 2017-02-14 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
405 * gimple-pretty-print.c (dump_unary_rhs): Adjust dump format for
406 ABS_EXPR for gimple dump.
408 2017-02-14 Jakub Jelinek <jakub@redhat.com>
410 PR target/79462
411 * config/sh/sh.c (expand_cbranchdi4): Don't clear operands[4].
413 PR tree-optimization/79408
414 * tree-vrp.c (simplify_div_or_mod_using_ranges): Handle also the
415 case when on TRUNC_MOD_EXPR op0 is INTEGER_CST.
416 (simplify_stmt_using_ranges): Call simplify_div_or_mod_using_ranges
417 also if rhs1 is INTEGER_CST.
419 2017-02-14 Richard Biener <rguenther@suse.de>
421 PR middle-end/79432
422 * tree-into-ssa.c (insert_phi_nodes): When the function can
423 have abnormal edges rewrite SSA names with broken use-def
424 dominance out of SSA and register them for PHI insertion.
426 2017-02-13 Martin Sebor <msebor@redhat.com>
428 PR middle-end/79496
429 * gimple-ssa-sprintf.c (pass_sprintf_length::handle_gimple_call): Avoid
430 clearing info.nowrite flag when snprintf size argument is a range.
432 2017-02-13 Jakub Jelinek <jakub@redhat.com>
434 * cprop.c (cprop_jump): Add missing space in string literal.
435 * tree-ssa-structalias.c (rewrite_constraints): Likewise.
436 (get_constraint_for_component_ref): Likewise.
437 * df-core.c (df_worklist_dataflow_doublequeue): Likewise.
438 * tree-outof-ssa.c (insert_partition_copy_on_edge): Likewise.
439 * lra-constraints.c (process_alt_operands): Likewise.
440 * ipa-inline.c (inline_small_functions): Likewise.
441 * tree-ssa-sccvn.c (visit_reference_op_store): Likewise.
442 * cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Likewise.
443 * trans-mem.c (diagnose_tm_1_op): Likewise.
444 * omp-grid.c (grid_find_single_omp_among_assignments): Likewise.
445 (grid_parallel_clauses_gridifiable): Likewise.
447 * config/nvptx/mkoffload.c (process): Add space in between
448 , and %d.
450 * config/i386/i386.h (REG_CLASS_NAMES): Add , in between
451 "MOD4_SSE_REGS" and "ALL_REGS".
453 * spellcheck.c (test_data): Add , in between "foo" and "food".
455 2017-02-13 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
457 PR target/79449
458 * config/rs6000/rs6000.c (expand_block_compare): Make sure runtime
459 boundary crossing check and subsequent code generation agree.
461 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
463 * config/aarch64/aarch64.c (has_memory_op): Delete.
464 (aarch64_madd_needs_nop): Use contains_mem_rtx_p instead of
465 has_memory_op.
467 2017-02-13 Jakub Jelinek <jakub@redhat.com>
469 PR rtl-optimization/79388
470 PR rtl-optimization/79450
471 * combine.c (distribute_notes): When removing TEM_INSN for which
472 corresponding dest has last value recorded, invalidate that last
473 value.
475 2017-02-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
477 * config/arm/arm.c (arm_print_tune_info): Use ASM_COMMENT_START instead
478 of explicit '@'. Add missing assembly comment marker on branch costs
479 printout.
481 2017-02-13 Nathan Sidwell <nathan@acm.org>
483 * gengtype-lex.l (<in_struct>): Add '/'.
485 2017-02-13 Martin Liska <mliska@suse.cz>
487 PR c/79471
488 * calls.c (expand_call): Replace XALLOCAVEC with XCNEWVEC.
490 2017-02-13 Richard Biener <rguenther@suse.de>
492 * configure.ac (HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS):
493 Remove.
494 * configure: Re-generate.
495 * config.in: Likewise.
496 * graphite-dependences.c: Simplify as if
497 HAVE_ISL_OPTIONS_SET_SCHEDULE_SERIALIZE_SCCS was defined.
498 * graphite-isl-ast-to-gimple.c: Likewise.
499 * graphite-optimize-isl.c: Likewise.
500 * graphite-poly.c: Likewise.
501 * graphite-sese-to-poly.c: Likewise.
502 * graphite.h: Likewise.
503 * toplev.c: Include isl/version.h and use isl_version () for
504 printing the ISL version.
505 * doc/install.texi: Update ISL requirement.
507 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
509 * doc/standards.texi (Standards): Update reference to
510 Objective-C 2.0.
512 2017-02-12 Gerald Pfeifer <gerald@pfeifer.com>
514 * doc/extend.texi (Named Address Spaces): sourceware.org now
515 defaults to https.
516 * doc/install.texi (Binaries): Ditto.
517 (Specific): Ditto.
519 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
521 * doc/cpp.texi: Replace "stringify"/"stringification" with C
522 standard terminology "stringize"/"stringizing" throughout.
523 * doc/cppinternals.texi: Likewise.
525 2017-02-11 Sandra Loosemore <sandra@codesourcery.com>
527 * doc/extend.texi: Fix some spelling mistakes and typos.
528 * doc/invoke.texi: Likewise.
530 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
532 PR ipa/79224
533 * params.def (inline-min-speedup) Change from 10 to 8.
535 2017-02-11 Jakub Jelinek <jakub@redhat.com>
537 * doc/invoke.texi (fopenmp): Bump OpenMP version from 4.0 to
538 4.5.
540 2017-02-11 Jan Hubicka <hubicka@ucw.cz>
542 PR ipa/79224
543 * ipa-inline-analysis.c (get_minimal_bb): New function.
544 (record_modified): Use it.
545 (remap_edge_change_prob): Handle also ancestor functions.
547 2017-02-11 Gerald Pfeifer <gerald@pfeifer.com>
549 * doc/contrib.texi (Contributors): Remove broken link into
550 the Mauve CVS repository.
552 2017-02-11 Jakub Jelinek <jakub@redhat.com>
554 PR middle-end/79454
555 * internal-fn.c (expand_vector_ubsan_overflow): Use piece-wise
556 result computation whenever lhs doesn't have vector mode, not
557 just when it has BLKmode.
559 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
561 * doc/makefile.texi (profiledbootstrap): Refer to the
562 installation instructions only in textual form.
564 2017-02-10 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
566 PR target/79295
567 * config/rs6000/altivec.md (bcd<bcd_add_sub>): Fix constraints.
569 2017-02-10 Gerald Pfeifer <gerald@pfeifer.com>
571 * doc/install.texi (Specific): Use https for blackfin.uclinux.org.
572 (Specific): Update mingw-w64 reference.
573 (Binaries): Ditto.
574 (Specific): Remove broken link to Renesas RX processor.
576 2017-02-10 Richard Biener <rguenther@suse.de>
578 * toplev.c (process_options): Do not mention obsolete graphite
579 options when printing sorry message about missing graphite support.
580 Mention -floop-nest-optimize.
582 2017-02-10 Christophe Lyon <christophe.lyon@linaro.org>
584 * config/aarch64/arm_neon.h (vtst_p8): Rewrite without asm.
585 (vtst_p16): Likewise.
586 (vtstq_p8): Likewise.
587 (vtstq_p16): Likewise.
588 (vtst_p64): New.
589 (vtstq_p64): Likewise.
590 * config/arm/arm_neon.h (vgetq_lane_p64): New.
591 (vset_lane_p64): New.
592 (vsetq_lane_p64): New.
594 2017-02-10 Jakub Jelinek <jakub@redhat.com>
596 PR tree-optimization/79411
597 * tree-ssa-reassoc.c (is_reassociable_op): Return false if
598 stmt operands are SSA_NAMEs used in abnormal phis.
599 (can_reassociate_p): Return false if op is SSA_NAME used in abnormal
600 phis.
602 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
604 PR ipa/70795
605 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
606 flag if needed.
608 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
610 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
612 2017-02-09 Jakub Jelinek <jakub@redhat.com>
614 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
615 to avoid warning.
617 PR c/79413
618 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
619 not arbitrary TREE_CONSTANT.
621 PR c/79431
622 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
623 "omp declare target link" attribute unless is_global_var.
624 * omp-offload.c (find_link_var_op): Likewise.
626 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
627 Chung-Lin Tang <cltang@codesourcery.com>
629 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
630 OMP_CLAUSE_TILE.
631 (gimplify_adjust_omp_clauses): Don't delete TILE.
632 (gimplify_omp_for): Deal with TILE.
633 * internal-fn.c (expand_GOACC_TILE): New function.
634 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
635 (GOACC_TILE): New.
636 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
637 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
638 element fields.
639 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
640 avoid DIV for outermost collapse var.
641 (expand_oacc_for): Insert tile element loop as needed. Adjust.
642 Remove out of date comments, fix whitespace.
643 * omp-general.c (omp_extract_for_data): Deal with tiling.
644 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
645 adjust OLF_DIM_BASE value.
646 (struct omp_for_data): Add tiling field.
647 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
648 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
649 for auto loops. Remove default auto determining, moved to
650 oacc_loop_fixed_partitions.
651 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
652 stmts, add e_mask field.
653 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
654 (oacc_thread_numbers): Use oacc_dim_call.
655 (oacc_xform_tile): New.
656 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
657 (finish_oacc_loop): Adjust for ifns vector.
658 (oacc_loop_discover_walk): Append loop abstraction sites to list,
659 add case for GOACC_TILE fns.
660 (oacc_loop_xform_loop): Delete.
661 (oacc_loop_process): Iterate over call list directly, and add
662 handling for GOACC_TILE fns.
663 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
664 dump partitioning.
665 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
666 vector partitioning to outer loops. Assign 2 partitions to loops
667 when available. Add TILE handling.
668 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
669 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
670 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
671 * tree.c (omp_clause_num_ops): Adjust TILE ops.
672 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
674 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
676 * configure.ac (ACX_BUGURL): Update.
677 * configure: Regenerate.
679 2017-02-09 Richard Biener <rguenther@suse.de>
681 PR tree-optimization/69823
682 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
683 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
685 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
687 * config/arc/arc-c.def: Add __NPS400__ definition.
688 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
689 (TARGET_NPS400): Define.
691 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
693 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
694 file.
695 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
696 pointer, arch_info.
697 (arc_cpu_types): Fill the arch_info field with a pointer into the
698 arc_arch_types table.
699 (arc_selected_cpu): Declare.
700 * config/arc/arc.c (arc_selected_cpu): Make global.
701 (arc_selected_arch): Delete.
702 (arc_base_cpu): Delete.
703 (arc_override_options): Remove references to deleted variables,
704 update access to arch information.
705 (ARC_OPT): Update access to arch information.
706 (ARC_OPTX): Likewise.
707 * config/arc/arc.h (arc_base_cpu): Remove declaration.
708 (TARGET_ARC600): Update access to arch information.
709 (TARGET_ARC601): Likewise.
710 (TARGET_ARC700): Likewise.
711 (TARGET_EM): Likewise.
712 (TARGET_HS): Likewise.
713 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
714 information.
716 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
718 PR target/78604
719 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
720 condition/operands for integer GE/LE/GEU/LEU operations.
722 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
724 PR translation/79397
725 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
726 of AltiVec.
728 2017-02-08 Martin Jambor <mjambor@suse.cz>
730 PR ipa/79375
731 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
732 whether allocation happened.
733 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
734 nothing was allocated.
736 2017-02-08 Jakub Jelinek <jakub@redhat.com>
738 PR tree-optimization/79408
739 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
740 constant, but SSA_NAME with a known integer range, use the minimum
741 of that range instead of op1 to determine if modulo can be replaced
742 with its first operand.
744 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
746 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
748 2017-02-08 Richard Biener <rguenther@suse.de>
750 PR tree-optimization/71824
751 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
752 Check all loops contained in the merged region.
754 2017-02-07 Andrew Pinski <apinski@cavium.com>
756 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
758 2017-02-07 Andrew Pinski <apinski@cavium.com>
760 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
761 (thunderxt88): Likewise.
762 (thunderxt81): Disable LSE and change v8.1 to v8.
763 (thunderxt83): Likewise.
765 2017-02-07 Jakub Jelinek <jakub@redhat.com>
766 Richard Biener <rguenther@suse.de>
768 PR middle-end/79399
769 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
770 type from int to size_t.
771 * ira-costs.c (struct_costs_size): Change type from int to size_t.
773 2017-02-07 Jakub Jelinek <jakub@redhat.com>
775 PR rtl-optimization/79386
776 * cprop.c (bypass_conditional_jumps): Initialize
777 bypass_last_basic_block already before splitting bbs after
778 unconditional traps...
779 (bypass_conditional_jumps): ... rather than here.
781 PR target/79299
782 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
783 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
784 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
785 fix -masm=intel patterns.
787 2017-02-07 Richard Biener <rguenther@suse.de>
789 PR tree-optimization/79256
790 PR middle-end/79278
791 * builtins.c (get_object_alignment_2): Use min_align_of_type
792 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
793 and ADJUST_FIELD_ALIGN.
795 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
796 type parameter.
797 * doc/tm.texi: Regenerate.
798 * stor-layout.c (layout_decl): Adjust.
799 (update_alignment_for_field): Likewise.
800 (place_field): Likewise.
801 (min_align_of_type): Likewise.
802 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
803 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
804 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
805 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
806 * config/frv/frv.c (frv_adjust_field_align): Likewise.
807 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
808 * config/i386/i386.c (x86_field_alignment): Likewise.
809 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
810 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
811 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
812 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
813 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
814 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
815 Likewise.
817 Revert
818 2017-01-30 Richard Biener <rguenther@suse.de>
820 PR tree-optimization/79256
821 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
822 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
823 alignment on TYPE.
825 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
827 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
828 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
829 builtins to SImode and emit a zero-extend, if necessary.
831 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
833 * docs/invoke.texi (RISC-V Options): Alphabetize.
835 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
837 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
838 options.
840 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
842 * config/riscv/riscv.c: New file.
843 * gcc/common/config/riscv/riscv-common.c: Likewise.
844 * config.gcc: Likewise.
845 * config/riscv/constraints.md: Likewise.
846 * config/riscv/elf.h: Likewise.
847 * config/riscv/generic.md: Likewise.
848 * config/riscv/linux.h: Likewise.
849 * config/riscv/multilib-generator: Likewise.
850 * config/riscv/peephole.md: Likewise.
851 * config/riscv/pic.md: Likewise.
852 * config/riscv/predicates.md: Likewise.
853 * config/riscv/riscv-builtins.c: Likewise.
854 * config/riscv/riscv-c.c: Likewise.
855 * config/riscv/riscv-ftypes.def: Likewise.
856 * config/riscv/riscv-modes.def: Likewise.
857 * config/riscv/riscv-opts.h: Likewise.
858 * config/riscv/riscv-protos.h: Likewise.
859 * config/riscv/riscv.h: Likewise.
860 * config/riscv/riscv.md: Likewise.
861 * config/riscv/riscv.opt: Likewise.
862 * config/riscv/sync.md: Likewise.
863 * config/riscv/t-elf-multilib: Likewise.
864 * config/riscv/t-linux: Likewise.
865 * config/riscv/t-linux-multilib: Likewise.
866 * config/riscv/t-riscv: Likewise.
867 * configure.ac: Likewise.
868 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
869 Waterman as RISC-V maintainers.
870 * doc/install.texi: Add RISC-V entries.
871 * doc/invoke.texi: Add RISC-V options section.
872 * doc/md.texi: Add RISC-V constraints section.
873 * configure: Regenerated.
875 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
877 PR target/66144
878 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
879 false values to be constant vectors with all 0 or all 1 bits set.
880 (vcondu<mode><mode>): Likewise.
881 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
882 predicate.
883 (fpmask_comparison_operator): Update comment.
884 (vecint_comparison_operator): New predicate.
885 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
886 vector conditionals when the true and false values are constant
887 vectors with all 0 bits or all 1 bits set.
889 2017-02-06 Martin Sebor <msebor@redhat.com>
891 PR tree-optimization/79376
892 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
894 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
896 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
897 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
898 to simplify split condition.
