1 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
4 * cgraphunit.c (cgraph_node::add_new_function): Set externally_visible
7 2017-02-09 Jan Hubicka <hubicka@ucw.cz>
9 * tree-ssa-loop-unswitch.c (hoist_guard): Update profile.
11 2017-02-09 Jakub Jelinek <jakub@redhat.com>
13 * omp-offload.c (oacc_loop_auto_partitions): Use || instead of |
17 * gimplify.h (is_gimple_sizepos): Only test for INTEGER_CST constants,
18 not arbitrary TREE_CONSTANT.
21 * gimplify.c (gimplify_adjust_omp_clauses): Ignore
22 "omp declare target link" attribute unless is_global_var.
23 * omp-offload.c (find_link_var_op): Likewise.
25 2017-02-09 Nathan Sidwell <nathan@codesourcery.com>
26 Chung-Lin Tang <cltang@codesourcery.com>
28 * gimplify.c (gimplify_scan_omp_clauses): No special handling for
30 (gimplify_adjust_omp_clauses): Don't delete TILE.
31 (gimplify_omp_for): Deal with TILE.
32 * internal-fn.c (expand_GOACC_TILE): New function.
33 * internal-fn.def (GOACC_DIM_POS): Comment may be overly conservative.
35 * omp-expand.c (struct oacc_collapse): Add tile and outer fields.
36 (expand_oacc_collapse_init): Add LOC paramter. Initialize tile
38 (expand_oacc_collapse_vars): Add INNER parm, adjust for tiling,
39 avoid DIV for outermost collapse var.
40 (expand_oacc_for): Insert tile element loop as needed. Adjust.
41 Remove out of date comments, fix whitespace.
42 * omp-general.c (omp_extract_for_data): Deal with tiling.
43 * omp-general.h (enum oacc_loop_flags): Add OLF_TILE flag,
44 adjust OLF_DIM_BASE value.
45 (struct omp_for_data): Add tiling field.
46 * omp-low.c (scan_sharing_clauses): Allow OMP_CLAUSE_TILE.
47 (lower_oacc_head_mark): Add OLF_TILE as appropriate. Ensure 2 levels
48 for auto loops. Remove default auto determining, moved to
49 oacc_loop_fixed_partitions.
50 * omp-offload.c (struct oacc_loop): Change 'ifns' to vector of call
51 stmts, add e_mask field.
52 (oacc_dim_call): New function, abstracted out from oacc_thread_numbers.
53 (oacc_thread_numbers): Use oacc_dim_call.
54 (oacc_xform_tile): New.
55 (new_oacc_loop_raw): Initialize e_mask, adjust for ifns vector.
56 (finish_oacc_loop): Adjust for ifns vector.
57 (oacc_loop_discover_walk): Append loop abstraction sites to list,
58 add case for GOACC_TILE fns.
59 (oacc_loop_xform_loop): Delete.
60 (oacc_loop_process): Iterate over call list directly, and add
61 handling for GOACC_TILE fns.
62 (oacc_loop_fixed_partitions): Determine default auto, deal with TILE,
64 (oacc_loop_auto_partitions): Add outer_assign parm. Assign all but
65 vector partitioning to outer loops. Assign 2 partitions to loops
66 when available. Add TILE handling.
67 (oacc_loop_partition): Adjust oacc_loop_auto_partitions call.
68 (execite_oacc_device_lower): Process GOACC_TILE fns, ignore unknown specs.
69 * tree-nested.c (convert_nonlocal_omp_clauses): Allow OMP_CLAUSE_TILE.
70 * tree.c (omp_clause_num_ops): Adjust TILE ops.
71 * tree.h (OMP_CLAUSE_TILE_ITERVAR, OMP_CLAUSE_TILE_COUNT): New.
73 2017-02-09 Gerald Pfeifer <gerald@pfeifer.com>
75 * configure.ac (ACX_BUGURL): Update.
76 * configure: Regenerate.
78 2017-02-09 Richard Biener <rguenther@suse.de>
80 PR tree-optimization/69823
81 * graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
82 Properly enumerate all BBs in the region. Use auto_vec/auto_bitmap.
84 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
86 * config/arc/arc-c.def: Add __NPS400__ definition.
87 * config/arc/arc.h (CPP_SPEC): Don't define __NPS400__ here.
88 (TARGET_NPS400): Define.
90 2017-02-09 Andrew Burgess <andrew.burgess@embecosm.com>
92 * config/arc/arc-arch.h (arc_arch_t): Move unchanged to earlier in
94 (arc_cpu_t): Change base_architecture field, arch, to a arc_arc_t
96 (arc_cpu_types): Fill the arch_info field with a pointer into the
98 (arc_selected_cpu): Declare.
99 * config/arc/arc.c (arc_selected_cpu): Make global.
100 (arc_selected_arch): Delete.
101 (arc_base_cpu): Delete.
102 (arc_override_options): Remove references to deleted variables,
103 update access to arch information.
104 (ARC_OPT): Update access to arch information.
105 (ARC_OPTX): Likewise.
106 * config/arc/arc.h (arc_base_cpu): Remove declaration.
107 (TARGET_ARC600): Update access to arch information.
108 (TARGET_ARC601): Likewise.
109 (TARGET_ARC700): Likewise.
110 (TARGET_EM): Likewise.
111 (TARGET_HS): Likewise.
112 * config/arc/driver-arc.c (arc_cpu_to_as): Update access to arch
115 2017-02-08 Pat Haugen <pthaugen@us.ibm.com>
118 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Invert
119 condition/operands for integer GE/LE/GEU/LEU operations.
121 2017-02-08 Segher Boessenkool <segher@kernel.crashing.org>
124 * config/rs6000/rs6000.opt (maltivec=le, maltivec=be): Fix spelling
127 2017-02-08 Martin Jambor <mjambor@suse.cz>
130 * ipa-prop.c (ipa_alloc_node_params): Make static, return bool
131 whether allocation happened.
132 (ipa_initialize_node_params): Do not call ipa_alloc_node_params if
133 nothing was allocated.
135 2017-02-08 Jakub Jelinek <jakub@redhat.com>
137 PR tree-optimization/79408
138 * tree-vrp.c (simplify_div_or_mod_using_ranges): If op1 is not
139 constant, but SSA_NAME with a known integer range, use the minimum
140 of that range instead of op1 to determine if modulo can be replaced
141 with its first operand.
143 2016-02-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
145 * config/riscv/riscv.c (riscv_build_integer_1): Avoid use of INT16_MAX.
147 2017-02-08 Richard Biener <rguenther@suse.de>
149 PR tree-optimization/71824
150 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
151 Check all loops contained in the merged region.
153 2017-02-07 Andrew Pinski <apinski@cavium.com>
155 * config/aarch64/aarch64.md (popcount<mode>2): New pattern.
157 2017-02-07 Andrew Pinski <apinski@cavium.com>
159 * config/aarch64/aarch64-cores.def (thunderx): Disable LSE.
160 (thunderxt88): Likewise.
161 (thunderxt81): Disable LSE and change v8.1 to v8.
162 (thunderxt83): Likewise.
164 2017-02-07 Jakub Jelinek <jakub@redhat.com>
165 Richard Biener <rguenther@suse.de>
168 * ira-int.h (struct target_ira_int): Change x_max_struct_costs_size
169 type from int to size_t.
170 * ira-costs.c (struct_costs_size): Change type from int to size_t.
172 2017-02-07 Jakub Jelinek <jakub@redhat.com>
174 PR rtl-optimization/79386
175 * cprop.c (bypass_conditional_jumps): Initialize
176 bypass_last_basic_block already before splitting bbs after
177 unconditional traps...
178 (bypass_conditional_jumps): ... rather than here.
181 * config/i386/sse.md (xtg_mode, gatherq_mode): New mode attrs.
182 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2,
183 *avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2): Use them,
184 fix -masm=intel patterns.
186 2017-02-07 Richard Biener <rguenther@suse.de>
188 PR tree-optimization/79256
190 * builtins.c (get_object_alignment_2): Use min_align_of_type
191 to extract alignment for MEM_REFs to honor BIGGEST_FIELD_ALIGNMENT
192 and ADJUST_FIELD_ALIGN.
194 * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional
196 * doc/tm.texi: Regenerate.
197 * stor-layout.c (layout_decl): Adjust.
198 (update_alignment_for_field): Likewise.
199 (place_field): Likewise.
200 (min_align_of_type): Likewise.
201 * config/arc/arc.h (ADJUST_FIELD_ALIGN): Adjust.
202 * config/epiphany/epiphany.h (ADJUST_FIELD_ALIGN): Likewise.
203 * config/epiphany/epiphany.c (epiphany_adjust_field_align): Likewise.
204 * config/frv/frv.h (ADJUST_FIELD_ALIGN): Likewise.
205 * config/frv/frv.c (frv_adjust_field_align): Likewise.
206 * config/i386/i386.h (ADJUST_FIELD_ALIGN): Likewise.
207 * config/i386/i386.c (x86_field_alignment): Likewise.
208 * config/rs6000/aix.h (ADJUST_FIELD_ALIGN): Likewise.
209 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Likewise.
210 * config/rs6000/freebsd64.h (ADJUST_FIELD_ALIGN): Likewise.
211 * config/rs6000/linux64.h (ADJUST_FIELD_ALIGN): Likewise.
212 * config/rs6000/sysv4.h (ADJUST_FIELD_ALIGN): Likewise.
213 * config/rs6000/rs6000.c (rs6000_special_adjust_field_align_p):
217 2017-01-30 Richard Biener <rguenther@suse.de>
219 PR tree-optimization/79256
220 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
221 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
224 2017-02-07 Toma Tabacu <toma.tabacu@imgtec.com>
226 * config/mips/mips.c (mips_expand_builtin_insn): Convert the QImode
227 argument of the pshufh, psllh, psllw, psrah, psraw, psrlh, psrlw
228 builtins to SImode and emit a zero-extend, if necessary.
230 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
232 * docs/invoke.texi (RISC-V Options): Alphabetize.
234 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
236 * doc/invoke.texi (RISC-V Options): Use two spaces to separate
239 2017-02-06 Palmer Dabbelt <palmer@dabbelt.com>
241 * config/riscv/riscv.c: New file.
242 * gcc/common/config/riscv/riscv-common.c: Likewise.
243 * config.gcc: Likewise.
244 * config/riscv/constraints.md: Likewise.
245 * config/riscv/elf.h: Likewise.
246 * config/riscv/generic.md: Likewise.
247 * config/riscv/linux.h: Likewise.
248 * config/riscv/multilib-generator: Likewise.
249 * config/riscv/peephole.md: Likewise.
250 * config/riscv/pic.md: Likewise.
251 * config/riscv/predicates.md: Likewise.
252 * config/riscv/riscv-builtins.c: Likewise.
253 * config/riscv/riscv-c.c: Likewise.
254 * config/riscv/riscv-ftypes.def: Likewise.
255 * config/riscv/riscv-modes.def: Likewise.
256 * config/riscv/riscv-opts.h: Likewise.
257 * config/riscv/riscv-protos.h: Likewise.
258 * config/riscv/riscv.h: Likewise.
259 * config/riscv/riscv.md: Likewise.
260 * config/riscv/riscv.opt: Likewise.
261 * config/riscv/sync.md: Likewise.
262 * config/riscv/t-elf-multilib: Likewise.
263 * config/riscv/t-linux: Likewise.
264 * config/riscv/t-linux-multilib: Likewise.
265 * config/riscv/t-riscv: Likewise.
266 * configure.ac: Likewise.
267 * doc/contrib.texi: Add Kito Cheng, Palmer Dabbelt, and Andrew
268 Waterman as RISC-V maintainers.
269 * doc/install.texi: Add RISC-V entries.
270 * doc/invoke.texi: Add RISC-V options section.
271 * doc/md.texi: Add RISC-V constraints section.
272 * configure: Regenerated.
274 2017-02-06 Michael Meissner <meissner@linux.vnet.ibm.com>
277 * config/rs6000/vector.md (vcond<mode><mode>): Allow the true and
278 false values to be constant vectors with all 0 or all 1 bits set.
279 (vcondu<mode><mode>): Likewise.
280 * config/rs6000/predicates.md (vector_int_reg_or_same_bit): New
282 (fpmask_comparison_operator): Update comment.
283 (vecint_comparison_operator): New predicate.
284 * config/rs6000/rs6000.c (rs6000_emit_vector_cond_expr): Optimize
285 vector conditionals when the true and false values are constant
286 vectors with all 0 bits or all 1 bits set.
288 2017-02-06 Martin Sebor <msebor@redhat.com>
290 PR tree-optimization/79376
291 * gimple-fold.c (get_range_strlen): Set the minimum length to zero.
293 2017-02-06 Uros Bizjak <ubizjak@gmail.com>
295 * config/i386/sse.md (vector modes -> vec_extract* splitter): Use
296 explicit subreg RTX with operand 1. Use VECTOR_MODE_P predicate
297 to simplify split condition.
299 2017-02-06 Jakub Jelinek <jakub@redhat.com>
301 * omp-expand.c (oxpand_omp_atomic_fetch_op,
302 expand_omp_atomic_pipeline): Return false if can_atomic_load_p is
305 2017-02-06 Segher Boessenkool <segher@kernel.crashing.org>
307 PR rtl-optimization/68664
308 * target.def (can_speculate_insn): New hook.
309 * doc/tm.texi.in (TARGET_SCHED_CAN_SPECULATE_INSN): New hook.
310 * doc/tm.texi: Regenerate.
311 * sched-rgn.c (can_schedule_ready_p): Use the new hook.
312 * config/rs6000/rs6000.c (TARGET_SCHED_CAN_SPECULATE_INSN): New macro.
313 (rs6000_sched_can_speculate_insn): New function.
315 2017-02-06 Jakub Jelinek <jakub@redhat.com>
317 PR tree-optimization/79284
318 * tree-vectorizer.h (VECT_SCALAR_BOOLEAN_TYPE_P): Define.
319 * tree-vect-stmts.c (vect_get_vec_def_for_operand,
320 vectorizable_mask_load_store, vectorizable_operation,
321 vect_is_simple_cond, get_same_sized_vectype): Use it instead
322 of comparing TREE_CODE of a type against BOOLEAN_TYPE.
323 * tree-vect-patterns.c (check_bool_pattern, search_type_for_mask_1,
324 vect_recog_bool_pattern, vect_recog_mask_conversion_pattern): Likewise.
325 * tree-vect-slp.c (vect_get_constant_vectors): Likewise.
326 * tree-vect-loop.c (vect_determine_vectorization_factor): Likewise.
327 Remove redundant gimple_code (stmt) == GIMPLE_ASSIGN test after
328 is_gimple_assign (stmt). Replace another such test with
329 is_gimple_assign (stmt).
331 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
334 * config/avr/avr.c (rtl-iter.h): Include it.
335 (TARGET_LEGITIMATE_COMBINED_INSN): New hook define...
336 (avr_legitimate_combined_insn): ...and implementation.
338 2017-02-06 Dominik Vogt <vogt@linux.vnet.ibm.com>
340 * config/s390/predicates.md ("larl_operand"): Use macros from hwint.h.
341 * config/s390/s390.c (s390_const_operand_ok)
342 (s390_canonicalize_comparison, s390_extract_part)
343 (s390_single_part, s390_contiguous_bitmask_nowrap_p)
344 (s390_contiguous_bitmask_p, s390_rtx_costs)
345 (legitimize_pic_address): Likewise.