900 2017-02-06 Jakub Jelinek <jakub@redhat.com>
902 * omp-expand.c (oxpand_omp_atomic_fetch_op,
903 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
904 false.
906 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
908 PR rtl-optimization/68664
909 * target.def (can_speculate_insn): New hook.
910 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
911 * doc/tm.texi: Regenerate.
912 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
913 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
914 (rs6000_sched_can_speculate_insn): New function.
916 2017-02-06 Jakub Jelinek <jakub@redhat.com>
918 PR tree-optimization/79284
919 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
920 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
921 vectorizable_mask_load_store, vectorizable_operation,
922 vect_is_simple_cond, get_same_sized_vectype): Use it instead
923 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
924 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
925 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
926 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
927 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
928 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
929 is_gimple_assign (stmt). Replace another such test with
930 is_gimple_assign (stmt).
932 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
934 PR target/78883
935 * config/avr/avr.c (rtl-iter.h): Include it.
936 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
937 (avr_legitimate_combined_insn): ...and implementation.
939 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
941 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
942 * config/s390/s390.c (s390_const_operand_ok)
943 (s390_canonicalize_comparison, s390_extract_part)
944 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
945 (s390_contiguous_bitmask_p, s390_rtx_costs)
946 (legitimize_pic_address): Likewise.
947 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
948 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
949 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
950 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
951 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
953 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
955 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
956 REGNO($0) == REGNO($1).
958 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
960 * config/s390/linux.h(SIZE_TYPE): Add comment.
962 2017-02-06 Julian Brown <julian@codesourcery.com>
963 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
964 Virendra Pathak <virendra.pathak@broadcom.com>
966 * config/aarch64/aarch64-cores.def: Change the scheduler
967 to Thunderx2t99.
968 * config/aarch64/aarch64.md: Include thunderx2t99.md.
969 * config/aarch64/thunderx2t99.md: New file.
971 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
973 * doc/standards.texi (Go Language): Update link to language
974 standard.
976 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
978 * tree-eh.c (lower_resx): Sanitize profile.
979 (cleanup_empty_eh_move_lp): Likewise.
981 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
983 PR tree-ssa/79347
984 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
985 ELSE_PROB.
986 * cfgloopmanip.h (loop_version): Update prototype.
987 * modulo-sched.c (sms_schedule): Update call of loop_version.
988 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
989 * tree-parloops.c (gen_parallel_loop): Likewise.
990 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
991 * tree-ssa-loop-split.c (split_loop): Likewise.
992 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
993 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
995 2017-02-05 Martin Liska <mliska@suse.cz>
997 PR bootstrap/78985
998 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
999 variable to NULL.
1000 (print_operand_address): Initialize a struct to zero.
1002 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
1004 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
1005 garbage collector only in textual form.
1007 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
1009 * doc/extend.texi (x86 specific memory model extensions for
1010 transactional memory): Simplify a phrase.
1012 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
1014 PR target/79353
1015 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
1016 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
1017 (atomic_storedi_1): Likewise.
1019 2017-02-04 Jakub Jelinek <jakub@redhat.com>
1021 PR tree-optimization/79338
1022 * tree-parloops.c (gather_scalar_reductions): Don't call
1023 vect_analyze_loop_form for loop->inner before destroying loop's
1024 loop_vinfo.
1026 2017-02-03 Martin Sebor <msebor@redhat.com>
1028 PR tree-optimization/79327
1029 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
1030 when precision has resulted in leading zeros.
1031 (format_integer): Adjust the likely counter to assume an unknown
1032 argument that may be zero is non-zero.
1034 2017-02-03 Jason Merrill <jason@redhat.com>
1036 PR c++/78689
1037 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
1038 avoid copying non-taken branch.
1040 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1042 PR tree-optimization/79340
1043 * tree-vect-loop.c (vectorizable_reduction): Release
1044 vec_defs elements after safe_splicing them into other vectors.
1045 Formatting fixes.
1047 PR tree-optimization/79327
1048 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
1049 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
1050 dirtype.
1051 (format_integer): Use wide_int_to_tree instead of build_int_cst
1052 + to_?hwi. If argmin is NULL, just set argmin and argmax to
1053 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
1054 of shortest and longest sequence.
1056 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
1058 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
1059 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
1061 2017-02-03 Walter Lee <walt@tilera.com>
1063 PR target/78862
1064 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
1065 after initial stackframe link reg save.
1066 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
1068 2017-02-03 Jakub Jelinek <jakub@redhat.com>
1070 PR target/79354
1071 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
1072 wu for stxssp alternative.
1074 2017-02-03 Martin Sebor <msebor@redhat.com>
1076 PR tree-optimization/79352
1077 * gimple-fold.c (get_range_strlen): Add argument.
1078 (get_range_strlen): Change return type to bool.
1079 (get_maxval_strlen): Pass in a dummy argument.
1080 * gimple-fold.h (get_range_strlen): Change return type to bool.
1081 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
1082 * tree.h (array_at_struct_end_p): Add argument.
1083 * tree.c (array_at_struct_end_p): Handle it.
1085 2017-02-03 Martin Liska <mliska@suse.cz>
1087 PR lto/66295
1088 * multiple_target.c (create_dispatcher_calls): Redirect edge
1089 from a caller of a dispatcher.
1090 (expand_target_clones): Make the clones local.
1091 (ipa_target_clone): Do both target clones and resolvers.
1092 (ipa_dispatcher_calls): Remove the pass.
1093 (pass_dispatcher_calls::gate): Likewise.
1094 (make_pass_dispatcher_calls): Likewise.
1095 * passes.def (pass_target_clone): Put as very first IPA early
1096 pass.
1098 2017-02-03 Martin Liska <mliska@suse.cz>
1100 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
1101 in case of a function with ifunc attribute.
1103 2017-02-03 Martin Liska <mliska@suse.cz>
1105 * cgraph.c (cgraph_node::dump): Dump function version info.
1106 * symtab.c (symtab_node::dump_base): Add missing new line.
1108 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
1110 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
1111 (ifcombine_ifandif): Use it.
1113 2017-02-03 Martin Liska <mliska@suse.cz>
1115 * doc/invoke.texi: Document default value for
1116 use-after-scope-direct-emission-threshold.
1118 2017-02-03 Martin Liska <mliska@suse.cz>
1120 PR tree-optimization/79339
1121 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
1122 (format_floating): Likewise.
1124 2017-02-03 Martin Liska <mliska@suse.cz>
1126 PR ipa/79337
1127 * ipa-prop.c (ipa_node_params_t::insert): Remove current
1128 implementation.
1129 (ipa_node_params_t::remove): Likewise.
1130 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
1131 initialization from removed ipa_node_params_t::insert.
1132 (ipa_node_params::~ipa_node_params): Move from removed
1133 ipa_node_params_t::release.
1134 * symbol-summary.h (symbol_summary::m_released): New member.
1135 Do not release a summary twice. Do not allow to call finalizer
1136 for types of a summary that live in GGC memory.
1138 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
1140 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
1141 cmp_branch fusion.
1143 2017-02-02 Martin Sebor <msebor@redhat.com>
1145 PR middle-end/79275
1146 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
1147 (format_string): Tighten up the range of output for non-constant
1148 strings and correct the expected range for wide non-constant strings.
1150 2017-02-02 Martin Sebor <msebor@redhat.com>
1152 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
1154 PR middle-end/32003
1155 * doc/invoke.texi (-fdump-tree-): Remove pass-specific options from
1156 index.
1157 (-fdump-tree-@var): Add to index and document how to come up
1158 with pass-specific option and dump file names.
1159 (-fdump-passes): Clarify where to look for output.
1161 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
1163 PR middle-end/77445
1164 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
1165 statistics of the analyzed path; allow threading for speed when
1166 any of BBs along the path are optimized for speed.
1168 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
1170 PR middle-end/78468
1171 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
1172 settings of the virtual registers.
1174 Revert again
1175 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
1177 * explow.c (get_dynamic_stack_size): Take known alignment of stack
1178 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
1179 needed.
1181 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1183 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
1184 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
1186 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1188 * config/s390/s390.md: Add missing comments with the expanded
1189 mnemonics.
1190 * config/s390/vector.md: Likewise.
1191 * config/s390/vx-builtins.md: Likewise.
1193 2017-02-02 Jakub Jelinek <jakub@redhat.com>
1195 PR target/79197
1196 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
1197 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
1198 conditions on a single line.
1200 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1202 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
1203 __S390_VX__ to __VX__.
1205 2017-02-01 Andrew Pinski <apinski@cavium.com>
1207 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
1208 stmt_info to record_stmt_cost.
1209 (vect_get_known_peeling_cost): Pass stmt_info if known to
1210 record_stmt_cost.
1211 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
1212 cpu_vector_cost field into
1213 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
1214 field into vec_int_stmt_cost and vec_fp_stmt_cost.
1215 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
1216 splitting of scalar_stmt_cost and vec_stmt_cost.
1217 (thunderx_vector_cost): Likewise.
1218 (cortexa57_vector_cost): LIkewise.
1219 (exynosm1_vector_cost): Likewise.
1220 (xgene1_vector_cost): Likewise.
1221 (thunderx2t99_vector_cost): Improve after the splitting of the two
1222 fields.
1223 (aarch64_builtin_vectorization_cost): Update for the splitting of
1224 scalar_stmt_cost and vec_stmt_cost.
1226 2017-02-01 Torvald Riegel <triegel@redhat.com>
1227 Richard Henderson <rth@redhat.com>
1229 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
1230 conditional on existance of a fast atomic load.
1231 * optabs-query.c (can_atomic_load_p): New function.
1232 * optabs-query.h (can_atomic_load_p): Declare it.
1233 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
1234 no fast atomic load is available for the particular size of access.
1235 (expand_atomic_compare_and_swap): Likewise.
1236 (expand_atomic_load): Likewise.
1237 (expand_atomic_store): Likewise.
1238 (expand_atomic_fetch_op): Likewise.
1239 * testsuite/lib/target-supports.exp
1240 (check_effective_target_sync_int_128): Remove x86 because it provides
1241 no fast atomic load.
1242 (check_effective_target_sync_int_128_runtime): Likewise.
1244 2017-02-01 Richard Biener <rguenther@suse.de>
1246 * graphite.c: Include tree-vectorizer.h for find_loop_location.
1247 (graphite_transform_loops): Provide opt-info for optimized nests.
1248 * tree-parloop.c (parallelize_loops): Provide opt-info for
1249 parallelized loops.
1251 2017-02-01 Richard Biener <rguenther@suse.de>
1253 PR middle-end/79315
1254 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
1255 was not set before.
1257 2017-02-01 Richard Biener <rguenther@suse.de>
1259 PR tree-optimization/71824
1260 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
1261 Verify the loops are valid in the merged SESE region.
1262 (scop_detection::can_represent_loop_1): Check analyzing the
1263 evolution of the number of iterations in the region succeeds.
1265 2017-01-31 Ian Lance Taylor <iant@golang.org>
1267 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
1268 REG_ARGS_SIZE note to 32-bit push insns and call insn.
1270 2017-01-31 David Malcolm <dmalcolm@redhat.com>
1272 PR preprocessor/79210
1273 * input.c (get_substring_ranges_for_loc): Replace line_width
1274 assertion with error-handling.
1276 2017-01-31 Richard Biener <rguenther@suse.de>
1278 PR tree-optimization/77318
1279 * graphite-sese-to-poly.c (extract_affine): Fix assert.
1280 (create_pw_aff_from_tree): Take loop parameter.
1281 (add_condition_to_pbb): Pass loop of the condition to
1282 create_pw_aff_from_tree.
1284 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1286 * config/s390/s390.c (s390_asan_shadow_offset): New function.
1287 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
1289 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
1291 PR target/78597
1292 PR target/79038
1293 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
1294 no longer used.
1295 (convert_int_to_float128): Likewise.
1296 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
1297 (convert_int_to_float128): Likewise.
1298 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
1299 (UNSPEC_IEEE128_CONVERT): Likewise.
1300 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
1301 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
1302 Use local variables for IBM extended format.
1303 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
1304 (fix_trunc<mode>si2_fprs): Likewise.
1305 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
1306 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
1307 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
1308 to know that we can now have integers of all sizes in vector
1309 registers.
1310 (fix<uns>_<mode>di2_hw): Likewise.
1311 (float<uns>_<mode>si2_hw): Likewise.
1312 (fix_<mode>si2_hw): Likewise.
1313 (fixuns_<mode>si2_hw): Likewise.
1314 (float<uns>_<mode>di2_hw): Likewise.
1315 (float_<mode>di2_hw): Likewise.
1316 (float_<mode>si2_hw): Likewise.
1317 (floatuns_<mode>di2_hw): Likewise.
1318 (floatuns_<mode>si2_hw): Likewise.
1319 (xscvqp<su>wz_<mode>): Delete, no longer used.
1320 (xscvqp<su>dz_<mode>): Likewise.
1321 (xscv<su>dqp_<mode>): Likewise.
1322 (ieee128_mfvsrd_64bit): Likewise.
1323 (ieee128_mfvsrd_32bit): Likewise.
1324 (ieee128_mfvsrwz): Likewise.
1325 (ieee128_mtvsrw): Likewise.
1326 (ieee128_mtvsrd_64bit): Likewise.
1327 (ieee128_mtvsrd_32bit): Likewise.
1329 2017-01-31 Martin Liska <mliska@suse.cz>
1331 PR ipa/79285
1332 * ipa-prop.c (ipa_free_all_node_params): Call release method
1333 instead of ~sumbol_summary to not to trigger double times
1334 dtor of hash_map.
1336 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
1338 PR tree-optimization/71691
1339 * bitmap.h (class auto_bitmap): New.
1340 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
1341 is_maybe_undefined instead of ssa_undefined_value_p.
1343 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1345 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
1346 __S390_ARCH_LEVEL__ to __ARCH__.
1348 2017-01-31 Jakub Jelinek <jakub@redhat.com>
1350 PR tree-optimization/79267
1351 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
1352 if should_remove_lhs_p is true.
1354 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
1356 PR debug/63238
1357 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
1358 (add_alignment_attribute): New.
1359 (base_type_die): Add alignment attribute.
1360 (subrange_type_die): Likewise.
1361 (modified_type_die): Likewise.
1362 (gen_array_type_die): Likewise.
1363 (gen_descr_array_type_die: Likewise.
1364 (gen_enumeration_type_die): Likewise.
1365 (gen_subprogram_die): Likewise.
1366 (gen_variable_die): Likewise.
1367 (gen_field_die): Likewise.
1368 (gen_ptr_to_mbr_type_die): Likewise.
1369 (gen_struct_or_union_type_die): Likewise.
1370 (gen_subroutine_type_die): Likewise.
1371 (gen_typedef_die): Likewise.
1372 (base_type_cmp): Compare alignment attribute.
1374 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1376 PR target/79170
1377 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
1378 (setb_unsigned) New pattern for setb with CCUNS.
1379 * config/rs6000/rs6000.c (expand_block_compare): Use a different
1380 subfc./subfe sequence to avoid overflow problems. Generate a
1381 shorter sequence with cmpld/setb for power9.
1382 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
1383 for generating subfc. instruction.
1384 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
1385 now uses this instruction.
1387 2017-01-30 Ian Lance Taylor <iant@google.com>
1389 PR debug/79289
1390 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
1391 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
1393 2017-01-30 Martin Sebor <msebor@redhat.com>
1395 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
1396 Move constant to the right of a relational operator.
1397 (get_mpfr_format_length, format_character, format_string): Ditto.
1398 (should_warn_p, maybe_warn): Same.
1400 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
1402 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
1404 PR lto/79061
1405 * asan.c (get_translation_unit_decl): Remove function.
1406 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
1408 2017-01-30 Martin Liska <mliska@suse.cz>
1410 PR gcov-profile/79259
1411 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
1412 -fprofile-generate.