346 * config/s390/s390.md ("clzdi2", "clztidi2"): Likewise.
347 * config/s390/vx-builtins.md ("vec_genbytemaskv16qi")
348 ("vec_permi<mode>", "vfae<mode>", "*vfaes<mode>", "vstrc<mode>")
349 ("*vstrcs<mode>"): Use UINTVAL() to set unsigned HOST_WIDE_INT.
350 * config/s390/vector.md ("vec_vfenes<mode>"): Likewise.
352 2017-02-06 Georg-Johann Lay <avr@gjlay.de>
354 * config/avr/avr.md (*addhi3_zero_extend): Add alternative where
355 REGNO($0) == REGNO($1).
357 2017-02-06 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
359 * config/s390/linux.h(SIZE_TYPE): Add comment.
361 2017-02-06 Julian Brown <julian@codesourcery.com>
362 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
363 Virendra Pathak <virendra.pathak@broadcom.com>
365 * config/aarch64/aarch64-cores.def: Change the scheduler
367 * config/aarch64/aarch64.md: Include thunderx2t99.md.
368 * config/aarch64/thunderx2t99.md: New file.
370 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
372 * doc/standards.texi (Go Language): Update link to language
375 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
377 * tree-eh.c (lower_resx): Sanitize profile.
378 (cleanup_empty_eh_move_lp): Likewise.
380 2017-02-05 Jan Hubicka <hubicka@ucw.cz>
383 * cfgloopmanip.c (lv_adjust_loop_entry_edge, loop_version): Add
385 * cfgloopmanip.h (loop_version): Update prototype.
386 * modulo-sched.c (sms_schedule): Update call of loop_version.
387 * tree-if-conv.c(version_loop_for_if_conversion): Likewise.
388 * tree-parloops.c (gen_parallel_loop): Likewise.
389 * tree-ssa-loop-manip.c (tree_transform_and_unroll_loop): Likewise.
390 * tree-ssa-loop-split.c (split_loop): Likewise.
391 * tree-ssa-loop-unswitch.c (tree_unswitch_loop): Likewise.
392 * tree-vect-loop-manip.c (vect_loop_versioning): Likewise.
394 2017-02-05 Martin Liska <mliska@suse.cz>
397 * config/s390/s390.c (s390_gimplify_va_arg): Initialize local
399 (print_operand_address): Initialize a struct to zero.
401 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
403 * doc/contrib.texi (Contributors): Refer to Hans Boehm's
404 garbage collector only in textual form.
406 2017-02-05 Gerald Pfeifer <gerald@pfeifer.com>
408 * doc/extend.texi (x86 specific memory model extensions for
409 transactional memory): Simplify a phrase.
411 2017-02-05 Eric Botcazou <ebotcazou@adacore.com>
414 * config/sparc/sync.md (atomic_loaddi_1): Replace 'U' constraint with
415 'r', 'm' constraint with 'T' and !TARGET_ARCH64 with TARGET_ARCH32.
416 (atomic_storedi_1): Likewise.
418 2017-02-04 Jakub Jelinek <jakub@redhat.com>
420 PR tree-optimization/79338
421 * tree-parloops.c (gather_scalar_reductions): Don't call
422 vect_analyze_loop_form for loop->inner before destroying loop's
425 2017-02-03 Martin Sebor <msebor@redhat.com>
427 PR tree-optimization/79327
428 * gimple-ssa-sprintf.c (tree_digits): Avoid adding the base prefix
429 when precision has resulted in leading zeros.
430 (format_integer): Adjust the likely counter to assume an unknown
431 argument that may be zero is non-zero.
433 2017-02-03 Jason Merrill <jason@redhat.com>
436 * tree-inline.c (copy_tree_body_r) [COND_EXPR]: Revert change to
437 avoid copying non-taken branch.
439 2017-02-03 Jakub Jelinek <jakub@redhat.com>
441 PR tree-optimization/79340
442 * tree-vect-loop.c (vectorizable_reduction): Release
443 vec_defs elements after safe_splicing them into other vectors.
446 PR tree-optimization/79327
447 * gimple-ssa-sprintf.c (adjust_range_for_overflow): If returning
448 true, always set *argmin and *argmax to TYPE_{MIN,MAX}_VALUE of
450 (format_integer): Use wide_int_to_tree instead of build_int_cst
451 + to_?hwi. If argmin is NULL, just set argmin and argmax to
452 TYPE_{MIN,MAX}_VALUE of argtype. Simplify and fix computation
453 of shortest and longest sequence.
455 2017-02-03 Uros Bizjak <ubizjak@gmail.com>
457 * config/i386/i386.c (dimode_scalar_chain::convert_reg):
458 Use pextrd for TARGET_SSE4_1 when creating scalar copy.
460 2017-02-03 Walter Lee <walt@tilera.com>
463 * config/tilegx/tilegx.md (tilegx_expand_prologue): Add blockage
464 after initial stackframe link reg save.
465 * config/tilepro/tilepro.md (tilepro_expand_prologue): Likewise.
467 2017-02-03 Jakub Jelinek <jakub@redhat.com>
470 * config/rs6000/rs6000.md (movsi_from_sf): Use wb constraint instead of
471 wu for stxssp alternative.
473 2017-02-03 Martin Sebor <msebor@redhat.com>
475 PR tree-optimization/79352
476 * gimple-fold.c (get_range_strlen): Add argument.
477 (get_range_strlen): Change return type to bool.
478 (get_maxval_strlen): Pass in a dummy argument.
479 * gimple-fold.h (get_range_strlen): Change return type to bool.
480 * gimple-ssa-sprintf.c (get_string_length): Set unlikely counter.
481 * tree.h (array_at_struct_end_p): Add argument.
482 * tree.c (array_at_struct_end_p): Handle it.
484 2017-02-03 Martin Liska <mliska@suse.cz>
487 * multiple_target.c (create_dispatcher_calls): Redirect edge
488 from a caller of a dispatcher.
489 (expand_target_clones): Make the clones local.
490 (ipa_target_clone): Do both target clones and resolvers.
491 (ipa_dispatcher_calls): Remove the pass.
492 (pass_dispatcher_calls::gate): Likewise.
493 (make_pass_dispatcher_calls): Likewise.
494 * passes.def (pass_target_clone): Put as very first IPA early
497 2017-02-03 Martin Liska <mliska@suse.cz>
499 * symtab.c (symtab_node::binds_to_current_def_p): Bail out
500 in case of a function with ifunc attribute.
502 2017-02-03 Martin Liska <mliska@suse.cz>
504 * cgraph.c (cgraph_node::dump): Dump function version info.
505 * symtab.c (symtab_node::dump_base): Add missing new line.
507 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
509 * tree-ssa-ifcombine.c (update_profile_after_ifcombine): New function.
510 (ifcombine_ifandif): Use it.
512 2017-02-03 Martin Liska <mliska@suse.cz>
514 * doc/invoke.texi: Document default value for
515 use-after-scope-direct-emission-threshold.
517 2017-02-03 Martin Liska <mliska@suse.cz>
519 PR tree-optimization/79339
520 * gimple-ssa-sprintf.c (format_floating_max): Call mpfr_clear.
521 (format_floating): Likewise.
523 2017-02-03 Martin Liska <mliska@suse.cz>
526 * ipa-prop.c (ipa_node_params_t::insert): Remove current
528 (ipa_node_params_t::remove): Likewise.
529 * ipa-prop.h (ipa_node_params::ipa_node_params): Make default
530 initialization from removed ipa_node_params_t::insert.
531 (ipa_node_params::~ipa_node_params): Move from removed
532 ipa_node_params_t::release.
533 * symbol-summary.h (symbol_summary::m_released): New member.
534 Do not release a summary twice. Do not allow to call finalizer
535 for types of a summary that live in GGC memory.
537 2017-02-02 Naveen H.S <Naveen.Hurugalawadi@cavium.com>
539 * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and
542 2017-02-02 Martin Sebor <msebor@redhat.com>
545 * gimple-ssa-sprintf.c (get_string_length): Set lower bound to zero.
546 (format_string): Tighten up the range of output for non-constant
547 strings and correct the expected range for wide non-constant strings.
549 2017-02-02 Martin Sebor <msebor@redhat.com>
551 * doc/invoke.texi (-maccumulate-args): Fix bad grammar.
554 * doc/invoke.texi (-fdump-rtl-): Remove pass-specific options from
556 (-fdump-tree-@var): Add to index and document how to come up
557 with pass-specific option and dump file names.
558 (-fdump-passes): Clarify where to look for output.
560 2017-02-02 Jan Hubicka <hubicka@ucw.cz>
563 * tree-ssa-threadbackward.c (profitable_jump_thread_path): Dump
564 statistics of the analyzed path; allow threading for speed when
565 any of BBs along the path are optimized for speed.
567 2017-02-02 Eric Botcazou <ebotcazou@adacore.com>
570 * emit-rtl.c (init_emit): Add ??? comment for problematic alignment
571 settings of the virtual registers.
574 2016-08-23 Dominik Vogt <vogt@linux.vnet.ibm.com>
576 * explow.c (get_dynamic_stack_size): Take known alignment of stack
577 pointer + STACK_DYNAMIC_OFFSET into account when calculating the size
580 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
582 * config/s390/vx-builtins.md ("vec_ceil", "vec_floor")
583 ("vec_trunc", "vec_roundc", "vec_round"): Remove expanders.
585 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
587 * config/s390/s390.md: Add missing comments with the expanded
589 * config/s390/vector.md: Likewise.
590 * config/s390/vx-builtins.md: Likewise.
592 2017-02-02 Jakub Jelinek <jakub@redhat.com>
595 * config/rs6000/rs6000.md (*fixuns_trunc<mode>di2_fctiduz): Rename to ...
596 (fixuns_trunc<mode>di2): ... this, remove previous expander. Put all
597 conditions on a single line.
599 2017-02-02 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
601 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
602 __S390_VX__ to __VX__.
604 2017-02-01 Andrew Pinski <apinski@cavium.com>
606 * tree-vect-loop.c (vect_compute_single_scalar_iteration_cost): Pass
607 stmt_info to record_stmt_cost.
608 (vect_get_known_peeling_cost): Pass stmt_info if known to
610 * config/aarch64/aarch64-protos.h (cpu_vector_cost): Split
611 cpu_vector_cost field into
612 scalar_int_stmt_cost and scalar_fp_stmt_cost. Split vec_stmt_cost
613 field into vec_int_stmt_cost and vec_fp_stmt_cost.
614 * config/aarch64/aarch64.c (generic_vector_cost): Update for the
615 splitting of scalar_stmt_cost and vec_stmt_cost.
616 (thunderx_vector_cost): Likewise.
617 (cortexa57_vector_cost): LIkewise.
618 (exynosm1_vector_cost): Likewise.
619 (xgene1_vector_cost): Likewise.
620 (thunderx2t99_vector_cost): Improve after the splitting of the two
622 (aarch64_builtin_vectorization_cost): Update for the splitting of
623 scalar_stmt_cost and vec_stmt_cost.
625 2017-02-01 Torvald Riegel <triegel@redhat.com>
626 Richard Henderson <rth@redhat.com>
628 * builtins.c (fold_builtin_atomic_always_lock_free): Make "lock-free"
629 conditional on existance of a fast atomic load.
630 * optabs-query.c (can_atomic_load_p): New function.
631 * optabs-query.h (can_atomic_load_p): Declare it.
632 * optabs.c (expand_atomic_exchange): Always delegate to libatomic if
633 no fast atomic load is available for the particular size of access.
634 (expand_atomic_compare_and_swap): Likewise.
635 (expand_atomic_load): Likewise.
636 (expand_atomic_store): Likewise.
637 (expand_atomic_fetch_op): Likewise.
638 * testsuite/lib/target-supports.exp
639 (check_effective_target_sync_int_128): Remove x86 because it provides
641 (check_effective_target_sync_int_128_runtime): Likewise.
643 2017-02-01 Richard Biener <rguenther@suse.de>
645 * graphite.c: Include tree-vectorizer.h for find_loop_location.
646 (graphite_transform_loops): Provide opt-info for optimized nests.
647 * tree-parloop.c (parallelize_loops): Provide opt-info for
650 2017-02-01 Richard Biener <rguenther@suse.de>
653 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it
656 2017-02-01 Richard Biener <rguenther@suse.de>
658 PR tree-optimization/71824
659 * graphite-scop-detection.c (scop_detection::build_scop_breadth):
660 Verify the loops are valid in the merged SESE region.
661 (scop_detection::can_represent_loop_1): Check analyzing the
662 evolution of the number of iterations in the region succeeds.
664 2017-01-31 Ian Lance Taylor <iant@golang.org>
666 * config/i386/i386.c (ix86_expand_split_stack_prologue): Add
667 REG_ARGS_SIZE note to 32-bit push insns and call insn.
669 2017-01-31 David Malcolm <dmalcolm@redhat.com>
671 PR preprocessor/79210
672 * input.c (get_substring_ranges_for_loc): Replace line_width
673 assertion with error-handling.
675 2017-01-31 Richard Biener <rguenther@suse.de>
677 PR tree-optimization/77318
678 * graphite-sese-to-poly.c (extract_affine): Fix assert.
679 (create_pw_aff_from_tree): Take loop parameter.
680 (add_condition_to_pbb): Pass loop of the condition to
681 create_pw_aff_from_tree.
683 2017-01-31 Jakub Jelinek <jakub@redhat.com>
685 * config/s390/s390.c (s390_asan_shadow_offset): New function.
686 (TARGET_ASAN_SHADOW_OFFSET): Redefine.
688 2017-01-31 Michael Meissner <meissner@linux.vnet.ibm.com>
692 * config/rs6000/rs6000-protos.h (convert_float128_to_int): Delete,
694 (convert_int_to_float128): Likewise.
695 * config/rs6000/rs6000.c (convert_float128_to_int): Likewise.
696 (convert_int_to_float128): Likewise.
697 * config/rs6000/rs6000.md (UNSPEC_IEEE128_MOVE): Likewise.
698 (UNSPEC_IEEE128_CONVERT): Likewise.
699 (floatsi<mode>2, FLOAT128 iterator): Bypass calling
700 rs6000_expand_float128_convert if we have IEEE 128-bit hardware.
701 Use local variables for IBM extended format.
702 (fix_trunc<mode>si2, FLOAT128 iterator): Likewise.
703 (fix_trunc<mode>si2_fprs): Likewise.
704 (fixuns_trunc<IEEE128:mode><SDI:mode>2): Likewise.
705 (floatuns<IEEE128:mode>2, IEEE128 iterator): Likewise.
706 (fix<uns>_<mode>si2_hw): Rework the IEEE 128-bt hardware support
707 to know that we can now have integers of all sizes in vector
709 (fix<uns>_<mode>di2_hw): Likewise.
710 (float<uns>_<mode>si2_hw): Likewise.
711 (fix_<mode>si2_hw): Likewise.
712 (fixuns_<mode>si2_hw): Likewise.
713 (float<uns>_<mode>di2_hw): Likewise.
714 (float_<mode>di2_hw): Likewise.
715 (float_<mode>si2_hw): Likewise.
716 (floatuns_<mode>di2_hw): Likewise.
717 (floatuns_<mode>si2_hw): Likewise.