1414 2017-01-30 Martin Liska <mliska@suse.cz>
1416 PR bootstrap/78985
1417 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
1418 Initialize variables with NULL value.
1420 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
1422 PR target/79260
1423 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
1424 tm_p_file.
1425 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
1427 2017-01-30 Richard Biener <rguenther@suse.de>
1429 PR tree-optimization/79276
1430 * tree-vrp.c (process_assert_insertions): Properly adjust common
1431 when removing a duplicate.
1433 * gcc.dg/torture/pr79276.c: New testcase.
1435 2017-01-30 Richard Biener <rguenther@suse.de>
1437 PR tree-optimization/79256
1438 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
1439 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
1440 alignment on TYPE.
1441 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
1443 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
1445 PR target/79240
1446 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
1447 ("*r<noxa>sbg_<mode>_sll_bitmask")
1448 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
1449 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
1450 Use contiguous_bitmask_nowrap_operand.
1452 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1454 PR target/79268
1455 * config/rs6000/altivec.h (vec_xl): Revise #define.
1456 (vec_xst): Likewise.
1458 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
1460 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
1462 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
1464 PR rtl-optimization/79194
1465 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
1466 traps before call to bypass_conditional_jumps.
1468 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1470 PR tree-optimization/71374
1471 * lra-constraints.c (check_conflict_input_operands): New.
1472 (match_reload): Use it.
1474 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
1476 PR target/79131
1477 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
1478 account to calculate conflict_set.
1480 2017-01-27 Bin Cheng <bin.cheng@arm.com>
1482 PR rtl-optimization/78559
1483 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
1484 other_insn in combine.
1486 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
1488 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
1489 uint16_type_node for BT_UINT16.
1491 2017-01-27 David Malcolm <dmalcolm@redhat.com>
1493 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
1494 "RTL Tests" to menu.
1495 (GIMPLE Tests): New node.
1496 (RTL Tests): New node.
1498 2017-01-27 Richard Biener <rguenther@suse.de>
1500 PR tree-optimization/79245
1501 * tree-loop-distribution.c (distribute_loop): Apply cost
1502 modeling also to detected patterns.
1504 2017-01-27 Richard Biener <rguenther@suse.de>
1506 PR tree-optimization/71433
1507 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
1508 (compare_assert_loc): New function.
1509 (process_assert_insertions): Sort and optimize assert locations
1510 to remove duplicates and push down identical assertions on
1511 edges to their destination block.
1513 2017-01-27 Richard Biener <rguenther@suse.de>
1515 PR tree-optimization/79244
1516 * tree-vrp.c (remove_range_assertions): Forcefully propagate
1517 out SSA names even if abnormal.
1519 2017-01-27 Jakub Jelinek <jakub@redhat.com>
1521 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
1522 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
1523 instead of MPFR_RNDN.
1525 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
1527 PR target/79239
1528 * arm.c (arm_option_override): Don't call build_target_option_node
1529 until after doing all option overrides.
1530 (arm_valid_target_attribute_tree): Likewise.
1532 2017-01-27 Martin Liska <mliska@suse.cz>
1534 * doc/invoke.texi (-fprofile-arcs): Document profiling support
1535 for {cd}tors and C++ {cd}tors.
1537 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
1539 * config/s390/s390.md ("*setmem_long_and")
1540 ("*setmem_long_and_31z"): Use zero_extend instead of and.
1542 2017-01-26 Martin Sebor <msebor@redhat.com>
1544 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
1545 of precision.
1547 2017-01-26 Martin Sebor <msebor@redhat.com>
1549 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
1550 HAVE_DFmode before using XFmode or DFmode.
1551 (parse_directive): Avoid using the z length modifier to avoid
1552 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
1554 PR middle-end/78703
1555 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
1556 to accept adjustment as an array.
1557 (get_int_range): New function.
1558 (struct directive): Make width and prec arrays.
1559 (directive::set_width, directive::set_precision): Call get_int_range.
1560 (format_integer, format_floating): Handle width and precision ranges.
1561 (format_string, parse_directive): Same.
1563 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1565 PR debug/79129
1566 * dwarf2out.c (generate_skeleton_bottom_up): For children with
1567 comdat_type_p set, just clone them, but keep the children in the
1568 original DIE.
1570 PR debug/78835
1571 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
1572 which have direct callers with -fvar-tracking-assignments enabled
1573 in the current TU.
1574 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
1575 inside of type units.
1577 2017-01-26 Martin Sebor <msebor@redhat.com>
1579 PR middle-end/78703
1580 * gimple-ssa-sprintf.c (struct result_range): Add likely and
1581 unlikely counters.
1582 (struct format_result): Replace number_chars, number_chars_min,
1583 and number_chars_max with a single member of struct result_range.
1584 Remove bounded.
1585 (format_result::operator+=): Adjust.
1586 (struct fmtresult): Remove bounded. Handle likely and unlikely
1587 counters.
1588 (fmtresult::adjust_for_width_or_precision): New function.
1589 (fmtresult:type_max_digits): New function.
1590 (bytes_remaining): Handle likely and unlikely counters.
1591 (min_bytes_remaining): Remove.
1592 (format_percent): Simplify.
1593 (format_integer, format_floating): Set likely and unlikely counters.
1594 (get_string_length, format_character, format_string): Same.
1595 (format_plain, should_warn_p): New function.
1596 (maybe_warn): Call should_warn_p. Update diagnostic messages
1597 and handle those for all directives, including plain strings.
1598 (format_directive): Handle likely and unlikely counters.
1599 Remove unnecessary quoting from diagnostics. Add an informational
1600 note.
1601 (add_bytes): Remove.
1602 (pass_sprintf_length::compute_format_length): Simplify.
1603 (try_substitute_return_value): Handle likely and unlikely counters.
1605 2017-01-26 Carl Love <cel@us.ibm.com>
1607 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
1608 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
1610 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
1612 PR target/79131
1613 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
1614 endianess for subregs into account.
1615 * lra-constraints.c (lra_constraints): Do risky transformations
1616 always on the first iteration.
1617 * lra-lives.c (check_pseudos_live_through_calls): Add arg
1618 last_call_used_reg_set.
1619 (process_bb_lives): Define and use last_call_used_reg_set.
1620 * lra.c (lra): Always continue after lra_constraints on the first
1621 iteration.
1623 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
1625 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
1626 constant.
1627 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
1629 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1631 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
1632 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
1633 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
1634 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
1635 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
1636 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
1637 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
1638 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
1639 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
1641 2017-01-26 Marek Polacek <polacek@redhat.com>
1643 PR c/79199
1644 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
1645 for the third operand.
1647 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1649 PR middle-end/79236
1650 * omp-low.c (struct omp_context): Add simt_stmt field.
1651 (scan_omp_for): Return omp_context *.
1652 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
1653 context to the _simt_ SIMD stmt.
1654 (lower_omp_for): For combined SIMD with sibling _simt_
1655 SIMD, make sure to use the same decls in _looptemp_
1656 clauses as in the sibling.
1658 2017-01-26 David Sherwood <david.sherwood@arm.com>
1660 PR middle-end/79212
1661 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
1662 all contexts.
1664 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1666 PR target/70465
1667 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
1668 emit fld b; fld a; if possible.
1670 * brig-builtins.def: Update copyright years.
1671 * config/arm/arm_acle_builtins.def: Update copyright years.
1673 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
1675 PR target/79179
1676 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
1677 constraint instead of o for the stxsd instruction.
1679 2017-01-25 Carl Love <cel@us.ibm.com>
1681 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
1682 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
1684 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
1686 * doc/invoke.texi (C++ Dialect Options): Fix typo.
1688 2017-01-25 Richard Biener <rguenther@suse.de>
1690 PR tree-optimization/69264
1691 * target.def (vector_alignment_reachable): Improve documentation.
1692 * doc/tm.texi: Regenerate.
1693 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
1694 and add a comment.
1695 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
1696 earlier changes with respect to TYPE_USER_ALIGN.
1697 (vector_alignment_reachable_p): Likewise. Improve dumping.
1699 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1701 PR target/79145
1702 * config/arm/arm.md (xordi3): Force constant operand into a register
1703 for TARGET_IWMMXT.
1705 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1707 * doc/invoke.texi (-fstore-merging): Correct default optimization
1708 levels at which it is enabled.
1709 (-O): Move -fstore-merging from list to...
1710 (-O2): ... Here.
1712 2017-01-25 Richard Biener <rguenther@suse.de>
1714 PR debug/78363
1715 * omp-expand.c: Include debug.h.
1716 (expand_omp_taskreg): Make sure to generate early debug before
1717 outlining anything from a function.
1718 (expand_omp_target): Likewise.
1719 (grid_expand_target_grid_body): Likewise.
1721 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
1723 PR lto/79061
1724 * asan.c (get_translation_unit_decl): New function.
1725 (asan_add_global): Extract modules file name from globals
1726 TRANSLATION_UNIT_DECL name.
1728 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
1730 PR target/77439
1731 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
1732 for long calls with APCS frame and VFP.
1734 2017-01-24 David Malcolm <dmalcolm@redhat.com>
1736 * cfg.c (original_copy_tables_initialized_p): New function.
1737 * cfg.h (original_copy_tables_initialized_p): New decl.
1738 * cfgrtl.c (relink_block_chain): Guard the call to
1739 free_original_copy_tables with a call to
1740 original_copy_tables_initialized_p.
1741 * cgraph.h (symtab_node::native_rtl_p): New decl.
1742 * cgraphunit.c (symtab_node::native_rtl_p): New function.
1743 (symtab_node::needed_p): Don't assert for early assembly output
1744 for __RTL functions.
1745 (cgraph_node::finalize_function): Set "force_output" for __RTL
1746 functions.
1747 (cgraph_node::analyze): Bail out early for __RTL functions.
1748 (analyze_functions): Update assertion to support __RTL functions.
1749 (cgraph_node::expand): Bail out early for __RTL functions.
1750 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
1751 __RTL functions.
1752 * function.h (struct function): Update comment for field
1753 "pass_startwith".
1754 * gimple-expr.c: Include "tree-pass.h".
1755 (gimple_has_body_p): Return false for __RTL functions.
1756 * Makefile.in (OBJS): Add run-rtl-passes.o.
1757 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
1758 accessor.
1759 (gcc::pass_manager::get_clean_slate): New accessor.
1760 * passes.c: Include "insn-addr.h".
1761 (should_skip_pass_p): Add logging. Update logic for running
1762 "expand" to be compatible with both __GIMPLE and __RTL. Guard
1763 property-provider override so it is only done for gimple passes.
1764 Don't skip dfinit.
1765 (skip_pass): New function.
1766 (execute_one_pass): Call skip_pass when skipping passes.
1767 * read-md.c (md_reader::read_char): Support filtering
1768 the input to a subset of line numbers.
1769 (md_reader::md_reader): Initialize fields
1770 m_first_line and m_last_line.
1771 (md_reader::read_file_fragment): New function.
1772 * read-md.h (md_reader::read_file_fragment): New decl.
1773 (md_reader::m_first_line): New field.
1774 (md_reader::m_last_line): New field.
1775 * read-rtl-function.c (function_reader::create_function): Only
1776 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
1777 curr_properties. Set DECL_INITIAL to a dummy block.
1778 (read_rtl_function_body_from_file_range): New function.
1779 * read-rtl-function.h (read_rtl_function_body_from_file_range):
1780 New decl.
1781 * run-rtl-passes.c: New file.
1782 * run-rtl-passes.h: New file.
1784 2017-01-24 Jeff Law <law@redhat.com>
1786 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
1787 buffer size.
1789 2017-01-24 Bin Cheng <bin.cheng@arm.com>
1791 PR tree-optimization/79159
1792 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
1793 (record_nonwrapping_iv): Improve boundary using above function if no
1794 value range information.
1796 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
1797 Martin Jambor <mjambor@suse.cz>
1799 * brig-builtins.def: New file.
1800 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
1801 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
1802 (DEF_HSAIL_SAT_BUILTIN): Likewise.
1803 (DEF_HSAIL_INTR_BUILTIN): Likewise.
1804 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
1805 * builtin-types.def (BT_INT8): New.
1806 (BT_INT16): Likewise.
1807 (BT_UINT8): Likewise.
1808 (BT_UINT16): Likewise.
1809 (BT_FN_ULONG): Likewise.
1810 (BT_FN_UINT_INT): Likewise.
1811 (BT_FN_UINT_ULONG): Likewise.
1812 (BT_FN_UINT_LONG): Likewise.
1813 (BT_FN_UINT_PTR): Likewise.
1814 (BT_FN_ULONG_PTR): Likewise.
1815 (BT_FN_INT8_FLOAT): Likewise.
1816 (BT_FN_INT16_FLOAT): Likewise.
1817 (BT_FN_UINT32_FLOAT): Likewise.
1818 (BT_FN_UINT16_FLOAT): Likewise.
1819 (BT_FN_UINT8_FLOAT): Likewise.
1820 (BT_FN_UINT64_FLOAT): Likewise.
1821 (BT_FN_UINT16_UINT32): Likewise.
1822 (BT_FN_UINT32_UINT16): Likewise.
1823 (BT_FN_UINT16_UINT16_UINT16): Likewise.
1824 (BT_FN_INT_PTR_INT): Likewise.
1825 (BT_FN_UINT_PTR_UINT): Likewise.
1826 (BT_FN_LONG_PTR_LONG): Likewise.
1827 (BT_FN_ULONG_PTR_ULONG): Likewise.
1828 (BT_FN_VOID_UINT64_UINT64): Likewise.
1829 (BT_FN_UINT8_UINT8_UINT8): Likewise.
1830 (BT_FN_INT8_INT8_INT8): Likewise.
1831 (BT_FN_INT16_INT16_INT16): Likewise.
1832 (BT_FN_INT_INT_INT): Likewise.
1833 (BT_FN_UINT_FLOAT_UINT): Likewise.
1834 (BT_FN_FLOAT_UINT_UINT): Likewise.
1835 (BT_FN_ULONG_UINT_UINT): Likewise.
1836 (BT_FN_ULONG_UINT_PTR): Likewise.
1837 (BT_FN_ULONG_ULONG_ULONG): Likewise.
1838 (BT_FN_UINT_UINT_UINT): Likewise.
1839 (BT_FN_VOID_UINT_PTR): Likewise.
1840 (BT_FN_UINT_UINT_PTR: Likewise.
1841 (BT_FN_UINT32_UINT64_PTR): Likewise.
1842 (BT_FN_INT_INT_UINT_UINT): Likewise.
1843 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
1844 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
1845 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
1846 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
1847 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
1848 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
1849 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
1850 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
1851 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
1852 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
1853 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
1854 * doc/frontends.texi: List BRIG FE.
1855 * doc/install.texi (Testing): Add BRIG tesring requirements.
1856 * doc/invoke.texi (Overall Options): Mention BRIG.
1857 * doc/standards.texi (Standards): Doucment BRIG HSA version.
1859 2017-01-24 Richard Biener <rguenther@suse.de>
1861 PR translation/79208
1862 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
1864 2017-01-24 Martin Jambor <mjambor@suse.cz>
1866 PR bootstrap/79198
1867 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
1868 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
1869 and known_contexts.
1871 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
1873 PR middle-end/79123
1874 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
1875 casts from signed to unsigned really don't have a range.
1877 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1879 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
1880 GMP_RNDx for compatiblity.
1882 2017-01-24 Martin Liska <mliska@suse.cz>
1884 PR bootstrap/79132
1885 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
1886 that would prevent us to call alloca with -1 as argument.
1888 2017-01-24 Jakub Jelinek <jakub@redhat.com>
1890 * dwarf2out.c (output_compilation_unit_header, output_file_names):
1891 Avoid -Wformat-security warning.
1893 2017-01-23 Andrew Pinski <apinski@cavium.com>
1895 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
1896 cost table.
1898 2017-01-23 Martin Sebor <msebor@redhat.com>
1900 PR middle-end/78703
1901 * gimple-ssa-sprintf.c (warn_level): New global.
1902 (format_integer): Use it here and throughout the rest of the file.