718 (xscvqp<su>wz_<mode>): Delete, no longer used.
719 (xscvqp<su>dz_<mode>): Likewise.
720 (xscv<su>dqp_<mode>): Likewise.
721 (ieee128_mfvsrd_64bit): Likewise.
722 (ieee128_mfvsrd_32bit): Likewise.
723 (ieee128_mfvsrwz): Likewise.
724 (ieee128_mtvsrw): Likewise.
725 (ieee128_mtvsrd_64bit): Likewise.
726 (ieee128_mtvsrd_32bit): Likewise.
728 2017-01-31 Martin Liska <mliska@suse.cz>
731 * ipa-prop.c (ipa_free_all_node_params): Call release method
732 instead of ~sumbol_summary to not to trigger double times
735 2017-01-31 Aldy Hernandez <aldyh@redhat.com>
737 PR tree-optimization/71691
738 * bitmap.h (class auto_bitmap): New.
739 * tree-ssa-loop-unswitch.c (tree_may_unswitch_on): Call
740 is_maybe_undefined instead of ssa_undefined_value_p.
742 2017-01-31 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
744 * config/s390/s390-c.c (s390_cpu_cpp_builtins_internal): Rename
745 __S390_ARCH_LEVEL__ to __ARCH__.
747 2017-01-31 Jakub Jelinek <jakub@redhat.com>
749 PR tree-optimization/79267
750 * value-prof.c (gimple_ic): Only drop lhs for noreturn calls
751 if should_remove_lhs_p is true.
753 2017-01-30 Alexandre Oliva <aoliva@redhat.com>
756 * dwarf2out.c (clone_as_declaration): Drop DW_AT_alignment.
757 (add_alignment_attribute): New.
758 (base_type_die): Add alignment attribute.
759 (subrange_type_die): Likewise.
760 (modified_type_die): Likewise.
761 (gen_array_type_die): Likewise.
762 (gen_descr_array_type_die: Likewise.
763 (gen_enumeration_type_die): Likewise.
764 (gen_subprogram_die): Likewise.
765 (gen_variable_die): Likewise.
766 (gen_field_die): Likewise.
767 (gen_ptr_to_mbr_type_die): Likewise.
768 (gen_struct_or_union_type_die): Likewise.
769 (gen_subroutine_type_die): Likewise.
770 (gen_typedef_die): Likewise.
771 (base_type_cmp): Compare alignment attribute.
773 2017-01-30 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
776 * config/rs6000/altivec.md (*setb_internal): Rename to setb_signed.
777 (setb_unsigned) New pattern for setb with CCUNS.
778 * config/rs6000/rs6000.c (expand_block_compare): Use a different
779 subfc./subfe sequence to avoid overflow problems. Generate a
780 shorter sequence with cmpld/setb for power9.
781 * config/rs6000/rs6000.md (subf<mode>3_carry_dot2): Add a new pattern
782 for generating subfc. instruction.
783 (cmpstrsi): Add TARGET_POPCNTD predicate as the generate sequence
784 now uses this instruction.
786 2017-01-30 Ian Lance Taylor <iant@google.com>
789 * dwarf2out.c (gen_type_die_with_usage): When picking a variant
790 for FUNCTION_TYPE/METHOD_TYPE, use the first matching one.
792 2017-01-30 Martin Sebor <msebor@redhat.com>
794 * gimple-ssa-sprintf.c (fmtresult::adjust_for_width_or_precision):
795 Move constant to the right of a relational operator.
796 (get_mpfr_format_length, format_character, format_string): Ditto.
797 (should_warn_p, maybe_warn): Same.
799 * doc/invoke.texi (-Wformat-truncation=1): Fix typo.
801 2017-01-30 Maxim Ostapenko <m.ostapenko@samsung.com>
804 * asan.c (get_translation_unit_decl): Remove function.
805 (asan_add_global): Force has_dynamic_init to zero in LTO mode.
807 2017-01-30 Martin Liska <mliska@suse.cz>
809 PR gcov-profile/79259
810 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
813 2017-01-30 Martin Liska <mliska@suse.cz>
816 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
817 Initialize variables with NULL value.
819 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
822 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
824 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
826 2017-01-30 Richard Biener <rguenther@suse.de>
828 PR tree-optimization/79276
829 * tree-vrp.c (process_assert_insertions): Properly adjust common
830 when removing a duplicate.
832 * gcc.dg/torture/pr79276.c: New testcase.
834 2017-01-30 Richard Biener <rguenther@suse.de>
836 PR tree-optimization/79256
837 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
838 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
840 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
842 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
845 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
846 ("*r<noxa>sbg_<mode>_sll_bitmask")
847 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
848 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
849 Use contiguous_bitmask_nowrap_operand.
851 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
854 * config/rs6000/altivec.h (vec_xl): Revise #define.
857 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
859 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
861 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
863 PR rtl-optimization/79194
864 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
865 traps before call to bypass_conditional_jumps.
867 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
869 PR tree-optimization/71374
870 * lra-constraints.c (check_conflict_input_operands): New.
871 (match_reload): Use it.
873 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
876 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
877 account to calculate conflict_set.
879 2017-01-27 Bin Cheng <bin.cheng@arm.com>
881 PR rtl-optimization/78559
882 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
883 other_insn in combine.
885 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
887 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
888 uint16_type_node for BT_UINT16.
890 2017-01-27 David Malcolm <dmalcolm@redhat.com>
892 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
894 (GIMPLE Tests): New node.
895 (RTL Tests): New node.
897 2017-01-27 Richard Biener <rguenther@suse.de>
899 PR tree-optimization/79245
900 * tree-loop-distribution.c (distribute_loop): Apply cost
901 modeling also to detected patterns.
903 2017-01-27 Richard Biener <rguenther@suse.de>
905 PR tree-optimization/71433
906 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
907 (compare_assert_loc): New function.
908 (process_assert_insertions): Sort and optimize assert locations
909 to remove duplicates and push down identical assertions on
910 edges to their destination block.
912 2017-01-27 Richard Biener <rguenther@suse.de>
914 PR tree-optimization/79244
915 * tree-vrp.c (remove_range_assertions): Forcefully propagate
916 out SSA names even if abnormal.
918 2017-01-27 Jakub Jelinek <jakub@redhat.com>
920 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
921 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
922 instead of MPFR_RNDN.
924 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
927 * arm.c (arm_option_override): Don't call build_target_option_node
928 until after doing all option overrides.
929 (arm_valid_target_attribute_tree): Likewise.
931 2017-01-27 Martin Liska <mliska@suse.cz>
933 * doc/invoke.texi (-fprofile-arcs): Document profiling support
934 for {cd}tors and C++ {cd}tors.
936 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
938 * config/s390/s390.md ("*setmem_long_and")
939 ("*setmem_long_and_31z"): Use zero_extend instead of and.
941 2017-01-26 Martin Sebor <msebor@redhat.com>
943 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
946 2017-01-26 Martin Sebor <msebor@redhat.com>
948 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
949 HAVE_DFmode before using XFmode or DFmode.
950 (parse_directive): Avoid using the z length modifier to avoid
951 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
954 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
955 to accept adjustment as an array.
956 (get_int_range): New function.
957 (struct directive): Make width and prec arrays.
958 (directive::set_width, directive::set_precision): Call get_int_range.
959 (format_integer, format_floating): Handle width and precision ranges.
960 (format_string, parse_directive): Same.
962 2017-01-26 Jakub Jelinek <jakub@redhat.com>
965 * dwarf2out.c (generate_skeleton_bottom_up): For children with
966 comdat_type_p set, just clone them, but keep the children in the
970 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
971 which have direct callers with -fvar-tracking-assignments enabled
973 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
974 inside of type units.
976 2017-01-26 Martin Sebor <msebor@redhat.com>
979 * gimple-ssa-sprintf.c (struct result_range): Add likely and
981 (struct format_result): Replace number_chars, number_chars_min,
982 and number_chars_max with a single member of struct result_range.
984 (format_result::operator+=): Adjust.
985 (struct fmtresult): Remove bounded. Handle likely and unlikely
987 (fmtresult::adjust_for_width_or_precision): New function.
988 (fmtresult:type_max_digits): New function.
989 (bytes_remaining): Handle likely and unlikely counters.
990 (min_bytes_remaining): Remove.
991 (format_percent): Simplify.
992 (format_integer, format_floating): Set likely and unlikely counters.
993 (get_string_length, format_character, format_string): Same.
994 (format_plain, should_warn_p): New function.
995 (maybe_warn): Call should_warn_p. Update diagnostic messages
996 and handle those for all directives, including plain strings.
997 (format_directive): Handle likely and unlikely counters.
998 Remove unnecessary quoting from diagnostics. Add an informational
1000 (add_bytes): Remove.
1001 (pass_sprintf_length::compute_format_length): Simplify.
1002 (try_substitute_return_value): Handle likely and unlikely counters.
1004 2017-01-26 Carl Love <cel@us.ibm.com>
1006 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
1007 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
1009 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
1012 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
1013 endianess for subregs into account.
1014 * lra-constraints.c (lra_constraints): Do risky transformations
1015 always on the first iteration.
1016 * lra-lives.c (check_pseudos_live_through_calls): Add arg
1017 last_call_used_reg_set.
1018 (process_bb_lives): Define and use last_call_used_reg_set.
1019 * lra.c (lra): Always continue after lra_constraints on the first
1022 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
1024 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
1026 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
1028 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1030 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
1031 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
1032 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
1033 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
1034 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
1035 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
1036 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
1037 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
1038 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
1040 2017-01-26 Marek Polacek <polacek@redhat.com>
1043 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
1044 for the third operand.
1046 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1049 * omp-low.c (struct omp_context): Add simt_stmt field.
1050 (scan_omp_for): Return omp_context *.
1051 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
1052 context to the _simt_ SIMD stmt.
1053 (lower_omp_for): For combined SIMD with sibling _simt_
1054 SIMD, make sure to use the same decls in _looptemp_
1055 clauses as in the sibling.
1057 2017-01-26 David Sherwood <david.sherwood@arm.com>
1060 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
1063 2017-01-26 Jakub Jelinek <jakub@redhat.com>
1066 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
1067 emit fld b; fld a; if possible.
1069 * brig-builtins.def: Update copyright years.
1070 * config/arm/arm_acle_builtins.def: Update copyright years.
1072 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
1075 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
1076 constraint instead of o for the stxsd instruction.
1078 2017-01-25 Carl Love <cel@us.ibm.com>
1080 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
1081 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
1083 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
1085 * doc/invoke.texi (C++ Dialect Options): Fix typo.
1087 2017-01-25 Richard Biener <rguenther@suse.de>
1089 PR tree-optimization/69264
1090 * target.def (vector_alignment_reachable): Improve documentation.
1091 * doc/tm.texi: Regenerate.
1092 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
1094 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
1095 earlier changes with respect to TYPE_USER_ALIGN.
1096 (vector_alignment_reachable_p): Likewise. Improve dumping.
1098 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1101 * config/arm/arm.md (xordi3): Force constant operand into a register
1104 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1106 * doc/invoke.texi (-fstore-merging): Correct default optimization
1107 levels at which it is enabled.
1108 (-O): Move -fstore-merging from list to...
1111 2017-01-25 Richard Biener <rguenther@suse.de>
1114 * omp-expand.c: Include debug.h.
1115 (expand_omp_taskreg): Make sure to generate early debug before
1116 outlining anything from a function.
1117 (expand_omp_target): Likewise.
1118 (grid_expand_target_grid_body): Likewise.
1120 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
1123 * asan.c (get_translation_unit_decl): New function.
1124 (asan_add_global): Extract modules file name from globals
1125 TRANSLATION_UNIT_DECL name.
1127 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
1130 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
1131 for long calls with APCS frame and VFP.
1133 2017-01-24 David Malcolm <dmalcolm@redhat.com>
1135 * cfg.c (original_copy_tables_initialized_p): New function.
1136 * cfg.h (original_copy_tables_initialized_p): New decl.
1137 * cfgrtl.c (relink_block_chain): Guard the call to
1138 free_original_copy_tables with a call to
1139 original_copy_tables_initialized_p.
1140 * cgraph.h (symtab_node::native_rtl_p): New decl.
1141 * cgraphunit.c (symtab_node::native_rtl_p): New function.
1142 (symtab_node::needed_p): Don't assert for early assembly output
1143 for __RTL functions.
1144 (cgraph_node::finalize_function): Set "force_output" for __RTL
1146 (cgraph_node::analyze): Bail out early for __RTL functions.
1147 (analyze_functions): Update assertion to support __RTL functions.
1148 (cgraph_node::expand): Bail out early for __RTL functions.
1149 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
1151 * function.h (struct function): Update comment for field
1153 * gimple-expr.c: Include "tree-pass.h".
1154 (gimple_has_body_p): Return false for __RTL functions.
1155 * Makefile.in (OBJS): Add run-rtl-passes.o.
1156 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
1158 (gcc::pass_manager::get_clean_slate): New accessor.
1159 * passes.c: Include "insn-addr.h".
1160 (should_skip_pass_p): Add logging. Update logic for running
1161 "expand" to be compatible with both __GIMPLE and __RTL. Guard
1162 property-provider override so it is only done for gimple passes.
1164 (skip_pass): New function.
1165 (execute_one_pass): Call skip_pass when skipping passes.
1166 * read-md.c (md_reader::read_char): Support filtering
1167 the input to a subset of line numbers.
1168 (md_reader::md_reader): Initialize fields
1169 m_first_line and m_last_line.
1170 (md_reader::read_file_fragment): New function.
1171 * read-md.h (md_reader::read_file_fragment): New decl.
1172 (md_reader::m_first_line): New field.
1173 (md_reader::m_last_line): New field.
1174 * read-rtl-function.c (function_reader::create_function): Only
1175 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
1176 curr_properties. Set DECL_INITIAL to a dummy block.
1177 (read_rtl_function_body_from_file_range): New function.
1178 * read-rtl-function.h (read_rtl_function_body_from_file_range):
1180 * run-rtl-passes.c: New file.
1181 * run-rtl-passes.h: New file.
1183 2017-01-24 Jeff Law <law@redhat.com>
1185 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
1188 2017-01-24 Bin Cheng <bin.cheng@arm.com>
1190 PR tree-optimization/79159
1191 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
1192 (record_nonwrapping_iv): Improve boundary using above function if no
1193 value range information.
1195 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
1196 Martin Jambor <mjambor@suse.cz>
1198 * brig-builtins.def: New file.
1199 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
1200 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
1201 (DEF_HSAIL_SAT_BUILTIN): Likewise.
1202 (DEF_HSAIL_INTR_BUILTIN): Likewise.
1203 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
1204 * builtin-types.def (BT_INT8): New.
1205 (BT_INT16): Likewise.
1206 (BT_UINT8): Likewise.
1207 (BT_UINT16): Likewise.
1208 (BT_FN_ULONG): Likewise.
1209 (BT_FN_UINT_INT): Likewise.
1210 (BT_FN_UINT_ULONG): Likewise.
1211 (BT_FN_UINT_LONG): Likewise.
1212 (BT_FN_UINT_PTR): Likewise.
1213 (BT_FN_ULONG_PTR): Likewise.
1214 (BT_FN_INT8_FLOAT): Likewise.
1215 (BT_FN_INT16_FLOAT): Likewise.
1216 (BT_FN_UINT32_FLOAT): Likewise.