1903 Use the same switch to compute sign as base.
1904 (maybe_warn): New function.
1905 (format_directive): Factor out warnings into maybe_warn.
1906 Add debugging output. Use warn_level.
1907 (add_bytes): Use warn_level.
1908 (pass_sprintf_length::compute_format_length): Add debugging output.
1909 (try_substitute_return_value): Same.
1910 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
1912 PR middle-end/78703
1913 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
1914 (struct fmtresult, format_integer, format_floating): Adjust.
1915 (fmtresult::fmtresult): Set max correctly in two argument ctor.
1916 (get_string_length, format_string,format_directive): Same.
1917 (pass_sprintf_length::compute_format_length): Same.
1918 (try_substitute_return_value): Simplify slightly.
1920 PR middle-end/78703
1921 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
1922 (fmtresult::operator+=): Outlined.
1923 (struct fmtresult): Add ctors.
1924 (struct conversion_spec): Rename...
1925 (struct directive): ...to this. Add and remove data members.
1926 (directive::set_width, directive::set_precision): New functions.
1927 (format_percent): Use fmtresult ctor.
1928 (get_width_and_precision): Remove.
1929 (format_integer): Make naming changes. Avoid computing width and
1930 precision.
1931 (format_floating): Same. Adjust indentation.
1932 (format_character, format_none): New functions.
1933 (format_string): Moved character handling to format_character.
1934 (format_directive): Remove arguments, change return type.
1935 (parse_directive): New function.
1936 (pass_sprintf_length::compute_format_length): Move directive
1937 parsing to parse_directive.
1939 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1941 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
1942 (assign_assembler_name_if_needed): ... this.
1943 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
1944 (assign_assembler_name_if_needed): ... this.
1945 (free_lang_data_in_cgraph): Adjust callers.
1946 * cgraphunit.c (cgraph_node::analyze): Likewise.
1947 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
1948 Likewise.
1950 2017-01-23 Richard Biener <rguenther@suse.de>
1952 PR tree-optimization/79088
1953 PR tree-optimization/79188
1954 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
1955 resetting loop bounds after last path deletion. Reset loop
1956 bounds of the target loop, make code match the comments.
1957 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
1958 Make sure loops need no fixups.
1960 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1962 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1963 exponent support with double type for first argument.
1964 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1965 type returned by __builtin_vec_extract_sig,
1966 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1967 functions from "vector int" to "vector unsigned int" or from
1968 "vector long long int" to "vector unsigned long long int".
1969 Changed type returned by __builtin_vec_extract_exp,
1970 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1971 functions from "vector int" to "vector unsigned int" or from
1972 "vector long long int" to "vector unsigned long long int".
1973 Changed return type of __builtin_vec_test_data_class,
1974 __builtin_vec_test_data_class_sp, and
1975 __builtin_vec_test_data_class_dp from "vector int" to
1976 "vector bool int" or from "vector long long int" to "vector bool
1977 long long int" and changed second argument type from "unsigned
1978 int" to "int". Added new overloaded function forms "vector float
1979 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1980 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1981 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1982 double, vector unsigned long long int)" and "vector double
1983 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1984 long int)". Changed return type of
1985 __builtin_scalar_test_data_class and
1986 __builtin_scalar_test_data_class_sp and
1987 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1988 int" and changed second argument from "unsigned int" to "int".
1989 Changed type returned by __builtin_scalar_test_neg,
1990 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1991 from "int" to "bool int". Added new overloaded function form
1992 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1993 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1994 exponent double-precision with floating point first argument.
1995 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1996 documentation of scalar_test_data_class, scalar_test_neg,
1997 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1998 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1999 vec_test_data_class built-in functions to reflect refinements in
2000 their type signatures.
2002 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
2004 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
2005 size of buf.
2006 (aarch64_elf_asm_destructor): Likewise.
2008 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
2010 PR rtl-optimization/78634
2011 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
2012 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
2013 * ifcvt.c (noce_try_cmove): Add missing cost check.
2015 PR rtl-optimization/71724
2016 * combine.c (if_then_else_cond): Look for situations where it is
2017 beneficial to undo the work of one of the recursive calls.
2019 2017-01-23 Bin Cheng <bin.cheng@arm.com>
2021 PR tree-optimization/70754
2022 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
2023 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
2024 combined stmt before it if not NULL.
2025 (combine_chains): Process refs reversely and compute dominance point
2026 for root ref.
2028 2017-01-23 Martin Liska <mliska@suse.cz>
2030 PR tree-optimization/79196
2031 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
2032 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
2033 instead of memcmp.
2034 (strlen_optimize_stmt): Call the renamed function.
2036 2017-01-23 Michael Matz <matz@suse.de>
2038 PR tree-optimization/78384
2039 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
2041 2017-01-23 Richard Biener <rguenther@suse.de>
2043 PR tree-optimization/79186
2044 * tree-vrp.c (register_new_assert_for): Make sure we've seen
2045 both incoming edges before moving an assert.
2047 2017-01-23 Martin Jambor <mjambor@suse.cz>
2049 * ipa-prop.c (load_from_param_1): Removed.
2050 (load_from_unmodified_param): Bits from load_from_param_1 put back
2051 here.
2052 (load_from_param): Removed.
2053 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
2054 with stmt. Reverted back to use of load_from_unmodified_param.
2056 2017-01-23 Martin Jambor <mjambor@suse.cz>
2058 PR ipa/79108
2059 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
2060 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
2061 field a pointer to garbage collected vector, mark lattices and
2062 ipcp_orig_node with GTY((skip)).
2063 (ipa_get_param_count): Adjust to descriptors being a pointer.
2064 (ipa_get_param): Likewise.
2065 (ipa_get_type): Likewise.
2066 (ipa_get_param_move_cost): Likewise.
2067 (ipa_set_param_used): Likewise.
2068 (ipa_get_controlled_uses): Likewise.
2069 (ipa_set_controlled_uses): Likewise.
2070 (ipa_is_param_used): Likewise.
2071 (ipa_node_params_t): Move into garbage collector. New methods insert
2072 and remove.
2073 (ipa_node_params_sum): Annotate wth GTY(()).
2074 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
2075 garbage collected.
2076 (ipa_load_from_parm_agg): Adjust declaration.
2077 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
2078 * ipa-profile.c (ipa_profile): Likewise.
2079 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
2080 (ipa_populate_param_decls): Make descriptors parameter garbage
2081 collected.
2082 (ipa_dump_param): Adjust to descriptors being a pointer.
2083 (ipa_alloc_node_params): Likewise.
2084 (ipa_initialize_node_params): Likewise.
2085 (load_from_param_1): Make descriptors parameter garbage collected.
2086 (load_from_unmodified_param): Likewise.
2087 (load_from_param): Likewise.
2088 (ipa_load_from_parm_agg): Likewise.
2089 (ipa_node_params::~ipa_node_params): Removed.
2090 (ipa_free_all_node_params): Remove call to delete operator.
2091 (ipa_node_params_t::insert): New.
2092 (ipa_node_params_t::remove): Likewise.
2093 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
2094 copy known_csts and known_contexts vectors.
2095 (ipa_read_node_info): Adjust to descriptors being a pointer.
2096 (ipcp_modif_dom_walker): Make m_descriptors field garbage
2097 collected.
2098 (ipcp_transform_function): Make descriptors variable garbage
2099 collected.
2101 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
2103 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
2104 * config/i386/avx512dqintrin.h: Ditto.
2105 * config/i386/avx512fintrin.h: Ditto.
2106 * gcc/config/i386/i386.c: Handle new builtins.
2107 * config/i386/i386-builtin.def: Add new builtins.
2108 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
2109 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
2111 2017-01-23 Jakub Jelinek <jakub@redhat.com>
2112 Martin Liska <mliska@suse.cz>
2114 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
2115 * asan.c (asan_expand_poison_ifn): Support stores and use
2116 appropriate ASAN report function.
2117 * internal-fn.c (expand_ASAN_POISON_USE): New function.
2118 * internal-fn.def (ASAN_POISON_USE): Declare.
2119 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
2120 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
2121 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
2122 ASAN_POISON calls w/o LHS.
2123 * tree-ssa.c (execute_update_addresses_taken): Create clobber
2124 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
2125 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
2126 * gimplify.c (asan_poison_variables): Add attribute
2127 use_after_scope_memory to variables that really needs to live
2128 in memory.
2129 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
2130 having the attribute.
2132 2017-01-23 Martin Liska <mliska@suse.cz>
2134 * asan.c (create_asan_shadow_var): New function.
2135 (asan_expand_poison_ifn): Likewise.
2136 * asan.h (asan_expand_poison_ifn): New declaration.
2137 * internal-fn.c (expand_ASAN_POISON): Likewise.
2138 * internal-fn.def (ASAN_POISON): New builtin.
2139 * sanopt.c (pass_sanopt::execute): Expand
2140 asan_expand_poison_ifn.
2141 * tree-inline.c (copy_decl_for_dup_finish): Make function
2142 external.
2143 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
2144 * tree-ssa.c (is_asan_mark_p): New function.
2145 (execute_update_addresses_taken): Rewrite local variables
2146 (identified just by use-after-scope as addressable) into SSA.
2148 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
2150 * doc/install.texi (Specific): opensource.apple.com uses https
2151 now. Remove trailing slash.
2153 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
2155 * README.Portability: Remove note on an Irix compatibility issue.
2157 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
2159 * gcov.c (INCLUDE_ALGORITHM): Define.
2160 (INCLUDE_VECTOR): Define.
2161 No longer include <vector> and <algorithm> directly.
2163 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
2165 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
2166 to https.
2167 * doc/invoke.texi (Code Gen Options): Ditto.
2169 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
2171 PR lto/78407
2172 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
2174 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
2176 rtl-optimization/79125
2177 * cprop.c (local_cprop_pass): Handle cases where we make an
2178 unconditional trap.
2180 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
2182 PR target/61729
2183 PR target/77850
2184 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
2185 read from, for big endian.
2187 2017-01-20 Jiong Wang <jiong.wang@arm.com>
2189 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
2190 register pauth builtins for LP64 only.
2192 2017-01-20 Marek Polacek <polacek@redhat.com>
2194 PR c/79152
2195 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
2196 non-case labels.
2198 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
2200 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
2201 of safelen status.
2202 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
2203 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
2204 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
2206 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2208 PR target/71270
2209 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
2210 in big-endian mode when they are not a single duplicated value.
2212 2017-01-20 Richard Biener <rguenther@suse.de>
2214 * BASE-VER: Bump to 7.0.1.
2216 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
2218 * omp-low.c (omplow_simd_context): New struct. Use it...
2219 (lower_rec_simd_input_clauses): ...here and...
2220 (lower_rec_input_clauses): ...here to hold common data. Adjust all
2221 references to idx, lane, max_vf, is_simt.
2223 2017-01-20 Graham Markall <graham.markall@embecosm.com>
2225 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
2226 mcpu=nps400.
2228 2017-01-20 Martin Jambor <mjambor@suse.cz>
2230 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
2231 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
2232 gt-hsa-common.h.
2233 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
2234 (GTFILES): Rename hsa.c to hsa-common.c.
2235 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
2236 * hsa-dump.c: Likewise.
2237 * hsa-gen.c: Likewise.
2238 * hsa-regalloc.c: Likewise.
2239 * ipa-hsa.c: Likewise.
2240 * omp-expand.c: Likewise.
2241 * omp-low.c: Likewise.
2242 * toplev.c: Likewise.
2244 2017-01-20 Marek Polacek <polacek@redhat.com>
2246 PR c/64279
2247 * doc/invoke.texi: Document -Wduplicated-branches.
2248 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
2249 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
2250 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
2251 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
2252 return 0 only when not OEP_LEXICOGRAPHIC.
2253 (fold_build_cleanup_point_expr): Use the expression
2254 location when building CLEANUP_POINT_EXPR.
2255 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
2256 * tree.c (add_expr): Handle error_mark_node.
2258 2017-01-20 Martin Liska <mliska@suse.cz>
2260 PR lto/69188
2261 * tree-profile.c (init_ic_make_global_vars): Do not call
2262 finalize_decl.
2263 (gimple_init_gcov_profiler): Likewise.
2265 2017-01-20 Martin Liska <mliska@suse.cz>
2267 PR ipa/71190
2268 * cgraph.h (maybe_create_reference): Remove argument and
2269 update comment.
2270 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
2271 argument.
2272 * ipa-cp.c (create_specialized_node): Likewise.
2273 * symtab.c (symtab_node::maybe_create_reference): Handle
2274 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
2276 2017-01-20 Martin Liska <mliska@suse.cz>
2278 * read-rtl-function.c (function_reader::create_function): Use
2279 build_decl instread of build_decl_stat.
2281 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
2283 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
2284 * config/i386/avx512dqintrin.h: Ditto.
2285 * config/i386/avx512fintrin.h: Ditto.
2286 * config/i386/i386-builtin-types.def: Add new types.
2287 * gcc/config/i386/i386.c: Handle new types.
2288 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
2289 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
2290 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
2291 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
2292 (__builtin_ia32_kshiftridi): New.
2293 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
2295 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
2297 PR target/78875
2298 PR target/79140
2299 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
2300 define to rs6000_init_stack_protect_guard.
2301 (rs6000_init_stack_protect_guard): New function.
2303 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2304 Yunqiang Su <yunqiang.su@imgtec.com>
2306 * config.gcc (supported_defaults): Add madd4.
2307 (with_madd4): Add validation.
2308 (all_defaults): Add madd4.
2309 * config/mips/mips.opt (mmadd4): New option.
2310 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2311 mmadd4.
2312 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2313 __mips_no_madd4.
2314 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
2315 (ISA_HAS_FUSED_MADD4): Likewise.
2316 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
2317 * gcc/doc/install.texi (--with-madd4): Document the new option.
2319 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2321 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
2322 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
2323 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
2324 (aarch64_init_pauth_hint_builtins): New.
2325 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
2326 (aarch64_expand_builtin): Expand new builtins.
2328 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2330 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
2331 * combine-stack-adj.c (no_unhandled_cfa): Handle
2332 REG_CFA_TOGGLE_RA_MANGLE.
2333 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
2334 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
2335 info for return address signing.
2336 (aarch64_expand_epilogue): Likewise.
2338 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2340 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
2341 * config/aarch64/aarch64-protos.h
2342 (aarch64_return_address_signing_enabled): New declaration.
2343 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
2344 New function.
2345 (aarch64_expand_prologue): Sign return address before it's pushed onto
2346 stack.
2347 (aarch64_expand_epilogue): Authenticate return address fetched from
2348 stack.
2349 (aarch64_override_options): Sanity check for ILP32 and ISA level.
2350 (aarch64_attributes): New function attributes for "sign-return-address".
2351 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
2352 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
2353 ("*do_return"): Generate combined instructions according to key index.
2354 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
2355 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
2356 iterators.
2357 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
2358 * config/aarch64/aarch64.opt (msign-return-address=): New.
2359 * doc/extend.texi (AArch64 Function Attributes): Documents
2360 "sign-return-address=".
2361 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
2363 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
2365 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
2366 overall option summary.
2368 2017-01-19 Jiong Wang <jiong.wang@arm.com>
2370 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
2371 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
2372 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
2373 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
2375 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
2377 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
2378 -mpower9-minmax by default for -mcpu=power9.
2379 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
2380 128-bit floating point.
2382 2017-01-20 Alan Modra <amodra@gmail.com>
2384 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
2385 optimizing for size.
2387 2017-01-20 Alan Modra <amodra@gmail.com>
2389 PR target/79144
2390 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
2391 for strcmp and strncmp from corresponding builtin decl.
2393 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2395 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
2396 instead of i386/rtems-64.h.
2397 * config/i386/rtems-64.h: Remove.
2399 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
2401 PR target/78478
2402 Revert:
2403 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2405 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
2407 2017-01-19 Tamar Christina <tamar.christina@arm.com>
2409 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
2410 Change int to HOST_WIDE_INT.
2411 * config/aarch64/aarch64-protos.h
2412 (aarch64_simd_gen_const_vector_dup): Likewise.
2413 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
2415 2017-01-19 David Malcolm <dmalcolm@redhat.com>
2417 * langhooks-def.h (lhd_type_for_size): New decl.
2418 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
2419 * langhooks.c (lhd_type_for_size): New function, taken from
2420 lto_type_for_size.
2422 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
2424 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
2425 define_bypass for CR latency.
2426 (power9-cracked-alu): Update bypass latency and remove power9-branch.
2427 (power9-alu2): Add define_bypass for CR latency.
2428 (power9-cmp): New.
2429 (power9-mul): Update insn latency.
2430 (power9-mul-compare): Update insn latency, bypass latency and remove
2431 power9-branch.
2433 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2435 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
2436 Delete.
2437 * config/aarch64/aarch64.md
2438 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
2439 aarch64_nopcrelative_literal_loads.
2440 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
2442 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
2444 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
2445 TARGET_LOONGSON_3A.
2446 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
2448 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
2450 PR target/78176
2451 * config.gcc (supported_defaults): Add lxc1-sxc1.
2452 (with_lxc1_sxc1): Add validation.
2453 (all_defaults): Add lxc1-sxc1.
2454 * config/mips/mips.opt (mlxc1-sxc1): New option.
2455 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
2456 mlxc1-sxc1.
2457 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
2458 __mips_no_lxc1_sxc1.
2459 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
2460 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
2461 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
2463 2017-01-19 Richard Biener <rguenther@suse.de>
2465 PR tree-optimization/72488
2466 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
2467 sure to restore SSA info.
2468 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
2470 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
2472 PR rtl-optimization/79121
2473 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
2474 of the inner type when shifting an extended value.
2476 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2478 PR lto/78407
2479 * symtab.c (symtab_node::equal_address_to): Fix comparing of
2480 interposable aliases.
2482 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
2484 PR target/78516
2485 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
2486 Use the evmergelohi instruction.
2487 (mov_si<mode>_e500_subreg4_2_le): Likewise.
2488 (mov_sitf_e500_subreg8_2_be): Likewise.
2489 (mov_sitf_e500_subreg12_2_le): Likewise.
2490 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
2491 (mov_si<mode>_e500_subreg4_2_be): Likewise.
2492 (mov_sitf_e500_subreg8_2_le): Likewise.
2493 (mov_sitf_e500_subreg12_2_be): Likewise.
2495 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2497 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
2498 attribute from vecsimple to vecperm.
2499 (altivec_vbpermq2): Likewise.
2501 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2503 PR target/79040
2504 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
2506 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
2507 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
2508 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
2509 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
2510 case where N arg is SIZE_MAX.
2511 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
2512 (cmpstrsi): Add pattern.
2514 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2516 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2517 __builtin_vec_revb builtins.
2518 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
2519 built-in functions to support generation of the ISA 3.0 XXBR<x>
2520 vector byte reverse instructions.
2521 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
2522 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
2523 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
2524 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
2525 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
2526 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
2527 (P9V_BUILTIN_VEC_REVB): Likewise.
2528 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
2529 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
2530 (p9_xxbrq_v16qi): Likewise.
2531 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
2532 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
2533 (p9_xxbrh_v8hi): Likewise.
2534 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
2535 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
2536 vec_revb built-in functions.
2538 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
2540 PR rtl-optimization/78952
2541 * config/i386/i386.md (any_extract): New code iterator.
2542 (*insvqi_2): Use any_extract for source operand.
2543 (*insvqi_3): Use any_shiftrt for source operand.
2545 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
2547 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
2548 New function.
2549 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
2551 2017-01-18 Matthias Klose <doko@ubuntu.com>
2553 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
2555 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2557 * config/rs6000/altivec.h (vec_bperm): Change #define.
2558 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
2559 (altivec_vbpermq2): New define_insn.
2560 (altivec_vbpermd): Likewise.
2561 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
2562 function interface.
2563 (VBPERMD): Likewise.
2564 (VBPERM): New polymorphic function interface.
2565 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
2566 Add entries for P9V_BUILTIN_VEC_VBPERM.
2567 * doc/extend.texi: Add interfaces for vec_bperm.
2569 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2571 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
2572 first letter of error messages.
2573 (s390_resolve_overloaded_builtin): Likewise.
2574 * config/s390/s390.c (s390_expand_builtin): Likewise.
2575 (s390_invalid_arg_for_unprototyped_fn): Likewise.
2576 (s390_valid_target_attribute_inner_p): Likewise.
2577 * config/s390/s390.md ("tabort"): Likewise.
2579 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
2581 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
2582 (ISA_AVOID_DIV_HILO): New macro.
2583 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
2584 (ISA_HAS_DDIV): Likewise.
2586 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2588 * doc/invoke.texi (fabi-version): Correct number of occurrences.
2590 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2592 * doc/invoke.texi (fabi-version): Spelling fix.
2594 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2596 PR c++/70182
2597 * doc/invoke.texi (fabi-version): Mention mangling fix for
2598 operator names.
2600 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2602 PR c++/77489
2603 * doc/invoke.texi (fabi-version): Document discriminator mangling.
2605 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
2607 PR target/78875
2608 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
2609 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
2610 the new options.
2611 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
2612 flexible settings.
2613 (stack_protect_test): Ditto.
2614 * config/rs6000/rs6000.opt (mstack-protector-guard=,
2615 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
2616 options.
2617 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
2618 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
2619 -mstack-protector-guard-offset=.
2620 (RS/6000 and PowerPC Options): Ditto.
2622 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2624 * config/i386/i386.h (MASK_CLASS_P): New define.
2625 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
2626 there are no registers from different register sets also when
2627 mask registers are used. Update function comment.
2628 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
2629 to (*k/*r) and (*k/*km) alternatives.
2631 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
2633 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
2634 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
2635 (EH_RETURN_HANDLER_RTX): New define.
2636 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
2637 Force frame pointer in EH return functions.
2638 (aarch64_expand_epilogue): Add barrier for eh_return.
2639 (aarch64_final_eh_return_addr): Remove.
2640 (aarch64_eh_return_handler_rtx): New function.
2641 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
2642 Remove.
2643 (aarch64_eh_return_handler_rtx): New prototype.
2645 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2647 * config/rs6000/altivec.h (vec_rlmi): New #define.
2648 (vec_vrlnm): Likewise.
2649 (vec_rlnm): Likewise.
2650 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
2651 (UNSPEC_VRLNM): Likewise.
2652 (VIlong): New mode iterator.
2653 (altivec_vrl<VI_char>mi): New define_insn.
2654 (altivec_vrl<VI_char>nm): Likewise.
2655 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
2656 function entry.
2657 (VRLDNM): Likewise.
2658 (RLNM): New polymorphic function entry.
2659 (VRLWMI): New monomorphic function entry.
2660 (VRLDMI): Likewise.
2661 (RLMI): New polymorphic function entry.
2662 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
2663 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
2664 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
2665 vec_vrlnm.
2667 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2669 PR debug/78839
2670 * dwarf2out.c (field_byte_offset): Restore the
2671 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
2672 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
2673 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
2674 of build2 + fold.
2676 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2678 PR ada/67205
2679 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
2681 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2683 PR debug/71669
2684 * dwarf2out.c (add_data_member_location_attribute): For constant
2685 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
2686 instead of DW_AT_data_member_location, DW_AT_bit_offset and
2687 DW_AT_byte_size attributes.
2689 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2691 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
2692 after forcing to constant memory when the code model is medium.
2694 2017-01-17 Julia Koval <julia.koval@intel.com>
2696 PR target/76731
2697 * config/i386/avx512fintrin.h
2698 (_mm512_i32gather_ps): Change __addr type to void const*.
2699 (_mm512_mask_i32gather_ps): Ditto.
2700 (_mm512_i32gather_pd): Ditto.
2701 (_mm512_mask_i32gather_pd): Ditto.
2702 (_mm512_i64gather_ps): Ditto.
2703 (_mm512_mask_i64gather_ps): Ditto.
2704 (_mm512_i64gather_pd): Ditto.
2705 (_mm512_mask_i64gather_pd): Ditto.
2706 (_mm512_i32gather_epi32): Ditto.
2707 (_mm512_mask_i32gather_epi32): Ditto.
2708 (_mm512_i32gather_epi64): Ditto.
2709 (_mm512_mask_i32gather_epi64): Ditto.
2710 (_mm512_i64gather_epi32): Ditto.
2711 (_mm512_mask_i64gather_epi32): Ditto.
2712 (_mm512_i64gather_epi64): Ditto.
2713 (_mm512_mask_i64gather_epi64): Ditto.
2714 (_mm512_i32scatter_ps): Change __addr type to void*.
2715 (_mm512_mask_i32scatter_ps): Ditto.
2716 (_mm512_i32scatter_pd): Ditto.
2717 (_mm512_mask_i32scatter_pd): Ditto.
2718 (_mm512_i64scatter_ps): Ditto.
2719 (_mm512_mask_i64scatter_ps): Ditto.
2720 (_mm512_i64scatter_pd): Ditto.
2721 (_mm512_mask_i64scatter_pd): Ditto.
2722 (_mm512_i32scatter_epi32): Ditto.
2723 (_mm512_mask_i32scatter_epi32): Ditto.
2724 (_mm512_i32scatter_epi64): Ditto.
2725 (_mm512_mask_i32scatter_epi64): Ditto.
2726 (_mm512_i64scatter_epi32): Ditto.
2727 (_mm512_mask_i64scatter_epi32): Ditto.
2728 (_mm512_i64scatter_epi64): Ditto.
2729 (_mm512_mask_i64scatter_epi64): Ditto.
2730 * config/i386/avx512pfintrin.h
2731 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
2732 (_mm512_mask_prefetch_i32gather_ps): Ditto.
2733 (_mm512_mask_prefetch_i64gather_pd): Ditto.
2734 (_mm512_mask_prefetch_i64gather_ps): Ditto.
2735 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
2736 (_mm512_prefetch_i32scatter_ps): Ditto.
2737 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2738 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2739 (_mm512_prefetch_i64scatter_pd): Ditto.
2740 (_mm512_prefetch_i64scatter_ps): Ditto.
2741 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2742 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2743 * config/i386/avx512vlintrin.h
2744 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
2745 (_mm_mmask_i32gather_ps): Ditto.
2746 (_mm256_mmask_i32gather_pd): Ditto.
2747 (_mm_mmask_i32gather_pd): Ditto.
2748 (_mm256_mmask_i64gather_ps): Ditto.
2749 (_mm_mmask_i64gather_ps): Ditto.
2750 (_mm256_mmask_i64gather_pd): Ditto.
2751 (_mm_mmask_i64gather_pd): Ditto.
2752 (_mm256_mmask_i32gather_epi32): Ditto.
2753 (_mm_mmask_i32gather_epi32): Ditto.
2754 (_mm256_mmask_i32gather_epi64): Ditto.
2755 (_mm_mmask_i32gather_epi64): Ditto.
2756 (_mm256_mmask_i64gather_epi32): Ditto.
2757 (_mm_mmask_i64gather_epi32): Ditto.
2758 (_mm256_mmask_i64gather_epi64): Ditto.
2759 (_mm_mmask_i64gather_epi64): Ditto.
2760 (_mm256_i32scatter_ps): Change __addr type to void*.
2761 (_mm256_mask_i32scatter_ps): Ditto.
2762 (_mm_i32scatter_ps): Ditto.
2763 (_mm_mask_i32scatter_ps): Ditto.
2764 (_mm256_i32scatter_pd): Ditto.
2765 (_mm256_mask_i32scatter_pd): Ditto.
2766 (_mm_i32scatter_pd): Ditto.
2767 (_mm_mask_i32scatter_pd): Ditto.
2768 (_mm256_i64scatter_ps): Ditto.
2769 (_mm256_mask_i64scatter_ps): Ditto.
2770 (_mm_i64scatter_ps): Ditto.
2771 (_mm_mask_i64scatter_ps): Ditto.
2772 (_mm256_i64scatter_pd): Ditto.
2773 (_mm256_mask_i64scatter_pd): Ditto.
2774 (_mm_i64scatter_pd): Ditto.
2775 (_mm_mask_i64scatter_pd): Ditto.
2776 (_mm256_i32scatter_epi32): Ditto.
2777 (_mm256_mask_i32scatter_epi32): Ditto.
2778 (_mm_i32scatter_epi32): Ditto.
2779 (_mm_mask_i32scatter_epi32): Ditto.
2780 (_mm256_i32scatter_epi64): Ditto.
2781 (_mm256_mask_i32scatter_epi64): Ditto.
2782 (_mm_i32scatter_epi64): Ditto.
2783 (_mm_mask_i32scatter_epi64): Ditto.
2784 (_mm256_i64scatter_epi32): Ditto.
2785 (_mm256_mask_i64scatter_epi32): Ditto.
2786 (_mm_i64scatter_epi32): Ditto.
2787 (_mm_mask_i64scatter_epi32): Ditto.
2788 (_mm256_i64scatter_epi64): Ditto.
2789 (_mm256_mask_i64scatter_epi64): Ditto.
2790 (_mm_i64scatter_epi64): Ditto.
2791 (_mm_mask_i64scatter_epi64): Ditto.
2792 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
2793 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
2794 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
2795 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
2796 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
2797 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
2798 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
2799 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
2800 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
2801 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
2802 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
2803 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
2804 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
2805 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
2806 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
2807 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
2808 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
2809 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
2810 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
2811 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
2812 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
2813 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
2814 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
2815 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
2816 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
2817 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
2818 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
2819 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
2820 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
2821 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
2822 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
2823 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
2824 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
2825 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
2826 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
2827 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
2828 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
2829 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
2830 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
2831 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
2832 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
2833 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
2834 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
2835 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
2836 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
2837 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
2838 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
2839 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
2840 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
2841 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
2842 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
2843 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
2844 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
2845 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
2846 definitions accordingly.
2848 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
2849 Kuan-Lin Chen <kuanlinchentw@gmail.com>
2851 PR target/79079
2852 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
2853 gen_lowpart.
2855 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
2857 PR target/79058
2858 * ira-conflicts.c (ira_build_conflicts): Update total conflict
2859 hard regs for inner regno.
2861 2017-01-17 Martin Liska <mliska@suse.cz>
2863 PR ipa/71207
2864 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
2865 assumption and add comment.
2867 2017-01-17 Nathan Sidwell <nathan@acm.org>
2869 * ipa-visibility.c (localize_node): New function, broken out of ...
2870 (function_and_variable_visibility): ... here. Call it.
2872 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2874 PR middle-end/77445
2875 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
2876 correctly set frequency of oudgoing edge.
2877 (duplicate_thread_path): Fix profile updating.
2879 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2881 PR other/79046
2882 * configure.ac: Add GCC_BASE_VER.
2883 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
2884 version from BASE-VER file.
2885 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
2886 (gcc.o): Depend on $(BASEVER).
2887 * common.opt (dumpfullversion): New option.
2888 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
2889 * doc/invoke.texi: Document -dumpfullversion.
2890 * doc/install.texi: Document --with-gcc-major-version-only.
2891 * configure: Regenerated.
2893 2017-01-17 Richard Biener <rguenther@suse.de>
2895 PR tree-optimization/71433
2896 * tree-vrp.c (register_new_assert_for): Merge same asserts
2897 on all incoming edges.
2898 (process_assert_insertions_for): Handle insertions at the
2899 beginning of BBs.
2901 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
2903 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
2904 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
2906 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
2908 PR target/78633
2909 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
2910 RTL sharing.
2912 2017-01-17 Alan Modra <amodra@gmail.com>
2914 PR target/79066
2915 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
2916 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
2917 symbolic stack limit when pic.
2919 2017-01-16 Martin Sebor <msebor@redhat.com>
2921 PR tree-optimization/78608
2922 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
2924 2017-01-16 Jeff Law <law@redhat.com>
2926 Revert:
2927 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
2928 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
2929 for several include directories that may be relative to sysroot.