1217 (BT_FN_UINT16_FLOAT): Likewise.
1218 (BT_FN_UINT8_FLOAT): Likewise.
1219 (BT_FN_UINT64_FLOAT): Likewise.
1220 (BT_FN_UINT16_UINT32): Likewise.
1221 (BT_FN_UINT32_UINT16): Likewise.
1222 (BT_FN_UINT16_UINT16_UINT16): Likewise.
1223 (BT_FN_INT_PTR_INT): Likewise.
1224 (BT_FN_UINT_PTR_UINT): Likewise.
1225 (BT_FN_LONG_PTR_LONG): Likewise.
1226 (BT_FN_ULONG_PTR_ULONG): Likewise.
1227 (BT_FN_VOID_UINT64_UINT64): Likewise.
1228 (BT_FN_UINT8_UINT8_UINT8): Likewise.
1229 (BT_FN_INT8_INT8_INT8): Likewise.
1230 (BT_FN_INT16_INT16_INT16): Likewise.
1231 (BT_FN_INT_INT_INT): Likewise.
1232 (BT_FN_UINT_FLOAT_UINT): Likewise.
1233 (BT_FN_FLOAT_UINT_UINT): Likewise.
1234 (BT_FN_ULONG_UINT_UINT): Likewise.
1235 (BT_FN_ULONG_UINT_PTR): Likewise.
1236 (BT_FN_ULONG_ULONG_ULONG): Likewise.
1237 (BT_FN_UINT_UINT_UINT): Likewise.
1238 (BT_FN_VOID_UINT_PTR): Likewise.
1239 (BT_FN_UINT_UINT_PTR: Likewise.
1240 (BT_FN_UINT32_UINT64_PTR): Likewise.
1241 (BT_FN_INT_INT_UINT_UINT): Likewise.
1242 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
1243 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
1244 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
1245 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
1246 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
1247 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
1248 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
1249 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
1250 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
1251 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
1252 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
1253 * doc/frontends.texi: List BRIG FE.
1254 * doc/install.texi (Testing): Add BRIG tesring requirements.
1255 * doc/invoke.texi (Overall Options): Mention BRIG.
1256 * doc/standards.texi (Standards): Doucment BRIG HSA version.
1258 2017-01-24 Richard Biener <rguenther@suse.de>
1260 PR translation/79208
1261 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
1263 2017-01-24 Martin Jambor <mjambor@suse.cz>
1266 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
1267 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
1270 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
1273 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
1274 casts from signed to unsigned really don't have a range.
1276 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1278 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
1279 GMP_RNDx for compatiblity.
1281 2017-01-24 Martin Liska <mliska@suse.cz>
1284 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
1285 that would prevent us to call alloca with -1 as argument.
1287 2017-01-24 Jakub Jelinek <jakub@redhat.com>
1289 * dwarf2out.c (output_compilation_unit_header, output_file_names):
1290 Avoid -Wformat-security warning.
1292 2017-01-23 Andrew Pinski <apinski@cavium.com>
1294 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
1297 2017-01-23 Martin Sebor <msebor@redhat.com>
1300 * gimple-ssa-sprintf.c (warn_level): New global.
1301 (format_integer): Use it here and throughout the rest of the file.
1302 Use the same switch to compute sign as base.
1303 (maybe_warn): New function.
1304 (format_directive): Factor out warnings into maybe_warn.
1305 Add debugging output. Use warn_level.
1306 (add_bytes): Use warn_level.
1307 (pass_sprintf_length::compute_format_length): Add debugging output.
1308 (try_substitute_return_value): Same.
1309 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
1312 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
1313 (struct fmtresult, format_integer, format_floating): Adjust.
1314 (fmtresult::fmtresult): Set max correctly in two argument ctor.
1315 (get_string_length, format_string,format_directive): Same.
1316 (pass_sprintf_length::compute_format_length): Same.
1317 (try_substitute_return_value): Simplify slightly.
1320 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
1321 (fmtresult::operator+=): Outlined.
1322 (struct fmtresult): Add ctors.
1323 (struct conversion_spec): Rename...
1324 (struct directive): ...to this. Add and remove data members.
1325 (directive::set_width, directive::set_precision): New functions.
1326 (format_percent): Use fmtresult ctor.
1327 (get_width_and_precision): Remove.
1328 (format_integer): Make naming changes. Avoid computing width and
1330 (format_floating): Same. Adjust indentation.
1331 (format_character, format_none): New functions.
1332 (format_string): Moved character handling to format_character.
1333 (format_directive): Remove arguments, change return type.
1334 (parse_directive): New function.
1335 (pass_sprintf_length::compute_format_length): Move directive
1336 parsing to parse_directive.
1338 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1340 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
1341 (assign_assembler_name_if_needed): ... this.
1342 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
1343 (assign_assembler_name_if_needed): ... this.
1344 (free_lang_data_in_cgraph): Adjust callers.
1345 * cgraphunit.c (cgraph_node::analyze): Likewise.
1346 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
1349 2017-01-23 Richard Biener <rguenther@suse.de>
1351 PR tree-optimization/79088
1352 PR tree-optimization/79188
1353 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
1354 resetting loop bounds after last path deletion. Reset loop
1355 bounds of the target loop, make code match the comments.
1356 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
1357 Make sure loops need no fixups.
1359 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
1361 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
1362 exponent support with double type for first argument.
1363 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
1364 type returned by __builtin_vec_extract_sig,
1365 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
1366 functions from "vector int" to "vector unsigned int" or from
1367 "vector long long int" to "vector unsigned long long int".
1368 Changed type returned by __builtin_vec_extract_exp,
1369 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
1370 functions from "vector int" to "vector unsigned int" or from
1371 "vector long long int" to "vector unsigned long long int".
1372 Changed return type of __builtin_vec_test_data_class,
1373 __builtin_vec_test_data_class_sp, and
1374 __builtin_vec_test_data_class_dp from "vector int" to
1375 "vector bool int" or from "vector long long int" to "vector bool
1376 long long int" and changed second argument type from "unsigned
1377 int" to "int". Added new overloaded function forms "vector float
1378 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
1379 "vector float __builtin_vec_insert_exp_sp (vector float, vector
1380 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
1381 double, vector unsigned long long int)" and "vector double
1382 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
1383 long int)". Changed return type of
1384 __builtin_scalar_test_data_class and
1385 __builtin_scalar_test_data_class_sp and
1386 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
1387 int" and changed second argument from "unsigned int" to "int".
1388 Changed type returned by __builtin_scalar_test_neg,
1389 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
1390 from "int" to "bool int". Added new overloaded function form
1391 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
1392 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
1393 exponent double-precision with floating point first argument.
1394 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
1395 documentation of scalar_test_data_class, scalar_test_neg,
1396 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
1397 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
1398 vec_test_data_class built-in functions to reflect refinements in
1399 their type signatures.
1401 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
1403 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
1405 (aarch64_elf_asm_destructor): Likewise.
1407 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
1409 PR rtl-optimization/78634
1410 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
1411 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
1412 * ifcvt.c (noce_try_cmove): Add missing cost check.
1414 PR rtl-optimization/71724
1415 * combine.c (if_then_else_cond): Look for situations where it is
1416 beneficial to undo the work of one of the recursive calls.
1418 2017-01-23 Bin Cheng <bin.cheng@arm.com>
1420 PR tree-optimization/70754
1421 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
1422 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
1423 combined stmt before it if not NULL.
1424 (combine_chains): Process refs reversely and compute dominance point
1427 2017-01-23 Martin Liska <mliska@suse.cz>
1429 PR tree-optimization/79196
1430 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
1431 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
1433 (strlen_optimize_stmt): Call the renamed function.
1435 2017-01-23 Michael Matz <matz@suse.de>
1437 PR tree-optimization/78384
1438 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
1440 2017-01-23 Richard Biener <rguenther@suse.de>
1442 PR tree-optimization/79186
1443 * tree-vrp.c (register_new_assert_for): Make sure we've seen
1444 both incoming edges before moving an assert.
1446 2017-01-23 Martin Jambor <mjambor@suse.cz>
1448 * ipa-prop.c (load_from_param_1): Removed.
1449 (load_from_unmodified_param): Bits from load_from_param_1 put back
1451 (load_from_param): Removed.
1452 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
1453 with stmt. Reverted back to use of load_from_unmodified_param.
1455 2017-01-23 Martin Jambor <mjambor@suse.cz>
1458 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
1459 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
1460 field a pointer to garbage collected vector, mark lattices and
1461 ipcp_orig_node with GTY((skip)).
1462 (ipa_get_param_count): Adjust to descriptors being a pointer.
1463 (ipa_get_param): Likewise.
1464 (ipa_get_type): Likewise.
1465 (ipa_get_param_move_cost): Likewise.
1466 (ipa_set_param_used): Likewise.
1467 (ipa_get_controlled_uses): Likewise.
1468 (ipa_set_controlled_uses): Likewise.
1469 (ipa_is_param_used): Likewise.
1470 (ipa_node_params_t): Move into garbage collector. New methods insert
1472 (ipa_node_params_sum): Annotate wth GTY(()).
1473 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
1475 (ipa_load_from_parm_agg): Adjust declaration.
1476 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
1477 * ipa-profile.c (ipa_profile): Likewise.
1478 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
1479 (ipa_populate_param_decls): Make descriptors parameter garbage
1481 (ipa_dump_param): Adjust to descriptors being a pointer.
1482 (ipa_alloc_node_params): Likewise.
1483 (ipa_initialize_node_params): Likewise.
1484 (load_from_param_1): Make descriptors parameter garbage collected.
1485 (load_from_unmodified_param): Likewise.
1486 (load_from_param): Likewise.
1487 (ipa_load_from_parm_agg): Likewise.
1488 (ipa_node_params::~ipa_node_params): Removed.
1489 (ipa_free_all_node_params): Remove call to delete operator.
1490 (ipa_node_params_t::insert): New.
1491 (ipa_node_params_t::remove): Likewise.
1492 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
1493 copy known_csts and known_contexts vectors.
1494 (ipa_read_node_info): Adjust to descriptors being a pointer.
1495 (ipcp_modif_dom_walker): Make m_descriptors field garbage
1497 (ipcp_transform_function): Make descriptors variable garbage
1500 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
1502 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
1503 * config/i386/avx512dqintrin.h: Ditto.
1504 * config/i386/avx512fintrin.h: Ditto.
1505 * gcc/config/i386/i386.c: Handle new builtins.
1506 * config/i386/i386-builtin.def: Add new builtins.
1507 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
1508 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
1510 2017-01-23 Jakub Jelinek <jakub@redhat.com>
1511 Martin Liska <mliska@suse.cz>
1513 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
1514 * asan.c (asan_expand_poison_ifn): Support stores and use
1515 appropriate ASAN report function.
1516 * internal-fn.c (expand_ASAN_POISON_USE): New function.
1517 * internal-fn.def (ASAN_POISON_USE): Declare.
1518 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
1519 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
1520 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
1521 ASAN_POISON calls w/o LHS.
1522 * tree-ssa.c (execute_update_addresses_taken): Create clobber
1523 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
1524 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
1525 * gimplify.c (asan_poison_variables): Add attribute
1526 use_after_scope_memory to variables that really needs to live
1528 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
1529 having the attribute.
1531 2017-01-23 Martin Liska <mliska@suse.cz>
1533 * asan.c (create_asan_shadow_var): New function.
1534 (asan_expand_poison_ifn): Likewise.
1535 * asan.h (asan_expand_poison_ifn): New declaration.
1536 * internal-fn.c (expand_ASAN_POISON): Likewise.
1537 * internal-fn.def (ASAN_POISON): New builtin.
1538 * sanopt.c (pass_sanopt::execute): Expand
1539 asan_expand_poison_ifn.
1540 * tree-inline.c (copy_decl_for_dup_finish): Make function
1542 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
1543 * tree-ssa.c (is_asan_mark_p): New function.
1544 (execute_update_addresses_taken): Rewrite local variables
1545 (identified just by use-after-scope as addressable) into SSA.
1547 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1549 * doc/install.texi (Specific): opensource.apple.com uses https
1550 now. Remove trailing slash.
1552 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
1554 * README.Portability: Remove note on an Irix compatibility issue.
1556 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
1558 * gcov.c (INCLUDE_ALGORITHM): Define.
1559 (INCLUDE_VECTOR): Define.
1560 No longer include <vector> and <algorithm> directly.
1562 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
1564 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
1566 * doc/invoke.texi (Code Gen Options): Ditto.
1568 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
1571 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
1573 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
1575 rtl-optimization/79125
1576 * cprop.c (local_cprop_pass): Handle cases where we make an
1579 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
1583 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
1584 read from, for big endian.
1586 2017-01-20 Jiong Wang <jiong.wang@arm.com>
1588 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
1589 register pauth builtins for LP64 only.
1591 2017-01-20 Marek Polacek <polacek@redhat.com>
1594 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
1597 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1599 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
1601 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
1602 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
1603 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
1605 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1608 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
1609 in big-endian mode when they are not a single duplicated value.
1611 2017-01-20 Richard Biener <rguenther@suse.de>
1613 * BASE-VER: Bump to 7.0.1.
1615 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
1617 * omp-low.c (omplow_simd_context): New struct. Use it...
1618 (lower_rec_simd_input_clauses): ...here and...
1619 (lower_rec_input_clauses): ...here to hold common data. Adjust all
1620 references to idx, lane, max_vf, is_simt.
1622 2017-01-20 Graham Markall <graham.markall@embecosm.com>
1624 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
1627 2017-01-20 Martin Jambor <mjambor@suse.cz>
1629 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
1630 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
1632 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
1633 (GTFILES): Rename hsa.c to hsa-common.c.
1634 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
1635 * hsa-dump.c: Likewise.
1636 * hsa-gen.c: Likewise.
1637 * hsa-regalloc.c: Likewise.
1638 * ipa-hsa.c: Likewise.
1639 * omp-expand.c: Likewise.
1640 * omp-low.c: Likewise.
1641 * toplev.c: Likewise.
1643 2017-01-20 Marek Polacek <polacek@redhat.com>
1646 * doc/invoke.texi: Document -Wduplicated-branches.
1647 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
1648 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
1649 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
1650 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
1651 return 0 only when not OEP_LEXICOGRAPHIC.
1652 (fold_build_cleanup_point_expr): Use the expression
1653 location when building CLEANUP_POINT_EXPR.
1654 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
1655 * tree.c (add_expr): Handle error_mark_node.
1657 2017-01-20 Martin Liska <mliska@suse.cz>
1660 * tree-profile.c (init_ic_make_global_vars): Do not call
1662 (gimple_init_gcov_profiler): Likewise.
1664 2017-01-20 Martin Liska <mliska@suse.cz>
1667 * cgraph.h (maybe_create_reference): Remove argument and
1669 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
1671 * ipa-cp.c (create_specialized_node): Likewise.
1672 * symtab.c (symtab_node::maybe_create_reference): Handle
1673 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
1675 2017-01-20 Martin Liska <mliska@suse.cz>
1677 * read-rtl-function.c (function_reader::create_function): Use
1678 build_decl instread of build_decl_stat.
1680 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
1682 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
1683 * config/i386/avx512dqintrin.h: Ditto.
1684 * config/i386/avx512fintrin.h: Ditto.
1685 * config/i386/i386-builtin-types.def: Add new types.
1686 * gcc/config/i386/i386.c: Handle new types.
1687 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
1688 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
1689 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
1690 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
1691 (__builtin_ia32_kshiftridi): New.
1692 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
1694 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
1698 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
1699 define to rs6000_init_stack_protect_guard.
1700 (rs6000_init_stack_protect_guard): New function.
1702 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1703 Yunqiang Su <yunqiang.su@imgtec.com>
1705 * config.gcc (supported_defaults): Add madd4.
1706 (with_madd4): Add validation.
1707 (all_defaults): Add madd4.
1708 * config/mips/mips.opt (mmadd4): New option.
1709 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1711 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1713 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
1714 (ISA_HAS_FUSED_MADD4): Likewise.
1715 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
1716 * gcc/doc/install.texi (--with-madd4): Document the new option.
1718 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1720 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
1721 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
1722 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
1723 (aarch64_init_pauth_hint_builtins): New.
1724 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
1725 (aarch64_expand_builtin): Expand new builtins.
1727 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1729 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
1730 * combine-stack-adj.c (no_unhandled_cfa): Handle
1731 REG_CFA_TOGGLE_RA_MANGLE.
1732 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
1733 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
1734 info for return address signing.
1735 (aarch64_expand_epilogue): Likewise.
1737 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1739 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
1740 * config/aarch64/aarch64-protos.h
1741 (aarch64_return_address_signing_enabled): New declaration.
1742 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
1744 (aarch64_expand_prologue): Sign return address before it's pushed onto
1746 (aarch64_expand_epilogue): Authenticate return address fetched from
1748 (aarch64_override_options): Sanity check for ILP32 and ISA level.
1749 (aarch64_attributes): New function attributes for "sign-return-address".
1750 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
1751 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
1752 ("*do_return"): Generate combined instructions according to key index.
1753 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
1754 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
1756 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
1757 * config/aarch64/aarch64.opt (msign-return-address=): New.
1758 * doc/extend.texi (AArch64 Function Attributes): Documents
1759 "sign-return-address=".
1760 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
1762 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
1764 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
1765 overall option summary.
1767 2017-01-19 Jiong Wang <jiong.wang@arm.com>
1769 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
1770 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
1771 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
1772 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
1774 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
1776 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
1777 -mpower9-minmax by default for -mcpu=power9.
1778 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
1779 128-bit floating point.
1781 2017-01-20 Alan Modra <amodra@gmail.com>
1783 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
1784 optimizing for size.
1786 2017-01-20 Alan Modra <amodra@gmail.com>
1789 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
1790 for strcmp and strncmp from corresponding builtin decl.
1792 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1794 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
1795 instead of i386/rtems-64.h.
1796 * config/i386/rtems-64.h: Remove.
1798 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
1802 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1804 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1806 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1808 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1809 Change int to HOST_WIDE_INT.
1810 * config/aarch64/aarch64-protos.h
1811 (aarch64_simd_gen_const_vector_dup): Likewise.
1812 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1814 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1816 * langhooks-def.h (lhd_type_for_size): New decl.
1817 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1818 * langhooks.c (lhd_type_for_size): New function, taken from
1821 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1823 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1824 define_bypass for CR latency.
1825 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1826 (power9-alu2): Add define_bypass for CR latency.
1828 (power9-mul): Update insn latency.
1829 (power9-mul-compare): Update insn latency, bypass latency and remove
1832 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1834 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1836 * config/aarch64/aarch64.md
1837 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1838 aarch64_nopcrelative_literal_loads.
1839 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1841 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1843 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1845 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1847 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1850 * config.gcc (supported_defaults): Add lxc1-sxc1.
1851 (with_lxc1_sxc1): Add validation.
1852 (all_defaults): Add lxc1-sxc1.
1853 * config/mips/mips.opt (mlxc1-sxc1): New option.
1854 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1856 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1857 __mips_no_lxc1_sxc1.
1858 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1859 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1860 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1862 2017-01-19 Richard Biener <rguenther@suse.de>
1864 PR tree-optimization/72488
1865 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1866 sure to restore SSA info.
1867 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1869 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1871 PR rtl-optimization/79121
1872 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1873 of the inner type when shifting an extended value.
1875 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1878 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1879 interposable aliases.
1881 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1884 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1885 Use the evmergelohi instruction.
1886 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1887 (mov_sitf_e500_subreg8_2_be): Likewise.
1888 (mov_sitf_e500_subreg12_2_le): Likewise.
1889 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1890 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1891 (mov_sitf_e500_subreg8_2_le): Likewise.
1892 (mov_sitf_e500_subreg12_2_be): Likewise.
1894 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1896 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1897 attribute from vecsimple to vecperm.
1898 (altivec_vbpermq2): Likewise.
1900 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1903 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1905 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1906 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1907 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1908 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1909 case where N arg is SIZE_MAX.
1910 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1911 (cmpstrsi): Add pattern.
1913 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1915 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1916 __builtin_vec_revb builtins.
1917 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1918 built-in functions to support generation of the ISA 3.0 XXBR<x>
1919 vector byte reverse instructions.
1920 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1921 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1922 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1923 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1924 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1925 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1926 (P9V_BUILTIN_VEC_REVB): Likewise.
1927 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1928 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1929 (p9_xxbrq_v16qi): Likewise.
1930 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1931 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1932 (p9_xxbrh_v8hi): Likewise.
1933 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1934 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1935 vec_revb built-in functions.
1937 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1939 PR rtl-optimization/78952
1940 * config/i386/i386.md (any_extract): New code iterator.
1941 (*insvqi_2): Use any_extract for source operand.
1942 (*insvqi_3): Use any_shiftrt for source operand.
1944 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1946 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1948 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1950 2017-01-18 Matthias Klose <doko@ubuntu.com>
1952 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1954 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1956 * config/rs6000/altivec.h (vec_bperm): Change #define.
1957 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1958 (altivec_vbpermq2): New define_insn.
1959 (altivec_vbpermd): Likewise.
1960 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1962 (VBPERMD): Likewise.
1963 (VBPERM): New polymorphic function interface.
1964 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1965 Add entries for P9V_BUILTIN_VEC_VBPERM.
1966 * doc/extend.texi: Add interfaces for vec_bperm.
1968 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1970 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1971 first letter of error messages.
1972 (s390_resolve_overloaded_builtin): Likewise.
1973 * config/s390/s390.c (s390_expand_builtin): Likewise.
1974 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1975 (s390_valid_target_attribute_inner_p): Likewise.
1976 * config/s390/s390.md ("tabort"): Likewise.
1978 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1980 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1981 (ISA_AVOID_DIV_HILO): New macro.
1982 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1983 (ISA_HAS_DDIV): Likewise.
1985 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1987 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1989 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1991 * doc/invoke.texi (fabi-version): Spelling fix.
1993 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1996 * doc/invoke.texi (fabi-version): Mention mangling fix for
1999 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
2002 * doc/invoke.texi (fabi-version): Document discriminator mangling.
2004 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
2007 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
2008 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
2010 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
2012 (stack_protect_test): Ditto.
2013 * config/rs6000/rs6000.opt (mstack-protector-guard=,
2014 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
2016 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
2017 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
2018 -mstack-protector-guard-offset=.
2019 (RS/6000 and PowerPC Options): Ditto.
2021 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2023 * config/i386/i386.h (MASK_CLASS_P): New define.
2024 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
2025 there are no registers from different register sets also when
2026 mask registers are used. Update function comment.
2027 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
2028 to (*k/*r) and (*k/*km) alternatives.
2030 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
2032 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
2033 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
2034 (EH_RETURN_HANDLER_RTX): New define.
2035 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
2036 Force frame pointer in EH return functions.
2037 (aarch64_expand_epilogue): Add barrier for eh_return.
2038 (aarch64_final_eh_return_addr): Remove.
2039 (aarch64_eh_return_handler_rtx): New function.
2040 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
2042 (aarch64_eh_return_handler_rtx): New prototype.
2044 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2046 * config/rs6000/altivec.h (vec_rlmi): New #define.
2047 (vec_vrlnm): Likewise.
2048 (vec_rlnm): Likewise.
2049 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
2050 (UNSPEC_VRLNM): Likewise.
2051 (VIlong): New mode iterator.
2052 (altivec_vrl<VI_char>mi): New define_insn.
2053 (altivec_vrl<VI_char>nm): Likewise.
2054 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
2057 (RLNM): New polymorphic function entry.
2058 (VRLWMI): New monomorphic function entry.
2060 (RLMI): New polymorphic function entry.
2061 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
2062 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
2063 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
2066 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2069 * dwarf2out.c (field_byte_offset): Restore the
2070 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
2071 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
2072 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
2075 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2078 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
2080 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2083 * dwarf2out.c (add_data_member_location_attribute): For constant
2084 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
2085 instead of DW_AT_data_member_location, DW_AT_bit_offset and
2086 DW_AT_byte_size attributes.
2088 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
2090 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
2091 after forcing to constant memory when the code model is medium.
2093 2017-01-17 Julia Koval <julia.koval@intel.com>
2096 * config/i386/avx512fintrin.h
2097 (_mm512_i32gather_ps): Change __addr type to void const*.
2098 (_mm512_mask_i32gather_ps): Ditto.
2099 (_mm512_i32gather_pd): Ditto.
2100 (_mm512_mask_i32gather_pd): Ditto.
2101 (_mm512_i64gather_ps): Ditto.
2102 (_mm512_mask_i64gather_ps): Ditto.
2103 (_mm512_i64gather_pd): Ditto.
2104 (_mm512_mask_i64gather_pd): Ditto.
2105 (_mm512_i32gather_epi32): Ditto.
2106 (_mm512_mask_i32gather_epi32): Ditto.
2107 (_mm512_i32gather_epi64): Ditto.
2108 (_mm512_mask_i32gather_epi64): Ditto.
2109 (_mm512_i64gather_epi32): Ditto.
2110 (_mm512_mask_i64gather_epi32): Ditto.
2111 (_mm512_i64gather_epi64): Ditto.
2112 (_mm512_mask_i64gather_epi64): Ditto.
2113 (_mm512_i32scatter_ps): Change __addr type to void*.
2114 (_mm512_mask_i32scatter_ps): Ditto.
2115 (_mm512_i32scatter_pd): Ditto.
2116 (_mm512_mask_i32scatter_pd): Ditto.
2117 (_mm512_i64scatter_ps): Ditto.
2118 (_mm512_mask_i64scatter_ps): Ditto.
2119 (_mm512_i64scatter_pd): Ditto.
2120 (_mm512_mask_i64scatter_pd): Ditto.
2121 (_mm512_i32scatter_epi32): Ditto.
2122 (_mm512_mask_i32scatter_epi32): Ditto.
2123 (_mm512_i32scatter_epi64): Ditto.
2124 (_mm512_mask_i32scatter_epi64): Ditto.
2125 (_mm512_i64scatter_epi32): Ditto.
2126 (_mm512_mask_i64scatter_epi32): Ditto.
2127 (_mm512_i64scatter_epi64): Ditto.
2128 (_mm512_mask_i64scatter_epi64): Ditto.
2129 * config/i386/avx512pfintrin.h
2130 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
2131 (_mm512_mask_prefetch_i32gather_ps): Ditto.
2132 (_mm512_mask_prefetch_i64gather_pd): Ditto.
2133 (_mm512_mask_prefetch_i64gather_ps): Ditto.
2134 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
2135 (_mm512_prefetch_i32scatter_ps): Ditto.
2136 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
2137 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
2138 (_mm512_prefetch_i64scatter_pd): Ditto.
2139 (_mm512_prefetch_i64scatter_ps): Ditto.
2140 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
2141 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
2142 * config/i386/avx512vlintrin.h
2143 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
2144 (_mm_mmask_i32gather_ps): Ditto.
2145 (_mm256_mmask_i32gather_pd): Ditto.
2146 (_mm_mmask_i32gather_pd): Ditto.
2147 (_mm256_mmask_i64gather_ps): Ditto.
2148 (_mm_mmask_i64gather_ps): Ditto.
2149 (_mm256_mmask_i64gather_pd): Ditto.
2150 (_mm_mmask_i64gather_pd): Ditto.
2151 (_mm256_mmask_i32gather_epi32): Ditto.
2152 (_mm_mmask_i32gather_epi32): Ditto.
2153 (_mm256_mmask_i32gather_epi64): Ditto.
2154 (_mm_mmask_i32gather_epi64): Ditto.
2155 (_mm256_mmask_i64gather_epi32): Ditto.
2156 (_mm_mmask_i64gather_epi32): Ditto.
2157 (_mm256_mmask_i64gather_epi64): Ditto.
2158 (_mm_mmask_i64gather_epi64): Ditto.
2159 (_mm256_i32scatter_ps): Change __addr type to void*.
2160 (_mm256_mask_i32scatter_ps): Ditto.
2161 (_mm_i32scatter_ps): Ditto.
2162 (_mm_mask_i32scatter_ps): Ditto.
2163 (_mm256_i32scatter_pd): Ditto.
2164 (_mm256_mask_i32scatter_pd): Ditto.
2165 (_mm_i32scatter_pd): Ditto.
2166 (_mm_mask_i32scatter_pd): Ditto.
2167 (_mm256_i64scatter_ps): Ditto.
2168 (_mm256_mask_i64scatter_ps): Ditto.
2169 (_mm_i64scatter_ps): Ditto.
2170 (_mm_mask_i64scatter_ps): Ditto.
2171 (_mm256_i64scatter_pd): Ditto.
2172 (_mm256_mask_i64scatter_pd): Ditto.
2173 (_mm_i64scatter_pd): Ditto.
2174 (_mm_mask_i64scatter_pd): Ditto.
2175 (_mm256_i32scatter_epi32): Ditto.
2176 (_mm256_mask_i32scatter_epi32): Ditto.
2177 (_mm_i32scatter_epi32): Ditto.
2178 (_mm_mask_i32scatter_epi32): Ditto.
2179 (_mm256_i32scatter_epi64): Ditto.
2180 (_mm256_mask_i32scatter_epi64): Ditto.
2181 (_mm_i32scatter_epi64): Ditto.
2182 (_mm_mask_i32scatter_epi64): Ditto.
2183 (_mm256_i64scatter_epi32): Ditto.
2184 (_mm256_mask_i64scatter_epi32): Ditto.
2185 (_mm_i64scatter_epi32): Ditto.
2186 (_mm_mask_i64scatter_epi32): Ditto.
2187 (_mm256_i64scatter_epi64): Ditto.
2188 (_mm256_mask_i64scatter_epi64): Ditto.
2189 (_mm_i64scatter_epi64): Ditto.
2190 (_mm_mask_i64scatter_epi64): Ditto.
2191 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
2192 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
2193 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
2194 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
2195 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
2196 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
2197 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
2198 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
2199 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
2200 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
2201 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
2202 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
2203 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
2204 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
2205 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
2206 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
2207 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
2208 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
2209 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
2210 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
2211 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
2212 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
2213 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
2214 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
2215 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
2216 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
2217 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
2218 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
2219 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
2220 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
2221 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
2222 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
2223 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
2224 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
2225 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
2226 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
2227 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
2228 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
2229 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
2230 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
2231 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
2232 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
2233 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
2234 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
2235 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
2236 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
2237 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
2238 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
2239 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
2240 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
2241 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
2242 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
2243 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
2244 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
2245 definitions accordingly.