2930 * config/i386/x-mingw32 (gplus_includedir): Define.
2931 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
2932 (native_system_includedir): Likewise.
2933 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
2934 override if TARGET_SYSTEM_ROOT is defined.
2935 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
2937 PR tree-optimization/79090
2938 PR tree-optimization/33562
2939 PR tree-optimization/61912
2940 PR tree-optimization/77485
2941 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
2942 and computed trims into the dump file.
2944 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2946 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
2948 2017-01-16 Jakub Jelinek <jakub@redhat.com>
2950 PR c/79089
2951 * gimplify.c (gimplify_init_constructor): If want_value and
2952 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
2953 fix.
2955 PR target/79080
2956 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
2957 sequence. Formatting fixes.
2958 (doloop_optimize): Formatting fixes.
2960 PR driver/49726
2961 * gcc.c (debug_level_greater_than_spec_func): New function.
2962 (static_spec_functions): Add debug-level-gt spec function.
2963 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2964 !g0.
2965 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2966 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2967 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2968 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2969 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2970 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2972 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2974 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2975 QImode fixups to general and mask registers only.
2977 2017-01-16 Carl Love <cel@us.ibm.com>
2979 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2980 for built-in functions
2981 vector signed char vec_nabs (vector signed char)
2982 vector signed short vec_nabs (vector signed short)
2983 vector signed int vec_nabs (vector signed int)
2984 vector signed long long vec_nabs (vector signed long long)
2985 vector float vec_nabs (vector float)
2986 vector double vec_nabs (vector double)
2987 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2988 and NABS overload.
2989 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2990 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2991 * doc/extend.texi: Update the documentation file for the new built-in
2992 functions.
2994 2017-01-16 Martin Sebor <msebor@redhat.com>
2996 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2997 message.
2999 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3001 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
3002 UNSPEC_VSX__XXSPLTD to require special splat handling.
3004 2017-01-16 David Malcolm <dmalcolm@redhat.com>
3006 PR bootstrap/78616
3007 * system.h: Poison strndup.
3009 2017-01-16 Alan Modra <amodra@gmail.com>
3011 PR target/79098
3012 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
3013 use a switch.
3015 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
3017 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
3019 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
3021 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
3022 call recog here. Assert that INSN_CODE (insn) is non-negative.
3024 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
3026 PR target/72749
3027 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
3028 fallthrough.
3029 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
3030 in the currently scheduled RTL fragment.
3032 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
3034 PR rtl-optimization/78751
3035 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
3036 give up.
3038 2017-01-14 Jeff Law <law@redhat.com>
3040 PR tree-optimization/79090
3041 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
3042 variable length stores.
3043 (compute_trims): Delete dead assignment to *trim_tail.
3044 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
3045 zero length.
3047 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
3049 PR rtl-optimization/78626
3050 PR rtl-optimization/78727
3051 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
3052 of a block, and split such blocks after everything else is finished.
3054 2017-01-14 Alan Modra <amodra@gmail.com>
3056 PR target/72749
3057 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
3058 target legitimate_combined_insn.
3059 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
3060 (rs6000_legitimate_combined_insn): New function.
3061 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
3062 all uses.
3063 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
3064 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
3065 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
3067 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
3069 * doc/frontends.texi (G++ and GCC): Remove references to Java.
3071 2017-01-13 Jeff Law <law@redhat.com>
3073 PR tree-optimization/33562
3074 PR tree-optimization/61912
3075 PR tree-optimization/77485
3076 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
3077 a statement.
3078 (delete_dead_assignment): Likewise.
3079 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
3080 statement to delete_dead_call and delete_dead_assignment.
3082 2017-01-13 David Malcolm <dmalcolm@redhat.com>
3084 PR c/78304
3085 * substring-locations.c (format_warning_va): Strengthen case 1 so
3086 that both endpoints of the substring must be within the format
3087 range for just the substring to be printed.
3089 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
3091 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
3092 * config/i386/i386.c (ix86_target_string): Add missing options
3093 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
3094 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
3095 flags_other and ix86_target_other to flags2_other. Display unknown
3096 isa2 options.
3097 (ix86_valid_target_attribute_inner_p): Add missing options and
3098 reorder options by implied ISAs, as in ix86_target_string.
3100 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3102 * hash-table.h (hash_table::too_empty_p): New function.
3103 (hash_table::expand): Use it.
3104 (hash_table::traverse): Likewise.
3105 (hash_table::empty_slot): Use sizeof (value_type) instead of
3106 sizeof (PTR) to convert bytes to elements. Shrink the table
3107 if the current size is excessive for the current number of
3108 elements.
3110 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
3112 * ira-costs.c (record_reg_classes): Break from the inner loop
3113 early once alt_fail is known to be true. Update outer loop
3114 handling accordingly.
3116 2017-01-13 Jeff Law <law@redhat.com>
3118 * tree-ssa-dse.c (decrement_count): New function.
3119 (increment_start_addr, maybe_trim_memstar_call): Likewise.
3120 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
3121 when we know the partially dead statement is a mem* function.
3123 PR tree-optimization/61912
3124 PR tree-optimization/77485
3125 * tree-ssa-dse.c: Include expr.h.
3126 (maybe_trim_constructor_store): New function.
3127 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
3129 PR tree-optimization/33562
3130 PR tree-optimization/61912
3131 PR tree-optimization/77485
3132 * doc/invoke.texi: Document new dse-max-object-size param.
3133 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
3134 * tree-ssa-dse.c: Include params.h.
3135 (dse_store_status): New enum.
3136 (initialize_ao_ref_for_dse): New, partially extracted from
3137 dse_optimize_stmt.
3138 (valid_ao_ref_for_dse, normalize_ref): New.
3139 (setup_live_bytes_from_ref, compute_trims): Likewise.
3140 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
3141 (maybe_trim_partially_dead_store): Likewise.
3142 (maybe_trim_complex_store): Likewise.
3143 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
3144 Track what bytes live from the original store. Return tri-state
3145 for dead, partially dead or live.
3146 (dse_dom_walker): Add constructor, destructor and new private members.
3147 (delete_dead_call, delete_dead_assignment): New extracted from
3148 dse_optimize_stmt.
3149 (dse_optimize_stmt): Make a member of dse_dom_walker.
3150 Use initialize_ao_ref_for_dse.
3152 PR tree-optimization/33562
3153 PR tree-optimization/61912
3154 PR tree-optimization/77485
3155 * sbitmap.h (bitmap_count_bits): Prototype.
3156 (bitmap_clear_range, bitmap_set_range): Likewise.
3157 * sbitmap.c (bitmap_clear_range): New function.
3158 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
3160 2017-01-13 Martin Liska <mliska@suse.cz>
3162 PR ipa/79043
3163 * function.c (set_cfun): Add new argument force.
3164 * function.h (set_cfun): Likewise.
3165 * ipa-inline-transform.c (inline_call): Use the function when
3166 strict alising from is dropped for function we inline to.
3168 2017-01-13 Richard Biener <rguenther@suse.de>
3170 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
3171 for dumping GIMPLE INTEGER_CSTs.
3173 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3175 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
3176 to 201112L since C++17.
3178 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
3180 PR sanitizer/78887
3181 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
3182 if -fsanitize=kernel-address is present.
3184 2017-01-13 Richard Biener <rguenther@suse.de>
3186 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
3187 as _Literal ( type ) number in case usual suffixes do not
3188 preserve all information.
3190 2017-01-13 Richard Biener <rguenther@suse.de>
3192 PR tree-optimization/77283
3193 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
3194 and ssa-iterators.h.
3195 (is_feasible_trace): Implement a cost model based on joiner
3196 PHI node uses.
3198 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3200 PR target/79004
3201 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
3202 char or short to __float128/_Float128 directly.
3204 2017-01-12 Martin Sebor <msebor@redhat.com>
3206 to -Wformat-overflow.
3207 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
3208 (min_bytes_remaining): Same.
3209 (get_string_length): Same.
3210 (format_string): Same.
3211 (format_directive): Same.
3212 (add_bytes): Same.
3213 (pass_sprintf_length::handle_gimple_call): Same.
3215 2017-01-12 Jakub Jelinek <jakub@redhat.com>
3217 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
3218 info.nowrite calls with no lhs that can't throw. Return bool
3219 whether gsi_remove has been called or not.
3220 (pass_sprintf_length::handle_gimple_call): Return bool whether
3221 try_substitute_return_value called gsi_remove. Formatting fix.
3222 (pass_sprintf_length::execute): Don't use gsi_remove if
3223 handle_gimple_call returned true.
3225 PR bootstrap/79069
3226 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
3227 be removed due to side-effects, don't remove following barrier nor
3228 turn the successor edge into fallthru edge.
3230 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3232 PR target/79044
3233 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
3234 element-reversing loads and stores as not swappable.
3236 2017-01-12 Nathan Sidwell <nathan@acm.org>
3237 Nicolai Stange <nicstange@gmail.com>
3239 * combine.c (try_combine): Don't ignore result of overlap checking
3240 loop. Combine overlap & asm check into single loop.
3242 2017-01-12 Richard Biener <rguenther@suse.de>
3244 * tree-pretty-print.c (dump_generic_node): Provide -gimple
3245 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
3247 2017-01-12 Richard Biener <rguenther@suse.de>
3249 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
3250 and TS_TARGET_OPTION directly derive from TS_BASE.
3251 * tree-core.h (tree_optimization_option): Derive from tree_base.
3252 (tree_target_option): Likewise.
3254 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
3256 * config/i386/i386.c (memory_address_length): Increase len
3257 only when rip_relative_addr_p returns false.
3259 2017-01-11 Julia Koval <julia.koval@intel.com>
3261 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
3262 (OPTION_MASK_ISA_SGX_SET): New.
3263 (ix86_handle_option): Handle OPT_msgx.
3264 * config.gcc: Added sgxintrin.h.
3265 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
3266 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
3267 * config/i386/i386.c (ix86_target_string): Add -msgx.
3268 (PTA_SGX): New.
3269 (ix86_option_override_internal): Handle new options.
3270 (ix86_valid_target_attribute_inner_p): Add sgx.
3271 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
3272 * config/i386/i386.opt: Add msgx.
3273 * config/i386/sgxintrin.h: New file.
3274 * config/i386/x86intrin.h: Add sgxintrin.h.
3276 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3278 PR c++/71537
3279 * fold-const.c (maybe_nonzero_address): Return 1 for function
3280 local objects.
3281 (tree_single_nonzero_warnv_p): Don't handle function local objects
3282 here.
3284 PR c++/72813
3285 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
3286 of c-header.
3288 2017-01-11 David Malcolm <dmalcolm@redhat.com>
3290 PR driver/78877
3291 * opts.c: Include "spellcheck.h"
3292 (struct string_fragment): New struct.
3293 (struct edit_distance_traits<const string_fragment &>): New
3294 struct.
3295 (get_closest_sanitizer_option): New function.
3296 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
3298 2017-01-11 Jakub Jelinek <jakub@redhat.com>
3300 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
3301 by 12.
3302 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
3303 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
3304 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
3305 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
3306 for initial die_offset if dwarf_split_debug_info.
3307 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
3308 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
3309 fields.
3310 (output_skeleton_debug_sections): Formatting fix. Use
3311 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
3312 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
3314 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
3316 * config/arm/cortex-a53.md: Add bypasses for
3317 cortex_a53_r2f_cvt.
3318 (cortex_a53_r2f): Only use for transfers.
3319 (cortex_a53_f2r): Likewise.
3320 (cortex_a53_r2f_cvt): Add reservation for conversions.
3321 (cortex_a53_f2r_cvt): Likewise.
3323 2017-01-11 Tamar Christina <tamar.christina@arm.com>
3325 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
3326 to all inlined functions, change static to extern.
3328 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
3330 PR target/78253
3331 * config/arm/arm.c (legitimize_pic_address): Handle reference to
3332 weak symbol.
3333 (arm_assemble_integer): Likewise.
3335 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
3337 * config.gcc: Use new awk script to check CPU, FPU and architecture
3338 parameters for --with-... options.
3339 * config/arm/parsecpu.awk: New file
3340 * config/arm/arm-cpus.in: New file.
3341 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
3342 files.
3343 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
3344 files.
3345 * config/arm/t-arm: Update dependency rules.
3346 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
3347 of processing .def files.
3348 * config/arm/genopt.sh: Deleted.
3349 * config/arm/gentune.sh: Deleted.
3350 * config/arm/arm-cores.def: Deleted.
3351 * config/arm/arm-arches.def: Deleted.
3352 * config/arm/arm-fpus.def: Deleted.
3353 * config/arm/arm-tune.md: Regenerated.
3354 * config/arm/arm-tables.opt: Regenerated.
3355 * config/arm/arm-cpu.h: New generated file.
3356 * config/arm/arm-cpu-data.h: New generated file.
3357 * config/arm/arm-cpu-cdata.h: New generated file.
3359 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
3361 PR lto/79042
3362 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
3363 bit.
3364 (input_varpool_node): Unpack dynamically_initialized bit.
3366 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
3368 PR rtl-optimization/79032
3369 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
3370 the alignment of the adjusted memory reference against that of MODE,
3371 instead of the alignment of the original memory reference.
3373 2017-01-11 Martin Jambor <mjambor@suse.cz>
3375 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
3376 test.
3377 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
3378 decorated functions.
3380 2017-01-11 Richard Biener <rguenther@suse.de>
3382 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
3383 set range/nonnull info for PHI results. Do not set it on
3384 stmts marked for removal.
3386 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
3388 * expr.c (store_field): In the bitfield case, fetch the return value
3389 from the registers before applying a single big-endian adjustment.
3390 Always do a final load for a BLKmode value not larger than a word.
3392 2017-01-10 David Malcolm <dmalcolm@redhat.com>
3394 PR c++/77949
3395 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3396 that we correctly handle column numbers greater than
3397 LINE_MAP_MAX_COLUMN_NUMBER.
3399 2017-01-10 Martin Sebor <msebor@redhat.com>
3401 PR middle-end/78245
3402 * gimple-ssa-sprintf.c (get_destination_size): Call
3403 {init,fini}object_sizes.
3404 * tree-object-size.c (addr_object_size): Adjust.
3405 (pass_through_call): Adjust.
3406 (pass_object_sizes::execute): Adjust.
3407 * tree-object-size.h (fini_object_sizes): Declare.
3409 2017-01-10 Martin Sebor <msebor@redhat.com>
3411 PR tree-optimization/78775
3412 * builtins.c (get_size_range): Move...
3413 * calls.c: ...to here.
3414 (alloc_max_size): Accept zero argument.
3415 (operand_signed_p): Remove.
3416 (maybe_warn_alloc_args_overflow): Call get_size_range.
3417 * calls.h (get_size_range): Declare.
3419 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
3421 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
3422 from TI's devices.csv file as of September 2016.
3423 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
3425 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
3427 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
3428 * doc/invoke.texi: Likewise.
3429 * doc/md.texi: Likewise.
3430 * doc/objc.texi: Likewise.
3432 2017-01-10 Joshua Conner <joshconner@google.com>
3434 * config/arm/fuchsia-elf.h: New file.
3435 * config/fuchsia.h: New file.
3436 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
3437 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
3438 targets.
3439 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
3441 2016-01-10 Richard Biener <rguenther@suse.de>
3443 PR tree-optimization/79034
3444 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
3445 Propagate out degenerate PHIs in the joiner.
3447 2017-01-10 Martin Liska <mliska@suse.cz>
3449 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
3450 (sort_congruence_classes_by_decl_uid): Likewise.
3451 (sort_congruence_class_groups_by_decl_uid): Likewise.
3452 (sem_item_optimizer::merge_classes): Sort class, groups in these
3453 classes and members in the groups by DECL_UID of declarations.
3454 This would make merge operations stable.
3456 2017-01-10 Martin Liska <mliska@suse.cz>
3458 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
3459 usage of m_classes_vec.
3460 (sem_item_optimizer::~sem_item_optimizer): Likewise.