2247 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
2248 Kuan-Lin Chen <kuanlinchentw@gmail.com>
2251 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
2254 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
2257 * ira-conflicts.c (ira_build_conflicts): Update total conflict
2258 hard regs for inner regno.
2260 2017-01-17 Martin Liska <mliska@suse.cz>
2263 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
2264 assumption and add comment.
2266 2017-01-17 Nathan Sidwell <nathan@acm.org>
2268 * ipa-visibility.c (localize_node): New function, broken out of ...
2269 (function_and_variable_visibility): ... here. Call it.
2271 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
2274 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
2275 correctly set frequency of oudgoing edge.
2276 (duplicate_thread_path): Fix profile updating.
2278 2017-01-17 Jakub Jelinek <jakub@redhat.com>
2281 * configure.ac: Add GCC_BASE_VER.
2282 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
2283 version from BASE-VER file.
2284 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
2285 (gcc.o): Depend on $(BASEVER).
2286 * common.opt (dumpfullversion): New option.
2287 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
2288 * doc/invoke.texi: Document -dumpfullversion.
2289 * doc/install.texi: Document --with-gcc-major-version-only.
2290 * configure: Regenerated.
2292 2017-01-17 Richard Biener <rguenther@suse.de>
2294 PR tree-optimization/71433
2295 * tree-vrp.c (register_new_assert_for): Merge same asserts
2296 on all incoming edges.
2297 (process_assert_insertions_for): Handle insertions at the
2300 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
2302 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
2303 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
2305 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
2308 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
2311 2017-01-17 Alan Modra <amodra@gmail.com>
2314 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
2315 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
2316 symbolic stack limit when pic.
2318 2017-01-16 Martin Sebor <msebor@redhat.com>
2320 PR tree-optimization/78608
2321 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
2323 2017-01-16 Jeff Law <law@redhat.com>
2326 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
2327 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
2328 for several include directories that may be relative to sysroot.
2329 * config/i386/x-mingw32 (gplus_includedir): Define.
2330 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
2331 (native_system_includedir): Likewise.
2332 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
2333 override if TARGET_SYSTEM_ROOT is defined.
2334 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
2336 PR tree-optimization/79090
2337 PR tree-optimization/33562
2338 PR tree-optimization/61912
2339 PR tree-optimization/77485
2340 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
2341 and computed trims into the dump file.
2343 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
2345 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
2347 2017-01-16 Jakub Jelinek <jakub@redhat.com>
2350 * gimplify.c (gimplify_init_constructor): If want_value and
2351 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
2355 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
2356 sequence. Formatting fixes.
2357 (doloop_optimize): Formatting fixes.
2360 * gcc.c (debug_level_greater_than_spec_func): New function.
2361 (static_spec_functions): Add debug-level-gt spec function.
2362 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
2364 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2365 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
2366 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
2367 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
2368 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
2369 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
2371 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
2373 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
2374 QImode fixups to general and mask registers only.
2376 2017-01-16 Carl Love <cel@us.ibm.com>
2378 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
2379 for built-in functions
2380 vector signed char vec_nabs (vector signed char)
2381 vector signed short vec_nabs (vector signed short)
2382 vector signed int vec_nabs (vector signed int)
2383 vector signed long long vec_nabs (vector signed long long)
2384 vector float vec_nabs (vector float)
2385 vector double vec_nabs (vector double)
2386 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
2388 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
2389 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
2390 * doc/extend.texi: Update the documentation file for the new built-in
2393 2017-01-16 Martin Sebor <msebor@redhat.com>
2395 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
2398 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2400 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
2401 UNSPEC_VSX__XXSPLTD to require special splat handling.
2403 2017-01-16 David Malcolm <dmalcolm@redhat.com>
2406 * system.h: Poison strndup.
2408 2017-01-16 Alan Modra <amodra@gmail.com>
2411 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
2414 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
2416 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
2418 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
2420 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
2421 call recog here. Assert that INSN_CODE (insn) is non-negative.
2423 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2426 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
2428 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
2429 in the currently scheduled RTL fragment.
2431 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
2433 PR rtl-optimization/78751
2434 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
2437 2017-01-14 Jeff Law <law@redhat.com>
2439 PR tree-optimization/79090
2440 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
2441 variable length stores.
2442 (compute_trims): Delete dead assignment to *trim_tail.
2443 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
2446 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
2448 PR rtl-optimization/78626
2449 PR rtl-optimization/78727
2450 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
2451 of a block, and split such blocks after everything else is finished.
2453 2017-01-14 Alan Modra <amodra@gmail.com>
2456 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
2457 target legitimate_combined_insn.
2458 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
2459 (rs6000_legitimate_combined_insn): New function.
2460 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
2462 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
2463 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
2464 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
2466 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
2468 * doc/frontends.texi (G++ and GCC): Remove references to Java.
2470 2017-01-13 Jeff Law <law@redhat.com>
2472 PR tree-optimization/33562
2473 PR tree-optimization/61912
2474 PR tree-optimization/77485
2475 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
2477 (delete_dead_assignment): Likewise.
2478 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
2479 statement to delete_dead_call and delete_dead_assignment.
2481 2017-01-13 David Malcolm <dmalcolm@redhat.com>
2484 * substring-locations.c (format_warning_va): Strengthen case 1 so
2485 that both endpoints of the substring must be within the format
2486 range for just the substring to be printed.
2488 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
2490 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
2491 * config/i386/i386.c (ix86_target_string): Add missing options
2492 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
2493 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
2494 flags_other and ix86_target_other to flags2_other. Display unknown
2496 (ix86_valid_target_attribute_inner_p): Add missing options and
2497 reorder options by implied ISAs, as in ix86_target_string.
2499 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2501 * hash-table.h (hash_table::too_empty_p): New function.
2502 (hash_table::expand): Use it.
2503 (hash_table::traverse): Likewise.
2504 (hash_table::empty_slot): Use sizeof (value_type) instead of
2505 sizeof (PTR) to convert bytes to elements. Shrink the table
2506 if the current size is excessive for the current number of
2509 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
2511 * ira-costs.c (record_reg_classes): Break from the inner loop
2512 early once alt_fail is known to be true. Update outer loop
2513 handling accordingly.
2515 2017-01-13 Jeff Law <law@redhat.com>
2517 * tree-ssa-dse.c (decrement_count): New function.
2518 (increment_start_addr, maybe_trim_memstar_call): Likewise.
2519 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
2520 when we know the partially dead statement is a mem* function.
2522 PR tree-optimization/61912
2523 PR tree-optimization/77485
2524 * tree-ssa-dse.c: Include expr.h.
2525 (maybe_trim_constructor_store): New function.
2526 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
2528 PR tree-optimization/33562
2529 PR tree-optimization/61912
2530 PR tree-optimization/77485
2531 * doc/invoke.texi: Document new dse-max-object-size param.
2532 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
2533 * tree-ssa-dse.c: Include params.h.
2534 (dse_store_status): New enum.
2535 (initialize_ao_ref_for_dse): New, partially extracted from
2537 (valid_ao_ref_for_dse, normalize_ref): New.
2538 (setup_live_bytes_from_ref, compute_trims): Likewise.
2539 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
2540 (maybe_trim_partially_dead_store): Likewise.
2541 (maybe_trim_complex_store): Likewise.
2542 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
2543 Track what bytes live from the original store. Return tri-state
2544 for dead, partially dead or live.
2545 (dse_dom_walker): Add constructor, destructor and new private members.
2546 (delete_dead_call, delete_dead_assignment): New extracted from
2548 (dse_optimize_stmt): Make a member of dse_dom_walker.
2549 Use initialize_ao_ref_for_dse.
2551 PR tree-optimization/33562
2552 PR tree-optimization/61912
2553 PR tree-optimization/77485
2554 * sbitmap.h (bitmap_count_bits): Prototype.
2555 (bitmap_clear_range, bitmap_set_range): Likewise.
2556 * sbitmap.c (bitmap_clear_range): New function.
2557 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
2559 2017-01-13 Martin Liska <mliska@suse.cz>
2562 * function.c (set_cfun): Add new argument force.
2563 * function.h (set_cfun): Likewise.
2564 * ipa-inline-transform.c (inline_call): Use the function when
2565 strict alising from is dropped for function we inline to.
2567 2017-01-13 Richard Biener <rguenther@suse.de>
2569 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
2570 for dumping GIMPLE INTEGER_CSTs.
2572 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2574 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
2575 to 201112L since C++17.
2577 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
2580 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
2581 if -fsanitize=kernel-address is present.
2583 2017-01-13 Richard Biener <rguenther@suse.de>
2585 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
2586 as _Literal ( type ) number in case usual suffixes do not
2587 preserve all information.
2589 2017-01-13 Richard Biener <rguenther@suse.de>
2591 PR tree-optimization/77283
2592 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
2593 and ssa-iterators.h.
2594 (is_feasible_trace): Implement a cost model based on joiner
2597 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
2600 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
2601 char or short to __float128/_Float128 directly.
2603 2017-01-12 Martin Sebor <msebor@redhat.com>
2605 to -Wformat-overflow.
2606 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
2607 (min_bytes_remaining): Same.
2608 (get_string_length): Same.
2609 (format_string): Same.
2610 (format_directive): Same.
2612 (pass_sprintf_length::handle_gimple_call): Same.
2614 2017-01-12 Jakub Jelinek <jakub@redhat.com>
2616 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
2617 info.nowrite calls with no lhs that can't throw. Return bool
2618 whether gsi_remove has been called or not.
2619 (pass_sprintf_length::handle_gimple_call): Return bool whether
2620 try_substitute_return_value called gsi_remove. Formatting fix.
2621 (pass_sprintf_length::execute): Don't use gsi_remove if
2622 handle_gimple_call returned true.
2625 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
2626 be removed due to side-effects, don't remove following barrier nor
2627 turn the successor edge into fallthru edge.
2629 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2632 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
2633 element-reversing loads and stores as not swappable.
2635 2017-01-12 Nathan Sidwell <nathan@acm.org>
2636 Nicolai Stange <nicstange@gmail.com>
2638 * combine.c (try_combine): Don't ignore result of overlap checking
2639 loop. Combine overlap & asm check into single loop.
2641 2017-01-12 Richard Biener <rguenther@suse.de>
2643 * tree-pretty-print.c (dump_generic_node): Provide -gimple
2644 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
2646 2017-01-12 Richard Biener <rguenther@suse.de>
2648 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
2649 and TS_TARGET_OPTION directly derive from TS_BASE.
2650 * tree-core.h (tree_optimization_option): Derive from tree_base.
2651 (tree_target_option): Likewise.
2653 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
2655 * config/i386/i386.c (memory_address_length): Increase len
2656 only when rip_relative_addr_p returns false.
2658 2017-01-11 Julia Koval <julia.koval@intel.com>
2660 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
2661 (OPTION_MASK_ISA_SGX_SET): New.
2662 (ix86_handle_option): Handle OPT_msgx.
2663 * config.gcc: Added sgxintrin.h.
2664 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
2665 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
2666 * config/i386/i386.c (ix86_target_string): Add -msgx.
2668 (ix86_option_override_internal): Handle new options.
2669 (ix86_valid_target_attribute_inner_p): Add sgx.
2670 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
2671 * config/i386/i386.opt: Add msgx.
2672 * config/i386/sgxintrin.h: New file.
2673 * config/i386/x86intrin.h: Add sgxintrin.h.
2675 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2678 * fold-const.c (maybe_nonzero_address): Return 1 for function
2680 (tree_single_nonzero_warnv_p): Don't handle function local objects
2684 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
2687 2017-01-11 David Malcolm <dmalcolm@redhat.com>
2690 * opts.c: Include "spellcheck.h"
2691 (struct string_fragment): New struct.
2692 (struct edit_distance_traits<const string_fragment &>): New
2694 (get_closest_sanitizer_option): New function.
2695 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
2697 2017-01-11 Jakub Jelinek <jakub@redhat.com>
2699 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
2701 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
2702 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
2703 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
2704 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
2705 for initial die_offset if dwarf_split_debug_info.
2706 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
2707 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
2709 (output_skeleton_debug_sections): Formatting fix. Use
2710 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
2711 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
2713 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
2715 * config/arm/cortex-a53.md: Add bypasses for
2717 (cortex_a53_r2f): Only use for transfers.
2718 (cortex_a53_f2r): Likewise.
2719 (cortex_a53_r2f_cvt): Add reservation for conversions.
2720 (cortex_a53_f2r_cvt): Likewise.
2722 2017-01-11 Tamar Christina <tamar.christina@arm.com>
2724 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
2725 to all inlined functions, change static to extern.
2727 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
2730 * config/arm/arm.c (legitimize_pic_address): Handle reference to
2732 (arm_assemble_integer): Likewise.
2734 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
2736 * config.gcc: Use new awk script to check CPU, FPU and architecture
2737 parameters for --with-... options.
2738 * config/arm/parsecpu.awk: New file
2739 * config/arm/arm-cpus.in: New file.
2740 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
2742 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
2744 * config/arm/t-arm: Update dependency rules.
2745 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
2746 of processing .def files.
2747 * config/arm/genopt.sh: Deleted.
2748 * config/arm/gentune.sh: Deleted.
2749 * config/arm/arm-cores.def: Deleted.
2750 * config/arm/arm-arches.def: Deleted.
2751 * config/arm/arm-fpus.def: Deleted.
2752 * config/arm/arm-tune.md: Regenerated.
2753 * config/arm/arm-tables.opt: Regenerated.
2754 * config/arm/arm-cpu.h: New generated file.
2755 * config/arm/arm-cpu-data.h: New generated file.
2756 * config/arm/arm-cpu-cdata.h: New generated file.
2758 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
2761 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
2763 (input_varpool_node): Unpack dynamically_initialized bit.
2765 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
2767 PR rtl-optimization/79032
2768 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
2769 the alignment of the adjusted memory reference against that of MODE,
2770 instead of the alignment of the original memory reference.
2772 2017-01-11 Martin Jambor <mjambor@suse.cz>
2774 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
2776 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
2777 decorated functions.
2779 2017-01-11 Richard Biener <rguenther@suse.de>
2781 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
2782 set range/nonnull info for PHI results. Do not set it on
2783 stmts marked for removal.
2785 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
2787 * expr.c (store_field): In the bitfield case, fetch the return value
2788 from the registers before applying a single big-endian adjustment.
2789 Always do a final load for a BLKmode value not larger than a word.
2791 2017-01-10 David Malcolm <dmalcolm@redhat.com>
2794 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2795 that we correctly handle column numbers greater than
2796 LINE_MAP_MAX_COLUMN_NUMBER.
2798 2017-01-10 Martin Sebor <msebor@redhat.com>
2801 * gimple-ssa-sprintf.c (get_destination_size): Call
2802 {init,fini}object_sizes.
2803 * tree-object-size.c (addr_object_size): Adjust.
2804 (pass_through_call): Adjust.
2805 (pass_object_sizes::execute): Adjust.
2806 * tree-object-size.h (fini_object_sizes): Declare.
2808 2017-01-10 Martin Sebor <msebor@redhat.com>
2810 PR tree-optimization/78775
2811 * builtins.c (get_size_range): Move...
2812 * calls.c: ...to here.
2813 (alloc_max_size): Accept zero argument.
2814 (operand_signed_p): Remove.
2815 (maybe_warn_alloc_args_overflow): Call get_size_range.
2816 * calls.h (get_size_range): Declare.