3461 (sem_item_optimizer::get_group_by_hash): Likewise.
3462 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
3463 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
3464 (sem_item_optimizer::verify_classes): Likewise.
3465 (sem_item_optimizer::process_cong_reduction): Likewise.
3466 (sem_item_optimizer::dump_cong_classes): Likewise.
3467 (sem_item_optimizer::merge_classes): Likewise.
3468 * ipa-icf.h (congruence_class_hash): Rename from
3469 congruence_class_group_hash. Remove declaration of m_classes_vec.
3471 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
3473 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
3474 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
3475 * config.gcc: Add avx512vpopcntdqintrin.h.
3476 * config/i386/avx512vpopcntdqintrin.h: New.
3477 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
3478 * config/i386/i386-builtin-types.def: Add new types.
3479 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
3480 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
3481 __builtin_ia32_vpopcountq_v8di_mask): New.
3482 * config/i386/i386-c.c (ix86_target_macros_internal): Define
3483 __AVX512VPOPCNTDQ__.
3484 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
3485 (PTA_AVX512VPOPCNTDQ): Define.
3486 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
3487 TARGET_AVX512VPOPCNTDQ_P): Define.
3488 * config/i386/i386.opt: Add mavx512vpopcntdq.
3489 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
3490 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
3492 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3494 PR middle-end/77484
3495 * predict.def (PRED_CALL): Set to 67.
3497 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
3499 * expr.c (store_field): In the bitfield case, if the value comes from
3500 a function call and is of an aggregate type returned in registers, do
3501 not modify the field mode; extract the value in all cases if the mode
3502 is BLKmode and the size is not larger than a word.
3504 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
3506 PR target/71017
3507 * config/i386/cpuid.h: Fix undefined behavior.
3509 2017-01-04 Jeff Law <law@redhat.com>
3511 PR tree-optimization/79007
3512 PR tree-optimization/67955
3513 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
3514 conservative for pt.null when flag_non_call_exceptions is on.
3516 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3518 PR translation/79019
3519 PR translation/79020
3520 * params.def (PARAM_INLINE_MIN_SPEEDUP,
3521 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
3522 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
3523 in descriptions.
3524 * config/avr/avr.opt (maccumulate-args): Likewise.
3525 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
3526 * common.opt (freport-bug): Likewise.
3527 * cif-code.def (CIF_FINAL_ERROR): Likewise.
3528 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
3529 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
3530 translatable string.
3531 * config/i386/i386.c (function_value_32): Likewise.
3532 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
3533 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
3534 Likewise.
3535 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
3536 * common/config/msp430/msp430-common.c (msp430_handle_option):
3537 Likewise.
3538 * symtab.c (symtab_node::verify_base): Likewise.
3539 * opts.c (set_debug_level): Likewise.
3540 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
3541 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
3542 missing whitespace to translatable strings.
3543 * config/avr/avr.md (bswapsi2): Fix typo in comment.
3544 * config/sh/superh.h: Likewise.
3545 * config/i386/xopintrin.h: Likewise.
3546 * config/i386/znver1.md: Likewise.
3547 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
3548 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
3549 * double-int.h (struct double_int): Likewise.
3550 * double-int.c (div_and_round_double): Likewise.
3551 * wide-int.cc: Likewise.
3552 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3553 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
3554 * cfgcleanup.c (crossjumps_occured): Renamed to ...
3555 (crossjumps_occurred): ... this.
3556 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
3557 Adjust all uses.
3559 PR tree-optimization/78899
3560 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
3561 returning bool return struct loop *, NULL for failure and the new
3562 loop on success.
3563 (versionable_outer_loop_p): Don't version outer loop if it has
3564 dont_vectorized bit set.
3565 (tree_if_conversion): When versioning outer loop, ensure
3566 tree_if_conversion is performed also on the inner loop of the
3567 non-vectorizable outer loop copy.
3568 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
3569 LOOP_VECTORIZED in inner loop of the scalar outer loop and
3570 prevent vectorization of it.
3571 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
3572 the outer loop vectorization of the non-scalar version is attempted
3573 before vectorization of the inner loop in scalar version. If
3574 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
3575 vectorization of its inner loop.
3576 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
3577 has 2 inner loops, rename also on edges from bb whose single pred
3578 is outer_loop->header. Fix typo in function comment.
3580 2017-01-09 Martin Sebor <msebor@redhat.com>
3582 PR bootstrap/79033
3583 * asan.c (asan_emit_stack_protection): Increase local buffer size
3584 to avoid snprintf truncation warning.
3586 2017-01-09 Andrew Pinski <apinski@cavium.com>
3588 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
3589 to reference thunderx2t99 for the tuning structure
3590 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
3591 Rename to ...
3592 (thunderx2t99_extra_costs): This.
3593 * config/aarch64/aarch64-tune.md: Regenerate.
3594 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
3595 (vulcan_addrcost_table): This.
3596 (vulcan_regmove_cost): Rename to ...
3597 (thunderx2t99_regmove_cost): This.
3598 (vulcan_vector_cost): Rename to ...
3599 (thunderx2t99_vector_cost): this.
3600 (vulcan_branch_cost): Rename to ...
3601 (thunderx2t99_branch_cost): This.
3602 (vulcan_tunings): Rename to ...
3603 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
3604 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
3606 2017-01-09 Martin Jambor <mjambor@suse.cz>
3608 PR ipa/78365
3609 PR ipa/78599
3610 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
3611 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
3612 (propagate_vr_accross_jump_function): Use the above function for all
3613 value range computations for pass-through jump functions and type
3614 converasion from explicit value range values.
3615 (ipcp_propagate_stage): Do not attempt to deduce types of formal
3616 parameters from TYPE_ARG_TYPES.
3617 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
3618 (ipa_write_node_info): Stream type of the actual argument.
3619 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
3621 2017-01-09 Martin Liska <mliska@suse.cz>
3623 PR pch/78970
3624 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
3625 (lookup_compiler): Do not show error message with have_E.
3627 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3629 PR tree-optimization/78938
3630 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
3631 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
3632 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
3633 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
3634 fixes.
3636 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3638 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
3639 is const0_rtx.
3641 2017-01-09 Richard Biener <rguenther@suse.de>
3643 PR tree-optimization/78997
3644 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
3645 name condition properly.
3647 2017-01-09 Richard Biener <rguenther@suse.de>
3649 PR debug/79000
3650 * dwarf2out.c (is_cxx): New overload with context.
3651 (is_naming_typedef_decl): Use it.
3653 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3655 * invoke.texi (Option Summary): Correct spacing in option lists
3656 and add line breaks to fix over-long lines.
3658 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3660 PR middle-end/17660
3662 * extend.texi (Common Variable Attributes): Add xref to GCC
3663 Internals manual to explain mode attribute keywords.
3665 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3667 PR other/16519
3668 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
3669 and Preprocessor Options.
3670 (Options for Linking): Document -pthread here....
3671 (RS/6000 and PowerPC Options): ...not here.
3672 (Solaris 2 Options): ...or here.
3673 * doc/cppopts.texi: Document -pthread.
3675 2017-01-08 Martin Sebor <msebor@redhat.com>
3677 PR middle-end/77708
3678 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
3679 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
3680 New member functions.
3681 (format_directive): Used them.
3682 (add_bytes): Same.
3683 (pass_sprintf_length::handle_gimple_call): Same.
3684 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
3685 to avoid truncation for any argument.
3686 (extract_affine_mul): Same.
3687 * tree.c (get_file_function_name): Same.
3689 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3691 PR middle-end/77484
3692 * predict.def (PRED_INDIR_CALL): Set to 86.
3694 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3696 PR preprocessor/54124
3697 * doc/cppopts.texi: Reformat -d subtable to list the full name
3698 of the options. Add cross-reference to the docs for the general
3699 compiler -d options.
3700 * doc/invoke.texi (Developer Options): Add cross-reference to the
3701 preprocessor-specific -d option documentation.
3703 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3705 PR preprocessor/13498
3706 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
3707 redudant material, and reflect new command-line options.
3708 (System Headers): Likewise.
3710 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3712 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
3713 -isystem, and -idirafter. Copy-edit.
3714 * doc/cppopts.texi: Copy-edit. Remove contradiction about
3715 default for -ftrack-macro-expansion. Delete obsolete and
3716 badly-formatted implementation details about -fdebug-cpp output.
3717 * doc/cppwarnopts.texi: Copy-edit.
3719 2017-01-07 David Malcolm <dmalcolm@redhat.com>
3721 PR c++/72803
3722 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3723 that the transition from a max line width >= 1<<10 to narrower
3724 lines works correctly.
3726 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
3728 * doc/options.texi (PerFunction): New.
3729 * opt-functions.awk (switch_flags): Map both Optimization and
3730 PerFunction to CL_OPTIMIZATION.
3731 * opth-gen.awk: Test for PerFunction flag along with
3732 Optimization.
3733 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
3734 it only when the latter is present. Skip those that don't in
3735 the hash function generator.
3736 * common.opt (fvar-tracking): Mark as PerFunction instead of
3737 Optimization.
3738 (fvar-tracking-assignments): Likewise.
3739 (fvar-tracking-assignments-toggle): Likewise.
3740 (fvar-tracking-uninit): Likewise.
3742 2017-01-07 Jakub Jelinek <jakub@redhat.com>
3744 PR translation/79018
3745 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
3746 the and store.
3748 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
3750 PR target/57583
3751 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
3752 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
3753 TARGET_LONG_JUMP_TABLE_OFFSETS.
3754 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
3755 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
3756 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
3757 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
3758 * config/m68k/m68k.md (tablejump expander): Likewise.
3759 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
3760 TARGET_LONG_JUMP_TABLE_OFFSETS.
3761 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
3762 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
3764 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
3765 David Holsgrove <david.holsgrove@xilinx.com>
3767 * common/config/microblaze/microblaze-common.c
3768 (TARGET_EXCEPT_UNWIND_INFO): Remove.
3769 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
3770 New prototype.
3771 * config/microblaze/microblaze.c (microblaze_must_save_register)
3772 (microblaze_expand_epilogue, microblaze_return_addr): Handle
3773 calls_eh_return.
3774 (microblaze_eh_return): New function.
3775 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
3776 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
3777 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
3778 * config/microblaze/microblaze.md (eh_return): New pattern.
3780 2017-01-06 Jakub Jelinek <jakub@redhat.com>
3782 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
3783 GCC_DIAGNOSTIC_STRINGIFY): Define.
3785 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
3787 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3789 * config/arm/arm.md (<mcrr>): New.
3790 (<mrrc>): New.
3791 * config/arm/arm.c (arm_arch5te): New.
3792 (arm_option_override): Set arm_arch5te.
3793 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
3794 and mrrc2.
3795 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
3796 (arm_mcrr_qualifiers): ... this. New.
3797 (MRRC_QUALIFIERS): Define to...
3798 (arm_mrrc_qualifiers): ... this. New.
3799 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
3800 __arm_mrrc2): New.
3801 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
3802 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
3803 (MRRCI, mrrc, MRRC): New.
3804 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
3805 VUNSPEC_MRRC2): New.
3807 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3809 * config/arm/arm.md (<mcr>): New.
3810 (<mrc>): New.
3811 * config/arm/arm.c (arm_coproc_builtin_available): Add
3812 support for mcr, mrc, mcr2 and mrc2.
3813 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
3814 (arm_mcr_qualifiers): ... this. New.
3815 (MRC_QUALIFIERS): Define to ...
3816 (arm_mrc_qualifiers): ... this. New.
3817 (MCR_QUALIFIERS): Define to ...
3818 (arm_mcr_qualifiers): ... this. New.
3819 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
3820 __arm_mrc2): New.
3821 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
3822 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
3823 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
3824 VUNSPEC_MRC2): New.
3826 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3828 * config/arm/arm.md (*ldc): New.
3829 (*stc): New.
3830 (<ldc>): New.
3831 (<stc>): New.
3832 * config/arm/arm.c (arm_coproc_builtin_available): Add
3833 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
3834 (arm_coproc_ldc_stc_legitimate_address): New.
3835 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
3836 'qualifier_const_pointer'.
3837 (LDC_QUALIFIERS): Define to...
3838 (arm_ldc_qualifiers): ... this. New.
3839 (STC_QUALIFIERS): Define to...
3840 (arm_stc_qualifiers): ... this. New.
3841 * config/arm/arm-protos.h
3842 (arm_coproc_ldc_stc_legitimate_address): New.
3843 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
3844 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
3845 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
3846 stc2, stcl, stc2l): New.
3847 * config/arm/constraints.md (Uz): New.
3848 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
3849 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
3850 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
3851 VUNSPEC_STC2L): New.
3853 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3855 * config/arm/arm.md (<cdp>): New.
3856 * config/arm/arm.c (neon_const_bounds): Rename this ...
3857 (arm_const_bounds): ... this.
3858 (arm_coproc_builtin_available): New.
3859 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
3860 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
3861 (CDP_QUALIFIERS): Define to...
3862 (arm_cdp_qualifiers): ... this. New.
3863 (void_UP): Define.
3864 (arm_expand_builtin_args): Add case for 6 arguments.
3865 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
3866 (arm_const_bounds): ... this.
3867 (arm_coproc_builtin_available): New.
3868 * config/arm/arm_acle.h (__arm_cdp): New.
3869 (__arm_cdp2): New.
3870 * config/arm/arm_acle_builtins.def (cdp): New.
3871 (cdp2): New.
3872 * config/arm/iterators.md (CDPI,CDP,cdp): New.
3873 * config/arm/neon.md: Rename all 'neon_const_bounds' to
3874 'arm_const_bounds'.
3875 * config/arm/types.md (coproc): New.
3876 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
3877 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
3878 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
3879 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
3881 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3883 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
3884 (UBINOP_QUALIFIERS): New.
3885 (si_UP): Define.
3886 (acle_builtin_data): New. Change comment.
3887 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
3888 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
3889 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
3890 arm_acle_builtins.def.
3891 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
3892 (arm_init_acle_builtins): New.
3893 (CRC32_BUILTIN): Remove.
3894 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
3895 crc32cb, crc32ch and crc32cw.
3896 (arm_init_crc32_builtins): Remove.
3897 (arm_init_builtins): Use arm_init_acle_builtins rather
3898 than arm_init_crc32_builtins.
3899 (arm_expand_acle_builtin): New.
3900 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
3901 * config/arm/arm_acle_builtins.def: New.
3903 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3905 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
3906 (arm_builtin_datum): ... this.
3907 (arm_init_neon_builtin): Rename to ...
3908 (arm_init_builtin): ... this. Add a new parameters PREFIX
3909 and USE_SIG_IN_NAME.
3910 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
3911 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
3912 'arm_builtin_datum'.
3913 (arm_init_vfp_builtins): Likewise.
3914 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
3915 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
3916 (arm_expand_neon_args): Rename to ...
3917 (arm_expand_builtin_args): ... this. Rename builtin_arg
3918 enum values and differentiate between ARG_BUILTIN_MEMORY
3919 and ARG_BUILTIN_NEON_MEMORY.
3920 (arm_expand_neon_builtin_1): Rename to ...
3921 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
3922 values, arm_expand_builtin_args and add bool parameter NEON.
3923 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
3924 (arm_expand_vfp_builtin): Likewise.
3925 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
3927 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3929 PR middle-end/77484
3930 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
3931 * predict.c (tree_estimate_probability_bb): Reverse direction of
3932 polymorphic call predictor.
3934 2017-01-06 David Malcolm <dmalcolm@redhat.com>
3936 * passes.c (execute_one_pass): Split out pass-skipping logic into...
3937 (determine_pass_name_match): ...this new function and...
3938 (should_skip_pass_p): ...this new function.
3940 2017-01-06 Nathan Sidwell <nathan@acm.org>
3942 * ipa-visibility.c (function_and_variable_visibility): Reformat
3943 comments and long lines. Remove extrneous if.
3944 * symtab.c (symtab_node::make_decl_local): Fix code format.
3945 (symtab_node::set_section_for_node): Fix comment typo.
3947 2017-01-06 Martin Liska <mliska@suse.cz>
3949 PR bootstrap/79003
3950 * lra-constraints.c: Rename invariant to lra_invariant.
3951 * predict.c (set_even_probabilities): Initialize e to NULL.
3953 2017-01-05 Martin Sebor <msebor@redhat.com>
3955 PR tree-optimization/78910
3956 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
3957 (format_integer): Correct off-by-one error in the handling
3958 of precision with negative numbers in signed conversions..
3960 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3962 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3964 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3966 PR tree-optimization/71016
3967 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3968 factor_out_conditional_conversion. Formatting fix.
3969 (factor_out_conditional_conversion): Add cond_stmt argument.
3970 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3971 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3972 Formatting fix.
3974 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3976 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3977 read-rtl-function.o, and selftest-rtl.o.
3978 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3979 (selftest::aarch64_test_loading_full_dump): New function.
3980 (selftest::aarch64_run_selftests): New function.
3981 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3982 selftest::aarch64_run_selftests.
3983 * config/i386/i386.c
3984 (selftest::ix86_test_loading_dump_fragment_1): New function.
3985 (selftest::ix86_test_loading_call_insn): New function.
3986 (selftest::ix86_test_loading_full_dump): New function.
3987 (selftest::ix86_test_loading_unspec): New function.
3988 (selftest::ix86_run_selftests): Call the new functions.
3989 * emit-rtl.c (maybe_set_max_label_num): New function.
3990 * emit-rtl.h (maybe_set_max_label_num): New decl.
3991 * function.c (instantiate_decls): Guard call to
3992 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3993 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3994 "static".
3995 * gensupport.c (gen_reader::gen_reader): Pass "false"
3996 for new "compact" param of rtx_reader.
3997 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3998 rather than an empty string for NULL strings.
3999 * read-md.c: Potentially include config.h rather than bconfig.h.
4000 Wrap include of errors.h with #ifdef GENERATOR_FILE.
4001 (have_error): New global, copied from errors.c.
4002 (md_reader::read_name): Rename to...
4003 (md_reader::read_name_1): ...this, adding "out_loc" param,
4004 and converting "missing name or number" to returning false, rather
4005 than failing.
4006 (md_reader::read_name): Reimplement in terms of read_name_1.
4007 (md_reader::read_name_or_nil): New function.
4008 (md_reader::read_string): Handle "(nil)" by returning NULL.
4009 (md_reader::md_reader): Add new param "compact".
4010 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
4011 (md_reader::read_file): New method.
4012 * read-md.h (md_reader::md_reader): Add new param "compact".
4013 (md_reader::read_file): New method.
4014 (md_reader::is_compact): New accessor.
4015 (md_reader::read_name): Convert return type from void to file_location.
4016 (md_reader::read_name_or_nil): New decl.
4017 (md_reader::read_name_1): New decl.
4018 (md_reader::m_compact): New field.
4019 (noop_reader::noop_reader): Pass "false" for new "compact" param
4020 of rtx_reader.
4021 (rtx_reader::rtx_reader): Add new "compact" param.
4022 (rtx_reader::read_rtx_operand): Make virtual and convert return
4023 type from void to rtx.
4024 (rtx_reader::read_until): New decl.
4025 (rtx_reader::handle_any_trailing_information): New virtual function.
4026 (rtx_reader::postprocess): New virtual function.
4027 (rtx_reader::finalize_string): New virtual function.
4028 (rtx_reader::m_in_call_function_usage): New field.
4029 (rtx_reader::m_reuse_rtx_by_id): New field.
4030 * read-rtl-function.c: New file.
4031 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
4032 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
4033 (selftest::verify_three_block_rtl_cfg): New decl.
4034 * read-rtl-function.h: New file.
4035 * read-rtl.c: Potentially include config.h rather than bconfig.h.
4036 For host, include function.h, memmodel.h, and emit-rtl.h.
4037 (one_time_initialization): New function.
4038 (struct compact_insn_name): New struct.
4039 (compact_insn_names): New array.
4040 (find_code): Handle insn codes in compact dumps.
4041 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
4042 (bind_subst_iter_and_attr): Likewise.
4043 (add_condition_to_string): Likewise.
4044 (add_condition_to_rtx): Likewise.
4045 (apply_attribute_uses): Likewise.
4046 (add_current_iterators): Likewise.
4047 (apply_iterators): Likewise.
4048 (initialize_iterators): Guard usage of apply_subst_iterator with
4049 #ifdef GENERATOR_FILE.
4050 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
4051 (md_reader::read_mapping): Likewise.
4052 (add_define_attr_for_define_subst): Likewise.
4053 (add_define_subst_attr): Likewise.
4054 (read_subst_mapping): Likewise.
4055 (check_code_iterator): Likewise.
4056 (rtx_reader::read_rtx): Likewise. Move one-time initialization
4057 logic to...
4058 (one_time_initialization): New function.
4059 (rtx_reader::read_until): New method.
4060 (read_flags): New function.
4061 (parse_reg_note_name): New function.
4062 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
4063 Handle reuse_rtx ids.
4064 Wrap iterator lookup within #ifdef GENERATOR_FILE.
4065 Add parsing support for RTL dumps, mirroring the special-cases in
4066 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
4067 values, and calling handle_any_trailing_information.
4068 (rtx_reader::read_rtx_operand): Convert return type from void
4069 to rtx, returning return_rtx. Handle case 'e'. Call
4070 finalize_string on XSTR and XTMPL fields.
4071 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
4072 "(nil)" values were omitted. Call the postprocess vfunc on the
4073 return_rtx.
4074 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
4075 class ctor. Initialize m_in_call_function_usage. Call
4076 one_time_initialization.
4077 * rtl-tests.c (selftest::test_uncond_jump): Call
4078 set_new_first_and_last_insn.
4079 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
4080 * selftest-rtl.c: New file.
4081 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
4082 (selftest::get_insn_by_uid): New decl.
4083 * selftest-run-tests.c (selftest::run_tests): Call
4084 read_rtl_function_c_tests.
4085 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
4086 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
4087 dumps.
4089 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
4091 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
4092 operands in a special way. Assert that pos+len <= mode precision.
4094 2017-01-05 Jakub Jelinek <jakub@redhat.com>
4096 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
4097 3 argument Alias with unlimited for the negative form.
4098 (fno-vect-cost-model): Removed.
4100 2017-01-05 Martin Liska <mliska@suse.cz>
4102 * hsa-gen.c (gen_hsa_divmod): New function.
4103 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
4105 2017-01-05 Martin Liska <mliska@suse.cz>
4107 PR pch/78970
4108 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
4109 header.
4111 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4113 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
4114 small constant length operands.
4116 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
4118 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
4119 between loop iterations.
4121 2017-01-05 Martin Liska <mliska@suse.cz>
4123 PR sanitizer/78815
4124 * gimplify.c (gimplify_decl_expr): Compare to
4125 asan_poisoned_variables instread of checking flags.
4126 (gimplify_target_expr): Likewise.
4127 (gimplify_expr): Likewise.
4128 (gimplify_function_tree): Conditionally initialize
4129 asan_poisoned_variables.
4131 2017-01-04 Jeff Law <law@redhat.com>
4133 PR tree-optimizatin/78812
4134 * rtl.h (contains_mem_rtx_p): Prototype.
4135 * ifcvt.c (containts_mem_rtx_p): Move from here to...
4136 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
4137 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
4138 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
4139 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
4141 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4143 * input.c (assert_char_at_range): Default-initialize actual_range.
4145 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4147 * df-scan.c (df_ref_create_structure): Make regno unsigned,
4148 to match the caller.
4150 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4152 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
4153 insns after final jump in test to emit dummy move.
4155 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4157 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
4158 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
4160 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
4162 * multiple_target.c (create_dispatcher_calls): Init e_next.
4163 * tree-ssa-loop-split.c (split_loop): Init border.
4164 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
4165 scalar_type.
4167 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
4169 PR target/71977
4170 PR target/70568
4171 PR target/78823
4172 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
4173 (altivec_register_operand): Do not return true if the operand
4174 contains a SUBREG mixing SImode and SFmode.
4175 (vsx_register_operand): Likewise.
4176 (vsx_reg_sfsubreg_ok): New predicate.
4177 (vfloat_operand): Do not return true if the operand contains a
4178 SUBREG mixing SImode and SFmode.
4179 (vint_operand): Likewise.
4180 (vlogical_operand): Likewise.
4181 (gpc_reg_operand): Likewise.
4182 (int_reg_operand): Likewise.
4183 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
4184 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
4185 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
4186 SImode and SFmode.
4187 (rs6000_emit_move_si_sf_subreg): New helper function.
4188 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
4189 fixup SUBREGs involving SImode and SFmode.
4190 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
4191 numbers for the new peephole2 optimization.
4192 (peephole2 for SFmode unions): New peephole2 to optimize cases in
4193 the GLIBC math library that do AND/IOR/XOR operations on single
4194 precision floating point.
4195 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
4196 target macros to say whether we need to avoid SUBREGs mixing
4197 SImode and SFmode.
4198 (TARGET_ALLOW_SF_SUBREG): Likewise.
4199 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
4200 (UNSPEC_SI_FROM_SF): Likewise.
4201 (iorxor): Change spacing.
4202 (and_ior_xor): New iterator for AND, IOR, and XOR.
4203 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
4204 (movdi_from_sf_zero_ext): Likewise.
4205 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
4206 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
4207 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
4208 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
4209 (fms<mode>4): Likewise.
4210 (fnma<mode>4): Likewise.
4211 (fnms<mode>4): Likewise.
4212 (nfma<mode>4): Likewise.
4213 (nfms<mode>4): Likewise.
4215 2017-01-04 Marek Polacek <polacek@redhat.com>
4217 PR c++/64767
4218 * doc/invoke.texi: Document -Wpointer-compare.
4220 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4222 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
4223 RejectNegative.
4225 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
4226 descriptions for -gdwarf-5 and emit them as uleb128 instead of
4227 2-byte data.
4229 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
4231 PR target/78056
4232 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
4233 documentation of the powerpc_popcntb_ok attribute.
4234 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
4235 code to issue warning messages if a requested CPU configuration is
4236 not supported by the binary (assembler and loader) toolchain.
4237 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
4238 made to define a built-in function that has been disabled.
4239 (paired_init_builtins): Add assertion to prevent ICE if attempt is
4240 made to define a built-in function that has been disabled.
4241 (altivec_init_builtins): Add comment explaining why definition
4242 of the DST built-in functions is not preceded by an assertion
4243 check. Add assertions to prevent ICE if attempts are made to
4244 define an altivec predicate or an abs* built-in function that has
4245 been disabled.
4246 (htm_init_builtins): Add comment explaining why definition of the
4247 htm built-in functions is not preceded by an assertion check.
4249 2017-01-04 Jeff Law <law@redhat.com>
4251 PR tree-optimizatin/67955
4252 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
4253 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
4254 the points-to solution does not include pt_null. Use DECL_PT_UID
4255 unconditionally.
4257 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
4259 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
4260 Use gen_int_mode instead of gen_lopwart for const_int operands.
4262 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4264 PR tree-optimization/71563
4265 * match.pd: Simplify X << Y into X if Y is known to be 0 or
4266 out of range value - has low bits known to be zero.
4268 2017-01-04 Alan Modra <amodra@gmail.com>
4270 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
4271 * configure: Regenerate.
4272 * config.in: Regenerate.
4274 2017-01-04 Jakub Jelinek <jakub@redhat.com>
4276 PR bootstrap/77569
4277 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
4278 a substring of the message, but strcmp with the whole message. Ifdef
4279 ENABLE_NLS, translate the message first using dgettext.
4281 2017-01-03 Jeff Law <law@redhat.com>
4283 PR tree-optimizatin/78856
4284 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
4285 (mark_threaded_blocks): Remove code to truncate thread paths that
4286 cross multiple loop headers. Instead invalidate the cached loop
4287 iteration information and handle case of a thread path walking
4288 into an irreducible region.
4290 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
4292 PR target/78900
4293 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
4294 assertions. Add support for doing the signbit if the IEEE 128-bit
4295 floating point value is in a GPR.
4296 * config/rs6000/rs6000.md (Fsignbit): Delete.
4297 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
4298 Update the length attribute if the value is in a GPR.
4299 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
4300 the sign or zero extension instruction, since the value is always 0/1.
4301 (signbit<mode>2_dm2): Delete using <Fsignbit>.
4303 PR target/78953
4304 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
4305 extracting SImode to a GPR register so that we can generate a
4306 store, limit the vector to be in a traditional Altivec register
4307 for the vextuwrx instruction.
4309 2017-01-03 Ian Lance Taylor <iant@google.com>
4311 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
4313 2017-01-03 Martin Sebor <msebor@redhat.com>
4315 PR tree-optimization/78696
4316 * gimple-ssa-sprintf.c (format_floating): Correct handling of
4317 precision. Use MPFR for %f for greater fidelity. Correct handling
4318 of %g.
4319 (pass_sprintf_length::compute_format_length): Set width and precision
4320 specified by asrerisk to void_node for vararg functions.
4321 (try_substitute_return_value): Adjust dump output.
4323 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
4325 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
4327 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
4329 * doc/invoke.texi (SPARC options): Document -mlra as the default.
4330 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
4331 -mlra/-mno-lra was passed to the compiler.
4333 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
4335 PR rtl-optimization/65618
4336 * emit-rtl.c (try_split): Move initialization of "before" and
4337 "after" to just before the call to emit_insn_after_setloc.
4339 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
4341 * doc/md.texi (Standard Names): Remove reference to Java frontend.
4343 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
4345 * dwarf2out.c (gen_enumeration_type_die): When
4346 -gno-strict-dwarf, add a DW_AT_encoding attribute.
4348 2017-01-03 Jakub Jelinek <jakub@redhat.com>
4350 PR tree-optimization/78965
4351 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
4352 Change first argument from const call_info & to call_info &. For %n
4353 set info.nowrite to false.
4355 PR middle-end/78901
4356 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
4357 possibly throwing calls.
4359 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
4360 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
4361 and fns handling, rather than in a separate case SSA_NAME.
4363 2017-01-02 Jeff Law <law@redhat.com>
4365 * config/darwin-driver.c (darwin_driver_init): Const-correctness
4366 fixes for first_period and second_period variables.
4368 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
4370 PR target/78967
4371 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
4372 (*insvqi_1): New insn pattern.
4373 (*insvqi_1_mem_rex64): Ditto.
4374 (*insvqi_2): Ditto.
4375 (*insvqi_3): Rename from *insvqi.
4377 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
4379 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
4381 * doc/cfg.texi (Edges): Remove reference to Java.
4382 (Maintaining the CFG): Ditto.
4384 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4386 PR middle-end/77674
4387 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
4388 transparent aliases.
4390 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
4392 PR middle-end/77484
4393 * predict.def (PRED_CALL): Update hitrate.
4394 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
4395 * predict.c (tree_estimate_probability_bb): Split CALL predictor
4396 into direct/indirect/polymorphic variants.
4398 2017-01-01 Jakub Jelinek <jakub@redhat.com>
4400 Update copyright years.
4402 * gcc.c (process_command): Update copyright notice dates.
4403 * gcov-dump.c (print_version): Ditto.
4404 * gcov.c (print_version): Ditto.
4405 * gcov-tool.c (print_version): Ditto.
4406 * gengtype.c (create_file): Ditto.
4407 * doc/cpp.texi: Bump @copying's copyright year.
4408 * doc/cppinternals.texi: Ditto.
4409 * doc/gcc.texi: Ditto.
4410 * doc/gccint.texi: Ditto.
4411 * doc/gcov.texi: Ditto.
4412 * doc/install.texi: Ditto.
4413 * doc/invoke.texi: Ditto.
4414 \f
4415 Copyright (C) 2017 Free Software Foundation, Inc.
4417 Copying and distribution of this file, with or without modification,
4418 are permitted in any medium without royalty provided the copyright
4419 notice and this notice are preserved.