2818 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2820 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2821 from TI's devices.csv file as of September 2016.
2822 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2824 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2826 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2827 * doc/invoke.texi: Likewise.
2828 * doc/md.texi: Likewise.
2829 * doc/objc.texi: Likewise.
2831 2017-01-10 Joshua Conner <joshconner@google.com>
2833 * config/arm/fuchsia-elf.h: New file.
2834 * config/fuchsia.h: New file.
2835 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2836 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2838 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2840 2016-01-10 Richard Biener <rguenther@suse.de>
2842 PR tree-optimization/79034
2843 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2844 Propagate out degenerate PHIs in the joiner.
2846 2017-01-10 Martin Liska <mliska@suse.cz>
2848 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2849 (sort_congruence_classes_by_decl_uid): Likewise.
2850 (sort_congruence_class_groups_by_decl_uid): Likewise.
2851 (sem_item_optimizer::merge_classes): Sort class, groups in these
2852 classes and members in the groups by DECL_UID of declarations.
2853 This would make merge operations stable.
2855 2017-01-10 Martin Liska <mliska@suse.cz>
2857 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2858 usage of m_classes_vec.
2859 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2860 (sem_item_optimizer::get_group_by_hash): Likewise.
2861 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2862 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2863 (sem_item_optimizer::verify_classes): Likewise.
2864 (sem_item_optimizer::process_cong_reduction): Likewise.
2865 (sem_item_optimizer::dump_cong_classes): Likewise.
2866 (sem_item_optimizer::merge_classes): Likewise.
2867 * ipa-icf.h (congruence_class_hash): Rename from
2868 congruence_class_group_hash. Remove declaration of m_classes_vec.
2870 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2872 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2873 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2874 * config.gcc: Add avx512vpopcntdqintrin.h.
2875 * config/i386/avx512vpopcntdqintrin.h: New.
2876 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2877 * config/i386/i386-builtin-types.def: Add new types.
2878 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2879 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2880 __builtin_ia32_vpopcountq_v8di_mask): New.
2881 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2882 __AVX512VPOPCNTDQ__.
2883 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2884 (PTA_AVX512VPOPCNTDQ): Define.
2885 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2886 TARGET_AVX512VPOPCNTDQ_P): Define.
2887 * config/i386/i386.opt: Add mavx512vpopcntdq.
2888 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2889 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2891 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2894 * predict.def (PRED_CALL): Set to 67.
2896 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2898 * expr.c (store_field): In the bitfield case, if the value comes from
2899 a function call and is of an aggregate type returned in registers, do
2900 not modify the field mode; extract the value in all cases if the mode
2901 is BLKmode and the size is not larger than a word.
2903 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2906 * config/i386/cpuid.h: Fix undefined behavior.
2908 2017-01-04 Jeff Law <law@redhat.com>
2910 PR tree-optimization/79007
2911 PR tree-optimization/67955
2912 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2913 conservative for pt.null when flag_non_call_exceptions is on.
2915 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2917 PR translation/79019
2918 PR translation/79020
2919 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2920 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2921 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2923 * config/avr/avr.opt (maccumulate-args): Likewise.
2924 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2925 * common.opt (freport-bug): Likewise.
2926 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2927 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2928 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2929 translatable string.
2930 * config/i386/i386.c (function_value_32): Likewise.
2931 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2932 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2934 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2935 * common/config/msp430/msp430-common.c (msp430_handle_option):
2937 * symtab.c (symtab_node::verify_base): Likewise.
2938 * opts.c (set_debug_level): Likewise.
2939 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2940 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2941 missing whitespace to translatable strings.
2942 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2943 * config/sh/superh.h: Likewise.
2944 * config/i386/xopintrin.h: Likewise.
2945 * config/i386/znver1.md: Likewise.
2946 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2947 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2948 * double-int.h (struct double_int): Likewise.
2949 * double-int.c (div_and_round_double): Likewise.
2950 * wide-int.cc: Likewise.
2951 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2952 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2953 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2954 (crossjumps_occurred): ... this.
2955 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2958 PR tree-optimization/78899
2959 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2960 returning bool return struct loop *, NULL for failure and the new
2962 (versionable_outer_loop_p): Don't version outer loop if it has
2963 dont_vectorized bit set.
2964 (tree_if_conversion): When versioning outer loop, ensure
2965 tree_if_conversion is performed also on the inner loop of the
2966 non-vectorizable outer loop copy.
2967 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2968 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2969 prevent vectorization of it.
2970 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2971 the outer loop vectorization of the non-scalar version is attempted
2972 before vectorization of the inner loop in scalar version. If
2973 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2974 vectorization of its inner loop.
2975 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2976 has 2 inner loops, rename also on edges from bb whose single pred
2977 is outer_loop->header. Fix typo in function comment.
2979 2017-01-09 Martin Sebor <msebor@redhat.com>
2982 * asan.c (asan_emit_stack_protection): Increase local buffer size
2983 to avoid snprintf truncation warning.
2985 2017-01-09 Andrew Pinski <apinski@cavium.com>
2987 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2988 to reference thunderx2t99 for the tuning structure
2989 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2991 (thunderx2t99_extra_costs): This.
2992 * config/aarch64/aarch64-tune.md: Regenerate.
2993 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2994 (vulcan_addrcost_table): This.
2995 (vulcan_regmove_cost): Rename to ...
2996 (thunderx2t99_regmove_cost): This.
2997 (vulcan_vector_cost): Rename to ...
2998 (thunderx2t99_vector_cost): this.
2999 (vulcan_branch_cost): Rename to ...
3000 (thunderx2t99_branch_cost): This.
3001 (vulcan_tunings): Rename to ...
3002 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
3003 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
3005 2017-01-09 Martin Jambor <mjambor@suse.cz>
3009 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
3010 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
3011 (propagate_vr_accross_jump_function): Use the above function for all
3012 value range computations for pass-through jump functions and type
3013 converasion from explicit value range values.
3014 (ipcp_propagate_stage): Do not attempt to deduce types of formal
3015 parameters from TYPE_ARG_TYPES.
3016 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
3017 (ipa_write_node_info): Stream type of the actual argument.
3018 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
3020 2017-01-09 Martin Liska <mliska@suse.cz>
3023 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
3024 (lookup_compiler): Do not show error message with have_E.
3026 2017-01-09 Jakub Jelinek <jakub@redhat.com>
3028 PR tree-optimization/78938
3029 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
3030 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
3031 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
3032 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
3035 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3037 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
3040 2017-01-09 Richard Biener <rguenther@suse.de>
3042 PR tree-optimization/78997
3043 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
3044 name condition properly.
3046 2017-01-09 Richard Biener <rguenther@suse.de>
3049 * dwarf2out.c (is_cxx): New overload with context.
3050 (is_naming_typedef_decl): Use it.
3052 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3054 * invoke.texi (Option Summary): Correct spacing in option lists
3055 and add line breaks to fix over-long lines.
3057 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3061 * extend.texi (Common Variable Attributes): Add xref to GCC
3062 Internals manual to explain mode attribute keywords.
3064 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
3067 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
3068 and Preprocessor Options.
3069 (Options for Linking): Document -pthread here....
3070 (RS/6000 and PowerPC Options): ...not here.
3071 (Solaris 2 Options): ...or here.
3072 * doc/cppopts.texi: Document -pthread.
3074 2017-01-08 Martin Sebor <msebor@redhat.com>
3077 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
3078 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
3079 New member functions.
3080 (format_directive): Used them.
3082 (pass_sprintf_length::handle_gimple_call): Same.
3083 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
3084 to avoid truncation for any argument.
3085 (extract_affine_mul): Same.
3086 * tree.c (get_file_function_name): Same.
3088 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3091 * predict.def (PRED_INDIR_CALL): Set to 86.
3093 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3095 PR preprocessor/54124
3096 * doc/cppopts.texi: Reformat -d subtable to list the full name
3097 of the options. Add cross-reference to the docs for the general
3098 compiler -d options.
3099 * doc/invoke.texi (Developer Options): Add cross-reference to the
3100 preprocessor-specific -d option documentation.
3102 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3104 PR preprocessor/13498
3105 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
3106 redudant material, and reflect new command-line options.
3107 (System Headers): Likewise.
3109 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
3111 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
3112 -isystem, and -idirafter. Copy-edit.
3113 * doc/cppopts.texi: Copy-edit. Remove contradiction about
3114 default for -ftrack-macro-expansion. Delete obsolete and
3115 badly-formatted implementation details about -fdebug-cpp output.
3116 * doc/cppwarnopts.texi: Copy-edit.
3118 2017-01-07 David Malcolm <dmalcolm@redhat.com>
3121 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
3122 that the transition from a max line width >= 1<<10 to narrower
3123 lines works correctly.
3125 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
3127 * doc/options.texi (PerFunction): New.
3128 * opt-functions.awk (switch_flags): Map both Optimization and
3129 PerFunction to CL_OPTIMIZATION.
3130 * opth-gen.awk: Test for PerFunction flag along with
3132 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
3133 it only when the latter is present. Skip those that don't in
3134 the hash function generator.
3135 * common.opt (fvar-tracking): Mark as PerFunction instead of
3137 (fvar-tracking-assignments): Likewise.
3138 (fvar-tracking-assignments-toggle): Likewise.
3139 (fvar-tracking-uninit): Likewise.
3141 2017-01-07 Jakub Jelinek <jakub@redhat.com>
3143 PR translation/79018
3144 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
3147 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
3150 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
3151 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
3152 TARGET_LONG_JUMP_TABLE_OFFSETS.
3153 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
3154 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
3155 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
3156 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
3157 * config/m68k/m68k.md (tablejump expander): Likewise.
3158 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
3159 TARGET_LONG_JUMP_TABLE_OFFSETS.
3160 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
3161 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
3163 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
3164 David Holsgrove <david.holsgrove@xilinx.com>
3166 * common/config/microblaze/microblaze-common.c
3167 (TARGET_EXCEPT_UNWIND_INFO): Remove.
3168 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
3170 * config/microblaze/microblaze.c (microblaze_must_save_register)
3171 (microblaze_expand_epilogue, microblaze_return_addr): Handle
3173 (microblaze_eh_return): New function.
3174 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
3175 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
3176 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
3177 * config/microblaze/microblaze.md (eh_return): New pattern.
3179 2017-01-06 Jakub Jelinek <jakub@redhat.com>
3181 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
3182 GCC_DIAGNOSTIC_STRINGIFY): Define.
3184 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
3186 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3188 * config/arm/arm.md (<mcrr>): New.
3190 * config/arm/arm.c (arm_arch5te): New.
3191 (arm_option_override): Set arm_arch5te.
3192 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
3194 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
3195 (arm_mcrr_qualifiers): ... this. New.
3196 (MRRC_QUALIFIERS): Define to...
3197 (arm_mrrc_qualifiers): ... this. New.
3198 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
3200 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
3201 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
3202 (MRRCI, mrrc, MRRC): New.
3203 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
3204 VUNSPEC_MRRC2): New.
3206 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3208 * config/arm/arm.md (<mcr>): New.
3210 * config/arm/arm.c (arm_coproc_builtin_available): Add
3211 support for mcr, mrc, mcr2 and mrc2.
3212 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
3213 (arm_mcr_qualifiers): ... this. New.
3214 (MRC_QUALIFIERS): Define to ...
3215 (arm_mrc_qualifiers): ... this. New.
3216 (MCR_QUALIFIERS): Define to ...
3217 (arm_mcr_qualifiers): ... this. New.
3218 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
3220 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
3221 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
3222 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
3225 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3227 * config/arm/arm.md (*ldc): New.
3231 * config/arm/arm.c (arm_coproc_builtin_available): Add
3232 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
3233 (arm_coproc_ldc_stc_legitimate_address): New.
3234 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
3235 'qualifier_const_pointer'.
3236 (LDC_QUALIFIERS): Define to...
3237 (arm_ldc_qualifiers): ... this. New.
3238 (STC_QUALIFIERS): Define to...
3239 (arm_stc_qualifiers): ... this. New.
3240 * config/arm/arm-protos.h
3241 (arm_coproc_ldc_stc_legitimate_address): New.
3242 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
3243 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
3244 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
3245 stc2, stcl, stc2l): New.
3246 * config/arm/constraints.md (Uz): New.
3247 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
3248 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
3249 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
3250 VUNSPEC_STC2L): New.
3252 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3254 * config/arm/arm.md (<cdp>): New.
3255 * config/arm/arm.c (neon_const_bounds): Rename this ...
3256 (arm_const_bounds): ... this.
3257 (arm_coproc_builtin_available): New.
3258 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
3259 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
3260 (CDP_QUALIFIERS): Define to...
3261 (arm_cdp_qualifiers): ... this. New.
3263 (arm_expand_builtin_args): Add case for 6 arguments.
3264 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
3265 (arm_const_bounds): ... this.
3266 (arm_coproc_builtin_available): New.
3267 * config/arm/arm_acle.h (__arm_cdp): New.
3269 * config/arm/arm_acle_builtins.def (cdp): New.
3271 * config/arm/iterators.md (CDPI,CDP,cdp): New.
3272 * config/arm/neon.md: Rename all 'neon_const_bounds' to
3274 * config/arm/types.md (coproc): New.
3275 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
3276 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
3277 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
3278 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
3280 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3282 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
3283 (UBINOP_QUALIFIERS): New.
3285 (acle_builtin_data): New. Change comment.
3286 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
3287 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
3288 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
3289 arm_acle_builtins.def.
3290 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
3291 (arm_init_acle_builtins): New.
3292 (CRC32_BUILTIN): Remove.
3293 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
3294 crc32cb, crc32ch and crc32cw.
3295 (arm_init_crc32_builtins): Remove.
3296 (arm_init_builtins): Use arm_init_acle_builtins rather
3297 than arm_init_crc32_builtins.
3298 (arm_expand_acle_builtin): New.
3299 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
3300 * config/arm/arm_acle_builtins.def: New.
3302 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
3304 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
3305 (arm_builtin_datum): ... this.
3306 (arm_init_neon_builtin): Rename to ...
3307 (arm_init_builtin): ... this. Add a new parameters PREFIX
3308 and USE_SIG_IN_NAME.
3309 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
3310 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
3311 'arm_builtin_datum'.
3312 (arm_init_vfp_builtins): Likewise.
3313 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
3314 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
3315 (arm_expand_neon_args): Rename to ...
3316 (arm_expand_builtin_args): ... this. Rename builtin_arg
3317 enum values and differentiate between ARG_BUILTIN_MEMORY
3318 and ARG_BUILTIN_NEON_MEMORY.
3319 (arm_expand_neon_builtin_1): Rename to ...
3320 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
3321 values, arm_expand_builtin_args and add bool parameter NEON.
3322 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
3323 (arm_expand_vfp_builtin): Likewise.
3324 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
3326 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3329 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
3330 * predict.c (tree_estimate_probability_bb): Reverse direction of
3331 polymorphic call predictor.
3333 2017-01-06 David Malcolm <dmalcolm@redhat.com>
3335 * passes.c (execute_one_pass): Split out pass-skipping logic into...
3336 (determine_pass_name_match): ...this new function and...
3337 (should_skip_pass_p): ...this new function.
3339 2017-01-06 Nathan Sidwell <nathan@acm.org>
3341 * ipa-visibility.c (function_and_variable_visibility): Reformat
3342 comments and long lines. Remove extrneous if.
3343 * symtab.c (symtab_node::make_decl_local): Fix code format.
3344 (symtab_node::set_section_for_node): Fix comment typo.
3346 2017-01-06 Martin Liska <mliska@suse.cz>
3349 * lra-constraints.c: Rename invariant to lra_invariant.
3350 * predict.c (set_even_probabilities): Initialize e to NULL.
3352 2017-01-05 Martin Sebor <msebor@redhat.com>
3354 PR tree-optimization/78910
3355 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
3356 (format_integer): Correct off-by-one error in the handling
3357 of precision with negative numbers in signed conversions..
3359 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
3361 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
3363 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3365 PR tree-optimization/71016
3366 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
3367 factor_out_conditional_conversion. Formatting fix.
3368 (factor_out_conditional_conversion): Add cond_stmt argument.
3369 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
3370 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
3373 2017-01-05 David Malcolm <dmalcolm@redhat.com>
3375 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
3376 read-rtl-function.o, and selftest-rtl.o.
3377 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
3378 (selftest::aarch64_test_loading_full_dump): New function.
3379 (selftest::aarch64_run_selftests): New function.
3380 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
3381 selftest::aarch64_run_selftests.
3382 * config/i386/i386.c
3383 (selftest::ix86_test_loading_dump_fragment_1): New function.
3384 (selftest::ix86_test_loading_call_insn): New function.
3385 (selftest::ix86_test_loading_full_dump): New function.
3386 (selftest::ix86_test_loading_unspec): New function.
3387 (selftest::ix86_run_selftests): Call the new functions.
3388 * emit-rtl.c (maybe_set_max_label_num): New function.
3389 * emit-rtl.h (maybe_set_max_label_num): New decl.
3390 * function.c (instantiate_decls): Guard call to
3391 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
3392 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
3394 * gensupport.c (gen_reader::gen_reader): Pass "false"
3395 for new "compact" param of rtx_reader.
3396 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
3397 rather than an empty string for NULL strings.
3398 * read-md.c: Potentially include config.h rather than bconfig.h.
3399 Wrap include of errors.h with #ifdef GENERATOR_FILE.
3400 (have_error): New global, copied from errors.c.
3401 (md_reader::read_name): Rename to...
3402 (md_reader::read_name_1): ...this, adding "out_loc" param,
3403 and converting "missing name or number" to returning false, rather
3405 (md_reader::read_name): Reimplement in terms of read_name_1.
3406 (md_reader::read_name_or_nil): New function.
3407 (md_reader::read_string): Handle "(nil)" by returning NULL.
3408 (md_reader::md_reader): Add new param "compact".
3409 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
3410 (md_reader::read_file): New method.
3411 * read-md.h (md_reader::md_reader): Add new param "compact".
3412 (md_reader::read_file): New method.
3413 (md_reader::is_compact): New accessor.
3414 (md_reader::read_name): Convert return type from void to file_location.
3415 (md_reader::read_name_or_nil): New decl.
3416 (md_reader::read_name_1): New decl.
3417 (md_reader::m_compact): New field.
3418 (noop_reader::noop_reader): Pass "false" for new "compact" param
3420 (rtx_reader::rtx_reader): Add new "compact" param.
3421 (rtx_reader::read_rtx_operand): Make virtual and convert return
3422 type from void to rtx.
3423 (rtx_reader::read_until): New decl.
3424 (rtx_reader::handle_any_trailing_information): New virtual function.
3425 (rtx_reader::postprocess): New virtual function.
3426 (rtx_reader::finalize_string): New virtual function.
3427 (rtx_reader::m_in_call_function_usage): New field.
3428 (rtx_reader::m_reuse_rtx_by_id): New field.
3429 * read-rtl-function.c: New file.
3430 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
3431 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
3432 (selftest::verify_three_block_rtl_cfg): New decl.
3433 * read-rtl-function.h: New file.
3434 * read-rtl.c: Potentially include config.h rather than bconfig.h.
3435 For host, include function.h, memmodel.h, and emit-rtl.h.
3436 (one_time_initialization): New function.
3437 (struct compact_insn_name): New struct.
3438 (compact_insn_names): New array.
3439 (find_code): Handle insn codes in compact dumps.
3440 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
3441 (bind_subst_iter_and_attr): Likewise.
3442 (add_condition_to_string): Likewise.
3443 (add_condition_to_rtx): Likewise.
3444 (apply_attribute_uses): Likewise.
3445 (add_current_iterators): Likewise.
3446 (apply_iterators): Likewise.
3447 (initialize_iterators): Guard usage of apply_subst_iterator with
3448 #ifdef GENERATOR_FILE.
3449 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
3450 (md_reader::read_mapping): Likewise.
3451 (add_define_attr_for_define_subst): Likewise.
3452 (add_define_subst_attr): Likewise.
3453 (read_subst_mapping): Likewise.
3454 (check_code_iterator): Likewise.
3455 (rtx_reader::read_rtx): Likewise. Move one-time initialization
3457 (one_time_initialization): New function.
3458 (rtx_reader::read_until): New method.
3459 (read_flags): New function.
3460 (parse_reg_note_name): New function.
3461 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
3462 Handle reuse_rtx ids.
3463 Wrap iterator lookup within #ifdef GENERATOR_FILE.
3464 Add parsing support for RTL dumps, mirroring the special-cases in
3465 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
3466 values, and calling handle_any_trailing_information.
3467 (rtx_reader::read_rtx_operand): Convert return type from void
3468 to rtx, returning return_rtx. Handle case 'e'. Call
3469 finalize_string on XSTR and XTMPL fields.
3470 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
3471 "(nil)" values were omitted. Call the postprocess vfunc on the
3473 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
3474 class ctor. Initialize m_in_call_function_usage. Call
3475 one_time_initialization.
3476 * rtl-tests.c (selftest::test_uncond_jump): Call
3477 set_new_first_and_last_insn.
3478 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
3479 * selftest-rtl.c: New file.
3480 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
3481 (selftest::get_insn_by_uid): New decl.
3482 * selftest-run-tests.c (selftest::run_tests): Call
3483 read_rtl_function_c_tests.
3484 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
3485 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
3488 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
3490 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
3491 operands in a special way. Assert that pos+len <= mode precision.
3493 2017-01-05 Jakub Jelinek <jakub@redhat.com>
3495 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
3496 3 argument Alias with unlimited for the negative form.
3497 (fno-vect-cost-model): Removed.
3499 2017-01-05 Martin Liska <mliska@suse.cz>
3501 * hsa-gen.c (gen_hsa_divmod): New function.
3502 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
3504 2017-01-05 Martin Liska <mliska@suse.cz>
3507 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
3510 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3512 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
3513 small constant length operands.
3515 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
3517 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
3518 between loop iterations.
3520 2017-01-05 Martin Liska <mliska@suse.cz>
3523 * gimplify.c (gimplify_decl_expr): Compare to
3524 asan_poisoned_variables instread of checking flags.
3525 (gimplify_target_expr): Likewise.
3526 (gimplify_expr): Likewise.
3527 (gimplify_function_tree): Conditionally initialize
3528 asan_poisoned_variables.
3530 2017-01-04 Jeff Law <law@redhat.com>
3532 PR tree-optimizatin/78812
3533 * rtl.h (contains_mem_rtx_p): Prototype.
3534 * ifcvt.c (containts_mem_rtx_p): Move from here to...
3535 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
3536 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
3537 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
3538 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
3540 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3542 * input.c (assert_char_at_range): Default-initialize actual_range.
3544 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3546 * df-scan.c (df_ref_create_structure): Make regno unsigned,
3547 to match the caller.
3549 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3551 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
3552 insns after final jump in test to emit dummy move.
3554 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3556 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
3557 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
3559 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
3561 * multiple_target.c (create_dispatcher_calls): Init e_next.
3562 * tree-ssa-loop-split.c (split_loop): Init border.
3563 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
3566 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
3571 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
3572 (altivec_register_operand): Do not return true if the operand
3573 contains a SUBREG mixing SImode and SFmode.
3574 (vsx_register_operand): Likewise.
3575 (vsx_reg_sfsubreg_ok): New predicate.
3576 (vfloat_operand): Do not return true if the operand contains a
3577 SUBREG mixing SImode and SFmode.
3578 (vint_operand): Likewise.
3579 (vlogical_operand): Likewise.
3580 (gpc_reg_operand): Likewise.
3581 (int_reg_operand): Likewise.
3582 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
3583 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
3584 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
3586 (rs6000_emit_move_si_sf_subreg): New helper function.
3587 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
3588 fixup SUBREGs involving SImode and SFmode.
3589 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
3590 numbers for the new peephole2 optimization.
3591 (peephole2 for SFmode unions): New peephole2 to optimize cases in
3592 the GLIBC math library that do AND/IOR/XOR operations on single
3593 precision floating point.
3594 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
3595 target macros to say whether we need to avoid SUBREGs mixing
3597 (TARGET_ALLOW_SF_SUBREG): Likewise.
3598 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
3599 (UNSPEC_SI_FROM_SF): Likewise.
3600 (iorxor): Change spacing.
3601 (and_ior_xor): New iterator for AND, IOR, and XOR.
3602 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
3603 (movdi_from_sf_zero_ext): Likewise.
3604 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
3605 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
3606 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
3607 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
3608 (fms<mode>4): Likewise.
3609 (fnma<mode>4): Likewise.
3610 (fnms<mode>4): Likewise.
3611 (nfma<mode>4): Likewise.
3612 (nfms<mode>4): Likewise.
3614 2017-01-04 Marek Polacek <polacek@redhat.com>
3617 * doc/invoke.texi: Document -Wpointer-compare.
3619 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3621 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
3624 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
3625 descriptions for -gdwarf-5 and emit them as uleb128 instead of
3628 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
3631 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
3632 documentation of the powerpc_popcntb_ok attribute.
3633 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
3634 code to issue warning messages if a requested CPU configuration is
3635 not supported by the binary (assembler and loader) toolchain.
3636 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
3637 made to define a built-in function that has been disabled.
3638 (paired_init_builtins): Add assertion to prevent ICE if attempt is
3639 made to define a built-in function that has been disabled.
3640 (altivec_init_builtins): Add comment explaining why definition
3641 of the DST built-in functions is not preceded by an assertion
3642 check. Add assertions to prevent ICE if attempts are made to
3643 define an altivec predicate or an abs* built-in function that has
3645 (htm_init_builtins): Add comment explaining why definition of the
3646 htm built-in functions is not preceded by an assertion check.
3648 2017-01-04 Jeff Law <law@redhat.com>
3650 PR tree-optimizatin/67955
3651 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
3652 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
3653 the points-to solution does not include pt_null. Use DECL_PT_UID
3656 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
3658 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
3659 Use gen_int_mode instead of gen_lopwart for const_int operands.
3661 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3663 PR tree-optimization/71563
3664 * match.pd: Simplify X << Y into X if Y is known to be 0 or
3665 out of range value - has low bits known to be zero.
3667 2017-01-04 Alan Modra <amodra@gmail.com>
3669 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
3670 * configure: Regenerate.
3671 * config.in: Regenerate.
3673 2017-01-04 Jakub Jelinek <jakub@redhat.com>
3676 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
3677 a substring of the message, but strcmp with the whole message. Ifdef
3678 ENABLE_NLS, translate the message first using dgettext.
3680 2017-01-03 Jeff Law <law@redhat.com>
3682 PR tree-optimizatin/78856
3683 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
3684 (mark_threaded_blocks): Remove code to truncate thread paths that
3685 cross multiple loop headers. Instead invalidate the cached loop
3686 iteration information and handle case of a thread path walking
3687 into an irreducible region.
3689 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3692 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
3693 assertions. Add support for doing the signbit if the IEEE 128-bit
3694 floating point value is in a GPR.
3695 * config/rs6000/rs6000.md (Fsignbit): Delete.
3696 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
3697 Update the length attribute if the value is in a GPR.
3698 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
3699 the sign or zero extension instruction, since the value is always 0/1.
3700 (signbit<mode>2_dm2): Delete using <Fsignbit>.
3703 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
3704 extracting SImode to a GPR register so that we can generate a
3705 store, limit the vector to be in a traditional Altivec register
3706 for the vextuwrx instruction.
3708 2017-01-03 Ian Lance Taylor <iant@google.com>
3710 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
3712 2017-01-03 Martin Sebor <msebor@redhat.com>
3714 PR tree-optimization/78696
3715 * gimple-ssa-sprintf.c (format_floating): Correct handling of
3716 precision. Use MPFR for %f for greater fidelity. Correct handling
3718 (pass_sprintf_length::compute_format_length): Set width and precision
3719 specified by asrerisk to void_node for vararg functions.
3720 (try_substitute_return_value): Adjust dump output.
3722 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
3724 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
3726 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
3728 * doc/invoke.texi (SPARC options): Document -mlra as the default.
3729 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
3730 -mlra/-mno-lra was passed to the compiler.
3732 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
3734 PR rtl-optimization/65618
3735 * emit-rtl.c (try_split): Move initialization of "before" and
3736 "after" to just before the call to emit_insn_after_setloc.
3738 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
3740 * doc/md.texi (Standard Names): Remove reference to Java frontend.
3742 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
3744 * dwarf2out.c (gen_enumeration_type_die): When
3745 -gno-strict-dwarf, add a DW_AT_encoding attribute.
3747 2017-01-03 Jakub Jelinek <jakub@redhat.com>
3749 PR tree-optimization/78965
3750 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
3751 Change first argument from const call_info & to call_info &. For %n
3752 set info.nowrite to false.
3755 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
3756 possibly throwing calls.
3758 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
3759 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
3760 and fns handling, rather than in a separate case SSA_NAME.
3762 2017-01-02 Jeff Law <law@redhat.com>
3764 * config/darwin-driver.c (darwin_driver_init): Const-correctness
3765 fixes for first_period and second_period variables.
3767 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
3770 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
3771 (*insvqi_1): New insn pattern.
3772 (*insvqi_1_mem_rex64): Ditto.
3774 (*insvqi_3): Rename from *insvqi.
3776 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
3778 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3780 * doc/cfg.texi (Edges): Remove reference to Java.
3781 (Maintaining the CFG): Ditto.
3783 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3786 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
3787 transparent aliases.
3789 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
3792 * predict.def (PRED_CALL): Update hitrate.
3793 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
3794 * predict.c (tree_estimate_probability_bb): Split CALL predictor
3795 into direct/indirect/polymorphic variants.
3797 2017-01-01 Jakub Jelinek <jakub@redhat.com>
3799 Update copyright years.
3801 * gcc.c (process_command): Update copyright notice dates.
3802 * gcov-dump.c (print_version): Ditto.
3803 * gcov.c (print_version): Ditto.
3804 * gcov-tool.c (print_version): Ditto.
3805 * gengtype.c (create_file): Ditto.
3806 * doc/cpp.texi: Bump @copying's copyright year.
3807 * doc/cppinternals.texi: Ditto.
3808 * doc/gcc.texi: Ditto.
3809 * doc/gccint.texi: Ditto.
3810 * doc/gcov.texi: Ditto.
3811 * doc/install.texi: Ditto.
3812 * doc/invoke.texi: Ditto.
3814 Copyright (C) 2017 Free Software Foundation, Inc.
3816 Copying and distribution of this file, with or without modification,
3817 are permitted in any medium without royalty provided the copyright
3818 notice and this notice are preserved.